2014 Book BiologicallyInspiredDesign

Ashok K.
Goel
Daniel A. McAdams
Robert B. Stone Editors
Biologically Inspired
Design
Computational Methods and Tools
Biologically Inspired Design
Ashok K. Goel Daniel A. McAdams
•
Robert B. Stone
Editors
Biologically Inspired
Design
Computational Methods and Tools
123
Editors
Ashok K. Goel Robert B. Stone
Georgia Institute of Technology School of Mechanical, Industrial and
Atlanta, GA Manufacturing Engineering
USA Oregon State University
Corvallis, OR
Daniel A. McAdams USA
Department of Mechanical Engineering
Texas A&M University
College Station, TX
USA
ISBN 978-1-4471-5247-7 ISBN 978-1-4471-5248-4 (eBook)

DOI 10.1007/978-1-4471-5248-4
Springer London Heidelberg New York Dordrecht
Library of Congress Control Number: 2013943573
Ó Springer-Verlag London 2014

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief
excerpts in connection with reviews or scholarly analysis or material supplied specifically for the
purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the
work. Duplication of this publication or parts thereof is permitted only under the provisions of
the Copyright Law of the Publisher’s location, in its current version, and permission for use must
always be obtained from Springer. Permissions for use may be obtained through RightsLink at the
Copyright Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)

In memory of my late parents, Urmil Goel
and Satya Prakash Goel
Ashok K. Goel
I would like to thank the students I have had

the opportunity and pleasure to teach. Their
constant query of what, why, and how has
kept me curious and energized to pursue my
own line of what, why, and how
Daniel A. McAdams
Foreword: Curating Nature’s Patent Database
The Moment of Truth
It is an odd hobby for a biologist, I admit, but I enjoy reading patents. Patents tell
the meticulous story of how humans have solved the conundrums of their era, from
catching mice to circling planets. This record of ingenuity is more than a legal
necessity; it is an inspiration. Inventors read patents not just to avoid reinventing
the wheel, but also to glimpse, from a mesa of inventive shoulders, the adjacent
possible.
The volume you are about to read describes a Googlesque quest to develop
another kind of patent database, one that describes nature’s 3.8 billion years of
adaptations. These adaptations are a record of life’s long march to become well
adapted to the particularities of this planet. While biologists ponder how adapta-
tions help individual lilies, plankton, and pelicans survive, biomimics ask: ‘‘How
might this adaptation, and the technology it inspires, help the human species fit in
here over the long haul?’’
In the last few decades, life’s adaptations have inspired a series of game-
changing technologies. A refrigeration-free vaccine inspired by the rugged
Tardigrade, a coral-inspired way to sequester tons of carbon dioxide in concrete,
and a material that captures fog as cleverly as a desert-dwelling beetle. Biomimics
are working on ways to reduce pesticides in farmer’s fields, ease traffic jams in
cities, and prevent antibiotic resistance in our hospitals. Biomimetic products are
doubling each year, and papers published in the field are doubling every 2–3 years,
much faster than the 13-year doubling rate of other sciences (Lepora 2013).
A 2010 economic study predicted that biomimicry could represent $1 trillion of
global Gross Domestic Product by 2025,1 and in 2012, biomimicry topped the
1
Fermanian Business & Economic Institute, Point Loma Nazarene University. Global
biomimicry efforts: an economic game changer, (2010).
vii
viii Foreword: Curating Nature’s Patent Database
Society of Manufacturing Engineers’ annual list of ‘‘innovations that could change

the way you manufacture.’’2
Though the prospects for bioinspired design have never been better, the dis-
cipline’s moment of truth is here as well. A chasm still exists for many innovators,
and unless we can cross it, biomimicry will remain the domain of the few inno-
vators skilled and interested enough to decipher the primary biological literature.
This leaves a knowledge divide for the millions of non-biologically trained
engineers, architects, product designers, planners, chemists, material scientists,
even policy makers for whom nature’s strategies would be a revelation.
It is not that relevant biological information does not exist; we are, in fact,
awash in it. If you can afford to access the full-text literature databases, and if you
are fluent in the jargon, you have a chance of keeping up with the science. But as a
designer, you are apt to have neither literature access nor a Rosetta Stone. For
biomimicry to realize its potential, we need a Biological Information System (BIS)
that is as ubiquitous and accessible as the Geographic Information System (GIS).
That platform needs to deliver curated knowledge at the moment of creation, in a
form tailored to fit the working styles of the people who invent our world. Like
GIS, the success of BIS will depend on software tools that intelligently make sense
of the raw data, augmented by apps that further extend its usefulness.
Transferable Ideas and Downloadable Beaks
Building a biological intelligence tool for inventors is the quintessential exercise in

spanning disciplines. Biologists and inventors not only speak different lan-
guages—they ask different questions. Biologists might write a paper about the
evolutionary significance of sharkskin that reduces biofouling, but often the
‘‘how’’ information—the dimensions and placement of the denticles—is buried in
a paper devoted to the ‘‘why.’’ Uncovering these gems of innovation from the
continual blizzard of papers is a challenge, requiring enabling technologies like
those collected in this volume: a way to describe biological phenomenon in
machine readable language, an engineering-to-biology thesaurus, natural language
query, near-clairvoyant search algorithms, and more.
Mining the literature, even clairvoyantly, is just the beginning. My biomimicry
consulting colleagues at Biomimicry 3.8 have spent 15 years bringing biology to
the design tables of companies such as Boeing, General Electric, General Mills,
HOK, Nike, InterfaceFLOR, and Procter and Gamble. We have learned that most
inventors are not interested in reading biological papers. They prefer that we
synthesize and translate the papers into a taxonomy of promising mechanisms.
Ultimately, they want a set of transferable ideas—design principles that will help
them approach their challenge in a completely novel way. Our researchers can
2
Society of Manufacturing Engineers http://www.sme.org/innovations12/#biomimicry.
Foreword: Curating Nature’s Patent Database ix
easily read 10,000 papers to answer a question such as ‘‘How does nature contain
liquids?’’ ‘‘How does nature manage vibration?’’ ‘‘How does nature store energy?’’
Building a taxonomy and extracting the design principles is a skill that takes years
to master.
Once inventors are equipped with bioinspired design principles, there are still
miles to go before these are translated into a product or process. This is where
interactive tools could help, walking inventors through an iterative design process
and giving them access to nature’s ideas every step of the way. How, where, and
when in the creative process this knowledge is delivered will mean the difference
between inspiration and execution. Ideally, actionable plug-ins will be accessible
right from the digital screens that designers, engineers, and architects use every-
day, e.g., an AskNature button embedded in CAD/CAM or BIM tools.
While designing a roofing system, for instance, a building engineer would be
able to visually browse reinforcement strategies in the natural world, and down-
load actual truss designs based on this information. While laying out the HVAC
system, he or she could run a branching algorithm to generate a fluid distribution
system that keeps frictional losses to a minimum. Framing designs could be light
weighted with the use of software that equalizes stress along surfaces, inspired by
the growth of trees and bones. A genetic algorithm that uses natural selection
protocols could optimize the entire design, all within the same program.
These digital modules are what our colleagues at Autodesk have described as
the difference between ‘‘concept’’ and ‘‘content.’’ Rather than read about a con-
cept, inventors want to access biological information as content that they can use
immediately. They would like to be able to download a biological library of
forms—3D models of life’s most streamlined, lightweight, or multifunctional
designs. Imagine if Eiji Nakatsu, the JR West engineer who mimicked the king-
fisher’s beak to create Japan’s Shinkansen train, had been able to download a 3D
model of the beak before building a physical model. He could have attached the
beak model to the train body, stretched and scaled it, even tested it in silico with
computational fluid dynamic tools.
Building a biological library of forms would help biologists as well as inven-
tors. With today’s reality-capture software for cameras, it is possible to imagine
‘‘scan jams’’ where volunteers would digitize the artifacts of the world’s natural
history museums, freeing them from molding drawers so they can enliven the next
generation of sustainable designs.
The internal blueprints of biomaterials will prove equally important, especially
as we move to computer-controlled additive manufacturing (3D printing).
Organisms add structure to common polymers to achieve extraordinary func-
tionality, e.g., beetles layer chitin composites in a plywood hatch to achieve
strength and toughness. A different structural design is used to create color,
resilience, or water repellency, all from the same material. A biomimetic structure-
function catalog could allow additive manufacturers to streamline their supply
chain as nature does, using a small palette of easily recyclable polymers in unique
architectures to achieve a wide range of functions.
x Foreword: Curating Nature’s Patent Database
Of course, each discipline will have different inventive needs. Chemists might
prefer a ‘‘substitution engine’’ that allows them to replace an industrial chemical
synthesis with a biochemical alternative, achieving similar effect without waste or
toxic by-products. Organizational managers will want yet another slice of bio-
logical information, pertaining to topics like communication, cooperation, net-
works, or resilience. For each category of human endeavor, new user-centric
applications will need to be created atop the BIS data.
Helping Innovators Meet Their Mentor
At the end of the day, even the cleverest information tools will not guarantee that a
new invention, even one inspired by nature, will be sustainable in terms of energy
and material use, toxicity, end-of-life fate, etc.
To help innovators create in ways that are deeply biomimetic, we find it useful
to use systems-thinking tools such as Biomimicry 3.8’s Life’s Principles3 in the
scoping, creating, and evaluating phases of invention. These are meta principles
common to most species on earth, and include reminders such as ‘‘build from the
bottom up,’’ ‘‘use a safe subset of the periodic table,’’ ‘‘perform chemistry in
aqueous solution,’’ ‘‘embed feedback mechanisms to continually evolve your
design,’’ etc. Interactive software tools that screen for how well a design is meeting
Life’s Principles could help innovators solve problems without creating new ones.
If you look at all the ways that nature can influence decision-making, you
realize that biomimicry is more than just a new way to innovate. It is a new way to
think. University and professional training courses that prepare designers and
engineers to ask ‘‘How would nature solve this?’’ are vital, as are techniques,
described here, that help students make the all-important cognitive leap from
design principle to application. The professors pioneering in this field are in a
unique role; they have an opportunity to encourage the highest and best use of this
new and powerful methodology, hopefully to solve the worthy conundrums of our
era.
A prescient Steve Jobs said: ‘‘I think the biggest innovations of the twenty-first
century will be at the intersection of biology and technology. A new era is
beginning.’’ If the age of biology is to keep its promise, the people who make our
world will need to become biologically literate. But they will not want to become
biologists themselves. Instead, they will want to know the key principles, the best
practices, the operating codes of the natural world. They will want to understand
ubiquitous patterns as well as the strange and wonderful curiosities in nature’s
patent database. Ultimately, they will want to understand how life has managed to
enhance this planet, and how our innovations might do the same.
3
http://biomimicry.net/about/biomimicry/lifes-principles/
Foreword: Curating Nature’s Patent Database xi
A full-function tool to bring biological wisdom into human design is on its way,
and the people in this volume will be instrumental in delivering it to us. They
know that the key to wide-scale adoption of biomimicry is user-centric, curated
knowledge, available at the moment of creation. Their efforts to help innovators
learn from and emulate other species will one day be remembered as a pivotal leap
in the evolution of our own.
Reference
Lepora, N. F., Ver schure, P. & Prescott, T. J. The state of the art in biomimetics.
Bioinspir. Biomim. 8, 013001 (2013)
Janine Benyus
Co-Founder and Institute Board President,
Biomimicry 3.8
Preface
Biologically inspired design is a promising paradigm for design innovation as well

as sustainable design. The scientific challenge now is to transform it into a
repeatable and scalable methodology. This requires addressing several big chal-
lenges, including the following four: the first and foremost of course is to use the
paradigm to address increasing numbers of real problems that translate into real
products in the market. A second challenge is to document the best practices of
successful applications of the paradigm and develop a theory of biologically
inspired design. A third challenge is to develop computational methods and tools
that can make biologically inspired design repeatable and scalable. A fourth
challenge is to educate new generations of would-be-designers in the paradigm of
biologically inspired design. These four challenges are interconnected and build on
one another: success at one likely will spur success at others.
This volume brings together a dozen chapters that together address all four of
the above challenges at least to some degree, while emphasizing computational
methods and tools for biologically inspired design. We are pleased to bring
together these articles by some of the leading researchers and practitioners of
biologically inspired design into a single volume.
Chapter 1 provides a brief review of two workshops sponsored by the United
States National Science Foundation (NSF). These workshops served as the initial
catalysis and formation of this book. Taken together, the two workshops brought
together some 50 researchers in biologically inspired design, helped establish a
stronger sense of research community, and led to the formulation of a research
agenda outlined in the chapter.
Chapter 2 by Jon-Michael Deldin and Megan Schuknecht describe AskNature,
Biomimicry 3.8 Institute’s publicly available webportal that provides a function-
ally indexed database of biological design strategies and systems. Insofar as we
know, this is the first scholarly article describing AskNature in detail, and thus
adds an important piece to the growing literature on biologically inspired design.
In Chap. 3, Li Shu and Hyunmin Cheong describe a natural language approach
to finding biological analogies and applying them to design problems. They review
a decade long research program on developing the natural language approach to
biologically inspired design, and also provide several examples of its application.
xiii
xiv Preface
In Chap. 4, Jacquelyn Nagel presents an engineering-to-biology thesaurus,

along with several examples of its use in addressing design problems. This kind of
thesaurus can be a very useful tool for designers in finding biological analogies to
their design problems.
Nagel, McAdams, and Stone in Chap. 5 describe the big picture of their several
years of research on biologically inspired design. In particular, they present their
information-processing theory of biologically inspired design, and illustrate it with
several examples. They also describe a suite of tools that match several tasks in the
process theory.
Goel, Swaroop Vattam, Bryan Wiltgen, and Michael Helms in Chap. 6 present
their information-processing theory of biologically inspired design. They also
compare their theory with similar theories such as Design Spiral and BioTRIZ, and
examine what makes biologically inspired design different from other paradigms.
In Chap. 7, Jeannette Yen, Helms, Goel, Craig Tovey, and Marc Weissburg
describe the evolution of a college-level interdisciplinary course on biologically
inspired design. Their chapter reviews many lessons from teaching the course for
several years. These lessons should be useful for potential teachers of similar
courses.
In Chap. 8, Amaresh Chakrabarti focuses on analogical transfer from biology to
engineering. He proposes guidelines for supporting this analogical transfer and
describes an interactive tool that implements the guidelines. Comparative studies
indicate that use of the tool increases the number of transferred designs.
Julie Linsey and Vimal Viswanathan in Chap. 9 study the cognitive challenges
in biologically inspired design, focusing on design fixation. They describe several
heuristics for addressing the challenges, including fixation. These heuristics should
be useful for designing interactive tools for supporting biologically inspired
design.
In Chap. 10, Wojciech Bejgerowski, John Gerdes, James Hopkins, Lengfeng
Lee, Madusudanan Sathia Narayanan, Frank Mendel, Venkat Krovi, and Satyandra
Gupta focus on bioinspired robotics. Building on several years of research, they
present case studies ranging from bird-inspired robots to snake-inspired robots.
They also describe a process by which biological features are selected and sim-
plified for application using existing technologies for robot construction.
Julian Vincent in Chap. 11 specifies a need for identifying design principles that
may help produce good technical designs without requiring biological expertise.
He proposes four such principles derived from TRIZ: local quality, merging,
dynamics, and prior cushioning. He calls for identification of design principles,
especially those pertaining to information and material.
Frank Fish and John Beneski in Chap. 12 analyze the limits of design by
evolution in biology and design by analogy in biologically inspired design. They
argue that biological designs are not necessarily optimal with respect to any
specific function. They advocate focusing on biological features that outperform
currently available technologies for incorporation into technical designs.
Preface xv
We are grateful to Janine Benyus for writing the Foreword to this volume. Her
work at Biomimicry 3.8 has long inspired the biomimicry movement, including
our own efforts.
We are grateful also to the US NSF for sponsoring the workshops that led to
this volume. In particular, we thank Christina Bloebaum for her support of this
work as a Director of the Engineering Design and Innovation Program (now
known as Engineering Systems Design) of the Civil, Mechanical, and Manufac-
turing Innovation Division of the Engineering Directorate at NSF.
We also thank the Springer publishing company, and especially Grace Quinn at
Springer, for working with us on the preparation of this volume. We can only hope
that scholars and practitioners of biologically inspired design, as well as design
teachers and students more generally, will find this volume useful.
Ashok K. Goel
Daniel A. McAdams
Robert B. Stone
Contents
1 Charting a Course for Computer-Aided Bio-Inspired Design . . . . 1

Robert B. Stone, Ashok K. Goel and Daniel A. McAdams
2 The AskNature Database: Enabling Solutions

in Biomimetic Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Jon-Michael Deldin and Megan Schuknecht
3 A Natural Language Approach to Biomimetic Design . . . . . . . . . 29

L. H. Shu and Hyunmin Cheong
4 A Thesaurus for Bioinspired Engineering Design. . . . . . . . . . . . . 63

Jacquelyn K. S. Nagel
5 Function-Based Biologically Inspired Design . . . . . . . . . . . . . . . . 95

Jacquelyn K. S. Nagel, Robert B. Stone
and Daniel A. McAdams
6 Information-Processing Theories of Biologically

Inspired Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Ashok K. Goel, Swaroop Vattam, Bryan Wiltgen
and Michael Helms
7 Adaptive Evolution of Teaching Practices in Biologically

Inspired Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Jeannette Yen, Michael Helms, Ashok Goel, Craig Tovey
and Marc Weissburg
8 Supporting Analogical Transfer in Biologically

Inspired Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Amaresh Chakrabarti
xvii
xviii Contents
9 Overcoming Cognitive Challenges in Bioinspired Design

and Analogy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Julie S. Linsey and Vimal K. Viswanathan
10 An Engineering Approach to Utilizing Bio-Inspiration

in Robotics Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Saytandra K. Gupta, Wojciech Bejgerowski, John Gerdes,
James Hopkins, Lengfeng Lee, Madusudanan Sathia Narayanan,
Frank Mendel and Venkat Krovi
11 An Ontology of Biomimetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

Julian F. V. Vincent
12 Evolution and Bio-Inspired Design: Natural Limitations . . . . . . . 287

Frank E. Fish and John T. Beneski
Author Biographies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
Chapter 1
Charting a Course for Computer-Aided
Bio-Inspired Design
Robert B. Stone, Ashok K. Goel and Daniel A. McAdams
Abstract Bio-inspired design (BID) is an emerging research area in design,

biology, computing, and engineering that seeks to systematically mine biological
knowledge to solve design problems. To promote BID research, and especially
research on computer-aided BID, the United States National Science Foundation
(NSF) recently sponsored two workshops. These workshops served as the catalysis
for this book. In this chapter, we review the discussions at the two workshops. We
also sketch the outline of a research program on computer-aided BID that emerged
from the workshops.

Keywords Biomimicry Biologically inspired design Computer-aided design

Engineering design Design computing
1.1 Introduction
Bio-inspired design (BID or biomimicry or bionics) is an emerging research area

in engineering design, computing, and biology that seeks to systematically mine
biological knowledge to solve existing design problems. However, the community
R. B. Stone (&)
School of Mechanical, Industrial, Oregon State University, 204 Rogers Hall,
Corvallis 97331, USA
e-mail: rob.stone@oregonstate.edu
A. K. Goel
School of Interactive Computing, Georgia Institute of Technology, Atlanta 30332, USA
e-mail: goel@cc.gatech.edu
D. A. McAdams
Department of Mechanical Engineering, Texas A&M University, 411 Mechanical
Engineering Office, 3123, College Station 77843-3123, USA
e-mail: dmcadams@tamu.edu
A. K. Goel et al. (eds.), Biologically Inspired Design, 1

DOI: 10.1007/978-1-4471-5248-4_1, Springer-Verlag London 2014
2 R. B. Stone et al.
of BID researchers at present is fragmented with no professional society, unifying

funding source, or recurring conference.
As it stands, BID research is active across many disciplines and has had
important and significant results. Nevertheless, BID remains largely a research
activity contained in universities. BID is not yet an activity practiced by design
engineers in the field. This ‘‘research-not-practice’’ status of BID exists because
most BID is the result of researchers studying a biological entity or system such
that their level of understanding allows an almost direct emulation but not nec-
essarily new inspiration. This point solution status of BID sets a large, if a little ill-
defined, scope for BID research: How do we transform BID from a point solution
effort to fundamental theories and methods?
As a first draft of scoping BID research, some initial problems are apparent that
served as motivation and agenda items for a workshop. How do we motivate and
facilitate the interaction between engineering-oriented design researchers and
science-oriented biology researchers—two fields with significantly different
research cultures? How do we scope and scale biology in the context of abstracting
engineering knowledge? How do we transmit biological knowledge to engineers
and engineering knowledge to biologists without extensive discipline-specific
training? How might computational theories of analogical reasoning inform the
transfer of knowledge from biology to engineering and vice versa? How might
artificial intelligence theories of knowledge representation and ontology support
the construction of shared mental models among engineers and biologists? How
might we teach biologically inspired design to engineers and biologists alike?
Of particular research interest to BID is how can we mine biology for solutions
to problems for which we have no current solution. For example, some biological
solutions exhibit superior sustainability to engineered solutions. Similarly, bio-
logical solutions often are complex both in their solution and the problem they
solve. As the design problems solved become more complex, and the engineered
solutions themselves also become more complex, biology may offer insights into
how to solve the problem.
In this chapter, we provide an overview of two national science foundation
(NSF)-sponsored workshops that started a discussion of these issues. The sub-
sequent chapters in this volume discuss some of these issues in detail.
1.1.1 NSF Workshops
National science foundation funded two workshops to chart a course for BID.
First, a one-day workshop in Palo Alto, California, was held on Sunday, March 20,
2011, in conjunction with the AAAI 2011 Spring Symposium on Artificial Intel-
ligence and Sustainable Design.1 A second, follow-up half-day workshop was held
1
http://designengineeringlab.org/BID-workshop/Workshop_1.html
1 Charting a Course for Computer-Aided Bio-Inspired Design 3
Table 1.1 Presenters and presentation titles at workshop 1

Speaker Title
Janine Benyus Biomimicry: sustainable design inspired by nature
Biomimicry guild/Biomimicry
Institute
Satyandra Gupta Bio-inspired robotics
University of Maryland
Daniel A. McAdams Bio-inspired design
Texas A&M University
Ashok Goel Computational methods and tools for biologically inspired
Georgia Institute of Technology design
Julian Vincent Methods of bridging biological information and engineering
University of Bath information
David Rosen SSS approach and creativity metrics for bio-inspired
Georgia Institute of Technology
Shapour Azarm Bio-inspired design optimization
University of Maryland
Thomasz Arciszewski Bio-inspiration
George Mason University
Jeannette Yen Center for biologically-inspired design
Amaresh Chakrabarti Biologically inspired design: an overview of research at ideas
Indian Institute of Science lab, IISc
in College Station, Texas, on June 5, 2012, in conjunction with the Fifth Inter-
national Conference on Design Computing and Cognition.2
The first workshop was a good match to the AAAI 2011 Spring Symposium as the
symposium had a topic on the related issue of biologically inspired and evolutionary
models of sustainable design.3 The objectives for the first workshop were to
1. Identify the research community.
2. Identify the topical coverage (overlap/gaps) of research.
3. Explore the gaps in current research efforts as it relates to a systematic appli-
cation of biological information for engineering design.
4. Formulate major themes under the BID research banner.
5. Articulate steps toward a sustainable research community for BID.
Table 1.1 summarizes the workshop presenters and the title of their presenta-
tions. The workshop was broken into three sessions where leading researchers in
the disciplines that intersect to form the BID research community presented the
state of the art in bio-inspired research topics.
2
http://designengineeringlab.org/BID-workshop/Workshop_2_%40_DCC.html
3
http://www.vuse.vanderbilt.edu/*dfisher/AISD-Program.html
Table 1.2 Workshop 1 Name Affiliation

participants and affiliation
Tom Arciszewski George Mason University
Shapour Azarm University of Maryland
Janine Benyus The Biomimicry Group
Amaresh Chakrabarti Indian Institute of Science (India)
Paul Egan Carnegie Mellon
Douglas Fisher Vanderbilt
Frank Fish West Chester University
Robert J. Full University of California, Berkeley
Erwin Gianchandani Computing Research Association
Satyandra Gupta University of Maryland
Norbert Hoeller Sustainable Innovation Network
Chris Jenkins Montana State University
Sangbae Kim Massachusetts Institute of Technology
Venkat Krovi University at Buffalo
Julie Linsey Georgia Institute of Technology
Mary Lou Maher University of Maryland
Julia O’Rourke The University of Texas at Austin
David Rosen Georgia Institute of Technology
Megan Schuknecht AskNature.com
Justin Seipel Purdue University
Jami Shah Arizona State University
Jacquelyn Stroble James Madison Univeristy
Srinivasan Arjun Tekalur Michigan State University
Craig Tovey Georgia Institute of Technology
Mohamed B. Trabia University of Nevada, Las Vegas
Irem Tumer Oregon State University
Swaroop Vattam Georgia Institute of Technology
Julian Vincent University of Bath (UK)
Anosh Wadia Texas A&M
Jeannette Yen Georgia Institute of Technology
Following the presentations, three breakout groups were formed to generate

answers (and thus recommendations) for three general questions:
1. If BID is going to contribute to sustainability it needs to…?
2. If BID is going to contribute to complex system design it needs to…?
3. If BID is going to contribute to design education and pedagogy it needs to…?
The three breakout groups were led by Mary Lou Maher, Jami Shah, and Craig
Tovey, respectively. The final session of the workshop featured debrief reports by
each of the breakout groups and a general discussion session for all participants.
The workshop included 30 researchers from the disciplines of engineering design,
computer science, and biology in addition to the three organizers. The participants
are listed in Table 1.2.
The second workshop, held in College Station, Texas, leveraged the compu-
tational design community at the 2012 Design Computing and Cognition
Table 1.3 Presenters and presentation titles at workshop 2

Speaker Title
Rob Stone Review of workshop 1 activities and findings
Oregon State
University
Ashok Goel Grounding bio-inspired design in cognition and computation
Georgia Institute of
Technology
Julie Linsey Word tree express—a tool for design by analogy and bio-inspired
Georgia Institute of design
Technology
Filipo Salustri Analogy and systems are at the heart of BID
Ryerson University
Marc Weissburg Pedagogical challenges to BID innovation
Georgia Institute of
Technology
Rob Stone Gathering designer feedback to generate requirements for intuitive
Oregon State biologically-inspired design tools
University
Conference to evaluate the first workshop’s findings and generate a draft

description of a program for research funding for BID. The objectives for the
second workshop were to
1. Review the findings from the first workshop.
2. Introduce new ideas and research directions for BID through invited talks and a
poster session with participants.
3. Draft a BID program description for a funding agency.
The outcome of the second workshop is captured in the later section entitled
‘‘Proposed NSF Program in BID.’’ Table 1.3 summarizes the workshop presenters
and the title of their presentations. The participants of the second workshop are
listed in Table 1.4.
In addition to the individual presentations, the second workshop held breakout
sessions with the goal of defining a potential NSF research program for funding
BID research. The leaders of the breakout sessions included Alice Agogino, Tom
Arciszewski, David Brown, Julie Linsey, and Marc Weissburg.
1.2 Current Research from the Disciplines
We begin with a brief summary of the state of the art from various subdisciplines
of BID, with pointers to workshop presenters whose presentations elaborate each
of those points.
Table 1.4 Workshop 2 Name Affiliation

participants and affiliation
Alice Agogino U. California at Berkeley
Kinda Al Sayed University College London (UK)
Fernando Alvarez Universidad de Bogotà (Columbia)
Tom Arciszewski George Mason University
Ryan Arlitt Oregon State University
David Brown Worcester Polytechnic Institute
Christopher Earl Open University (UK)
John Gero George Mason University
Michael Helms Georgia Institute of Technology
Julie Linsey Georgia Institute of Technology
Mijeong Kim Kyung He University (South Korea)
Pertti Saariluoma University of Jyväskylä (Finland)
Filipo Salustri Ryerson University (Canada)
Noe Vargas-Hernandez U. Texas—El Paso
Pieter Vermaas TU-Delft (Netherlands)
Marc Weissburg Georgia Institute of Technology
1.2.1 Bio-Inspiration
One of the key aspects of BID is utilizing the similarities noted in nature for a
particular problem to design and to bring inspiration to the designer. Inspiration
through the forms of nature can come in three different types: visual, conceptual,
and computational. Visual inspiration is widely used and understood. Pictures or
other visuals of a biological system are used to create engineering systems that
share the same visual appearance. Conceptual inspiration is the use of the
knowledge found in biology to form design rules, heuristics, principles, or pat-
terns. This type of inspiration requires an understanding in both nature and
engineering. Algorithmic bio-inspiration is searching through nature to find
algorithms like evolutionary computation and knowledge representations such as
generative representations.
There are three different sources of bio-inspiration: evolution which is the
gradual improvement in living systems in response to environmental stimuli,
coevolution or coadaptation of a species in response to evolution of other species
in the habitat, and morphogenesis—evolutionary development of an organism or
its parts (Areiszewski).
Another search for bio-inspiration has been in the development of 3D manu-
facturing. While there are a large number of polymers used in engineering to
complete various tasks, nature uses a small set. Taking inspiration from nature and
using a smaller set more efficiently can be beneficial in manufacturing (Benyus).
Studies have shown that 70 % of engineering problems are solved by energy
rather than by information. On the other hand, nature solves problems by using
information rather that energy, making the system more efficient (Vincent).
Novelty is how unique the solutions are, while variety is how many of the
solutions were new. A study containing a control group, a BID group, and an
engineering group allowed the analysis of the groups’ solutions using these met-
rics. The study showed that BID increases novelty but not necessarily variety.
There was noticeable fixation in the engineering group that was not seen in the
others (Rosen).
One can also use a creativity metric to analyze designs. By examining the
novelty and the usefulness of a design, a creativity score can be determined.
Biological systems may be a way to inspire more creative solutions when
undergoing design problems (Chakrabarti).
1.2.2 Bridging Biology and Engineering
Historically, BID has been rather anecdotal rather than systematic; therefore, a
way to make BID more systematic is needed to increase the benefits gained from
BID. There are two main methods to enhancing BID, stimuli, and transfer
guidelines. Stimuli is broken up into two different categories, structured and
unstructured. Structured information allows for a simpler search and easier transfer
of knowledge. However, the arrangement of the information into a structure can be
difficult and time-consuming. Unstructured information requires no effort to
arrange the information; however, the ‘‘search’’ carried out by the designer and the
transferring of the idea tend to be more difficult and time-consuming. Transfer
guidelines are broken into four general steps: formulate search objectives, search
for biological analogs, analyze biological analogs, and transfer relevant knowledge
to the target domain. Using the SAPPhIRE model, seven levels of abstraction can
be obtained. These abstractions can be used to inspire ideas. The SAPPhIRE model
excels at empirical findings and exploring the number of ideas which lead to a
higher levels of SAPPhIRE and greater novelty. One can also use design creativity
as a metric of the novelty and the usefulness of a design. Currently, not all levels of
SAPPhIRE have been explored and studies are being done to examine these. Use
of the SAPPhIRE model has shown that a systematic framework helps increase the
overall number of ideas (Chakrabarti).
Bio-inspired design is responsible for many useful, innovative designs: The
lotus leaf inspired self-cleaning water repellent surfaces, and the cocklebur
inspired velcro. These are just a few examples of the power of BID. Engineering
design is more problem driven, and the concepts are dominated by knowledge of
similar systems. One can recast BID as a problem-driven effort by combining it
with function-based design methods to create function-based BID. Function can be
used as the analogical connection between what an engineering system needs to do
and how the natural system completes that function. By using normal functional
modeling techniques, naturel systems can be modeled. This modeling framework
allows an analogous connection needed between engineering systems and natural
systems. A practical challenge of this approach is that engineers and biologists use
different terminologies when describing solutions to function; therefore, an engi-
neering-to-biology thesaurus is needed. The current BID methodology uses the
functional basis terms with biological function/flow correspondents. The main
steps for this start with creating a functional model for the design problem using a
terminology known as the functional basis. The next step is translating the func-
tional model into the associated biological keywords. After this is done, one can
search a specific biological knowledge repository or use Google, biology texts, or
other publications (McAdams).
Goel has done work in developing information processing theories and com-
putational methods for BID. This work starts with conducting empirical studies of
BID, then developing general information processing accounts of BID, next
constructing computational tools and techniques for aiding BID, and finally
deploying and evaluating the tools in realistic settings. One of their findings is that
BID often involves compound analogies, entailing intricate interaction between
problem decomposition and analogical reasoning. A second finding is that BID
engages not only in analogical transfer of functionally indexed mechanisms but
also in transfer of problem decompositions. A third finding is that biological
analogies are useful for several tasks of BID in addition to design concept gen-
eration, such as design analysis and explanation.
Other work bridges biological information with engineering information. Rosen
has been researching and developing a strategy–state–structure ontology to create
a formal language that represents designs. Strategy is function plus behaviors;
therefore, this would include a taxonomy of functions and behaviors that could be
used together to form strategies (Rosen).
At the University of Maryland, Gupta has been conducting studies with bio-
inspired robots. Bio-inspired robots are robots with the main inspiration coming
from a bird, animal, or fish. There are many applications of bio-inspired robots that
include but are not limited to medical, reconnaissance, mine detection, enter-
tainment, and space exploration. The traditional approach in creating these robots
is taking inspiration from nature and adding modeling- and simulation-based
optimization to create designs for these robots. Traditional manufacturing tech-
niques are then examined to determine which designs can be executed. There are
many successful robots larger than 100 mm; however, miniature robots (between
100 mm and 5 mm) usually come with limited capabilities. By finding new novel
manufacturing concepts, highly capable yet miniature bio-inspired robots could be
implemented. The first step is to approach design differently—to simplify by only
retaining features of the biological creature that are useful and to identify high-
value characteristics. Another need in the design phase is to amplify the useful
biological characteristics to improve the performance of the robot. The second step
is to approach the modeling and simulation of these robots differently using
metamodel synthesis. The last is to approach assembly differently (Gupta).
1.2.3 Techniques and Tools for Bio-Inspired Conceptual

Design
The development of design tools to help designers use BID effectively is crucial to
being able to teach BID in a classroom and implement BID in the workplace.
Asknature.org is an important tool for BID. Asknature.org takes organizes bio-
literature by function and allows the user to search through. The Web site is public
domain, part search engine and part social network, and helps people using BID
connect with each other. As of now, asknature.org contains 1,300 strategies for
design. These strategies contain links to biologist Web pages or Google Scholar
articles to help the user be able to research the strategy. Support from Autodesk
Inventor has been crucial as it implements BID into a commonly used design tool.
A current redesign of AskNature is underway to follow the path of idea from
nature–transferable idea–possible products–actual products–digital downloads–
Autodesk Inventor. This would allow an easier transfer of flows from BID into
products. Another possible tool would be implementing native ecosystem data in
city planning. By setting new ecological performance standards, architects and city
planners would use the ecosystem data to develop a city that is more sustainable
and functions together (Benyus).
Another tool of importance is the engineering-to-biology thesaurus. This the-
saurus helps engineers include bio-inspiration in the engineering design process by
relating the engineering functions and flows in the functional basis to biological
functions and flows. In this manner, the designer is allowed to use their functional
model to relate key functionality to the way nature also solves it (McAdams).
Another tool for BID is the creation of databases to organize information.
Goel described two knowledge bases: DANE and Biologue. Both DANE and
Biologue use structure–behavior–function models of biological and engineering
systems. DANE, which has already been released (http://dilab.cc.gatech.edu/dane/),
is a functionally organized database of biological solutions. This is useful to
designers because it allows them to look up the function they are trying to solve and
relate that to biological solutions. The more recent Biologue system is a database
that allows the indexing of biology-related documents using the structure–behavior–
function model. Using this, a designer can compare the functions of what they are
designing and also compare the structures and behaviors of their design to biological
systems.
Of note, bio-inspired optimization techniques are also used in design. Genetic,
ant, particle swarm, bee, and firefly algorithms all contain specific uses to help
engineers optimize in design. By studying the ways grouped animals move and
interact, techniques can be developed to mimic those interactions and create better
optimization techniques and algorithms. These techniques can be used to help
overcome challenges often found while trying to optimize designs. Such chal-
lenges include complexity, scalability, and convergence. There is a need to extend
the applicability of genetic algorithms for design optimization with regard to
uncertainty, system product design, and multi-objective genetic algorithms
(MOGA). Pros and cons exist for all of the BID optimization algorithms. The
desirable traits are that the algorithm is population based, can optimize non-convex
discontinuous functions, can handle discrete–continuous/combinatorial design
problems, and obtain global optimums. However, sometimes these optimization
techniques result in local optimums that can be undesirable. Nevertheless, there
are complications with these algorithms that include complexity, scalability, and
convergence (Azarm).
Vincent discussed another important BID tool, Bio-TRIZ. Bio-TRIZ is a
reduced form of TRIZ that relates the TRIZ inventive principles to biological
contradictions. Studying 5,000 examples, the conflict matrix was reduced from 39
conflict elements to 6 elements that appear in both biology and engineering, and a
6 9 6 contradiction matrix that contains all 40 of the inventive principles was
created. These 6 conflict elements are substance, structure, time, space, energy/
field, and information/regulation. Biological solutions were then studied to fill in
inventive principles aspect of the matrix (Vincent).
1.2.4 Education in Bio-Inspired Design
The Biomimicry Institute (TBI) has been working on implementing BID into
earlier education (K1-12) as well as into a professional masters program. An
important part of featuring BID in a classroom setting is combining biologists,
biomimetic scientists, engineers, and designers together. Academic settings tend to
be very different from work settings. Therefore, working to close the difference
between academic and work settings would lead to better transition from BID in
the classroom to BID in the workplace (Benyus).
Since 2005, Georgia Institute of Technology has offered a BID interdisciplinary
course for engineers, biologists, and other scientists. This course features *30
engineering students and *15 biology students that take part in semester-long,
self-defined design projects that include the help of faculty mentors. This has been
created to help promote BID practice and explore how BID aids in developing
products. The main goal of this course is to encourage ways of thinking about and
to explore nature that helps facilitate a designer’s ability to implement biological
strategies in engineered products and sustainable systems. The development results
in a fifteen-week course that includes the principles of biologically inspired design,
biomimetic materials, biologically inspired sensing and movement, system design
and optimization, and green technology. These lectures contain both content and
practice to help the students grasp the material. The final is a presentation by the
students on a novel BID (Yen).
An important aspect of BID is implementing it in a classroom setting so stu-
dents can not only learn and understand it but also implement it in industry. The
University of Maryland has offered bio-inspired robotics undergraduate classes.
These classes offer an excellent opportunity to teach students how to design a
complex and modern mechatronics system. The content of the course includes the
fundamentals of robotics, including kinematics, inverse kinematics, dynamics,

trajectory generation and controls, bio-inspired robots where design concepts,
sensors, and actuators are covered, and ending with robot building tutorials using
mechanical design, servo motors, microcontrollers, and programming. The class is
project based with bio-inspired robot outcome (Gupta).
1.3 Recommendations
Biologically inspired design has been practiced on an ad hoc basis throughout

human history. The breakout sessions at the first workshop were designed to
explore the research needed to move BID from ad hoc to intentional in its
application. The breakout sessions at the second workshop were targeted toward
defining an NSF program for funding BID research.
1.3.1 A Design Theory for Bio-Inspired Design
Bio-inspired design seeks to exploit biology for several different kinds of design
such as sustainable design, creative design, and complex system design. Of course,
these different kinds of design are mutually compatible and consistent: We can
have complex systems that are sustainable, sustainable designs that are creative,
and so on. Nevertheless, the three kinds of design have different emphases and
foci.
The goal of bio-inspired sustainable design is to use biology as an inspiration
for designing technological products that are ecologically sustainable. Although
biological systems are not always optimal, the argument goes that they typically
use only local and abundant resources and are often very efficient in terms of use
of resources such as energy and water. Of course, this does not guarantee that bio-
inspired products will be necessarily sustainable, but it promises that they may be
more sustainable than equivalent products available in the market today. Consider
the following specific cases:
The Biomimicry Institute’s work on BID is driven by the growing need for
sustainable design. An example of sustainable design at TBI is the novel design of
a water bottle with ‘‘ribs’’ on its plastic surface mimicking the ribs on trees and
providing strength to the bottle. This allows the bottle to use less plastic, which
makes it lighter than similar water bottles (Benyus).
Recent work on BioTRIZ indicates that for many functions for which techno-
logical products typically use energy, equivalent biological systems use infor-
mation instead. This suggests that we seek biological sources as inspiration to
design a new generation of technological products that use information in place of
energy to achieve as many functions as possible (Vincent).
Another line of research seeks to identify design patterns that biological sys-
tems use to achieve ecological sustainability. For example, the Namibian beetle
uses an interleaved pattern of the biological effects of hydrophobia and hydro-
phylia for harvesting water from dry air. Bio-inspired designers have now used the
same pattern for fog harvesting. This suggests building a classification of func-
tionally indexed design patterns for sustainable design (Goel).
The goal of bio-inspired complex system design is to use the characteristically
complex interactions found in nature as a guide to engineered systems that are
complex and integrated among their constituent components. Although biologists
welcome complexity, engineers typically are concerned about it and do what they
can to avoid it. Approaching complex system design from a biologist’s perspective
that complexity could be a positive aspect and allow a mechanism for coping with
product failures appears a promising avenue with the following observations:
On the engineering side of complex systems, one has relatively simple units
leading to an emergent phenomenon that includes many interacting parts exhibiting
nonlinear behavior, uncertainty, and multiple scales. The biology side of complex
designs contains heterogeneity, multiple different parts that fit together and provide
different outcomes dependent on initial conditions, and multi-functionality.
Biologists can help shift the paradigm of perceived complexity versus real
complexity. This shift can help show engineers how to manage complexity in their
systems which could lead to innovative products and elegant designs and help
predict the performance of complex systems.
Nature has the ability to self-repair, adapt, add redundancies, accommodate
failure, regenerate and reconfigure parts, as well as others. Engineering systems do
not accept the notion of failure as a positive feature, but small, intentional failures
to avoid a catastrophic failure can be a good thing.
1.3.2 A Funding Program for Bio-Inspired Design Research
As part of the second workshop, the participants reviewed the findings of the first
workshop and produced variants of a proposed call for proposals. For purposes of
context, it was assumed that the umbrella agency for such a program would be the
United States NSF. The variant ideas were aggregated into the following brief
description of an NSF program:
The NSF invites proposals for research in BID. Biologically inspired designs
and processes have much potential to solve urgent and complex challenges faced
by the United States and the planet such as those found in military, urban infra-
structures, climate change, sustainability, and space exploration domains. Pro-
posals must be from suitable multi-disciplinary teams (i.e., members might include
biologists, computer scientists, engineers, or psychologists), addressing small- to
medium-scale designs (such as household products or automotive systems), have
demonstrated educational and computational potential, and be well evaluated.
Suitable research areas include but are not limited to:
• BID system usability, interface design, visualization, and search;

• Identification of the role of biologically inspired tools and design methods for
each stage of the design process, for example, problem framing, conceptual
design, refinement, production (DfM for BIDs), marketing, reuse/recycle;
• Knowledge-base/database building and integration, ontologies (construction,
use, and evaluation), and test beds (computational and physical);
• Evaluation of analogy utility (before, after, and during use) of BIDs, design
methods, and the manufacturability of resulting designs;
• Roles of the scales (both spatial—from micro- to macro-, to systems of sys-
tems—and temporal—to promote desired emergent behavior over time) of
biological knowledge for problem identification, design decomposition, gener-
ation, evaluation, and explanation;
• Impact of BID on communication in multiple disciplinary teams, for example,
novice–expert studies, development of a community of practice and networks,
sociology of disciplinary norms;
• Teaching approaches and curricular development.
The work must be extensible and must be shared in order to promote BID
community building.
1.4 Summary
In summary, it is clear that recent research efforts across the disciplines of engi-
neering design, computer science, and biology have attempted to address the var-
ious problems associated with not only developing biologically inspired designs,
but also teaching students how to develop biologically inspired designs. However, it
is also evident that there is a need for additional work on refining the proposed
methods and tools as well as developing new methods to address current limita-
tions. The recommendation of the workshop organizers is that a new cross-cutting
NSF program in BID be established that seeks to fund transformative research as
described in the program brief above. Such a program is expected to support high-
risk/high-reward research that otherwise has no current home in the NSF.
Suggested Readings
Bio-Inspiration
Arciszewski T, Cornell J (2006) Bioinspiration: learning creative design principles. In intelligent

computing in engineering and architecture. Lect Notes Comput Sci 4200:32–53
Bar-Cohen Y (2006) (ed) Biomimetics–using nature to inspire human innovation. Bioinspir
Biomim 1(1):P1–P12. doi:10.1088/1748-3182/1/1/P01
Bar-Cohen Y (2011) (ed) Biomimetics: nature-based Innovation. CRC Press, Boca Raton
Benyus JM (1997) Biomimicry: innovation inspired by nature, 1st edn. Morrow, NewYork
Bonser R, Vincent J (2007) Technology trajectories, innovation, and the growth of biomimetics.
In: Proceedings of the institution of mechanical engineers, part C: journal of mechanical
engineering science, pp 1177–1180
Chakrabarti A, Shu L (2010) Biologically inspired design. AIEDAM 24:453–454
Helms M, Vattam S, Goel A (2009) Biologically inspired design: process and products. Des Stud
30(5):606–622
Lindemann U, Gramann J (2004) Engineering design using biological principles. Proc Int Des
Conf Des 5:18–21 (Dubrovnik)
Mak T, Shu L (2008) Using descriptions of biological phenomena for idea generation. Res Eng
Des 19:21–28
Shu, L, Ueda K, Chiu I, Cheong H (2011) Biologically inspired design. CIRP Ann Manuf
Technol
Vincent JFV (2005) Deconstructing the design of a biological material. J Theor Biol 236:73–78
Vincent J, Mann D (2002) Systematic technology transfer from biology to engineering,
philosophical transactions of the royal society. Physical Sci 360:159–173
Wilson J, Rosen D, Nelson B, Yen J (2010) The effects of biological examples in idea generation.
Des Stud 31(2):169–186
Yen J, Weissburg M (2007) Perspectives on biologically inspired design: introduction to the
collected contributions. J Bioinspiration Biomimetics 2
Bridging Biology and Engineering
Fish FE (1998) Imaginative solutions by marine organisms for drag reduction. In: Meng JCS (ed)
Proceedings of the international symposium on seawater drag reduction. Newport, Rhode
Island, pp 443–450
Fish FE (2006) Limits of nature and advances of technology in marine systems: What does
biomimetics have to offer to aquatic robots? Appl Bionics Biomech 3:49–60
French M (1994) Invention and evolution: design in nature and engineering, 2nd edn. Cambridge
University Press, Cambridge, 1988
Nagel JKS, Stone RB, McAdams DA (2010) An engineering-to-biology thesaurus for engineering
design. In: ASME IDETC/CIE 2010 DTM-28233, Montreal, Quebec, Canada, 2010
Vattam S, Goel A (2011) Foraging for inspiration: understanding and supporting the information
seeking practices of biologically inspired designers. In Proceedings of ASME DETC
conference on design theory and methods, Washington DC, August 2011
Vogel S (1988) Life’s devices. Princeton University Press, Princeton
Vogel S (1994) Life in moving fluids. Princeton University Press, Princeton
Techniques and Tools for Bio-Inspired Conceptual Design
Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju B (2005) A functional representation for

aiding biomimetic and artificial inspiration of new ideas. AIEDAM 19(2):113–132
Cheong H, Chiu I, Shu LH, Stone RB, McAdams DA (2011) Biologically meaningful keywords
for terms of the functional basis. J Mech Des 132(2):02001-7–02001-11
Chiu I, Shu L (2007a) Biomimetic design through natural language analysis to facilitate cross-
domain information retrieval. AIEDAM 21:45–59
Chiu I, Shu L (2007b) Using language as related stimuli for concept generation. AIEDAM
21:103–121
Goel A, Vattam S, Wiltgen B, Helms M (2012) Cognitive, collaborative, conceptual and
creative—four characteristics of the next generation of knowledge-based CAD systems: a
study in biologically inspired design. Comput Aided Des 44(10):879–900
Lenau T (2009) Biomimetics as a design methodology—possibilities and challenges, interna-

tional conference on engineering design, ICED’09 24–27 August 2009. Stanford University,
Stanford
Nagel JKS, Stone RB (2012) A computational approach to biologically-inspired design.
AIEDAM 26(2)
Nagel RL, Midha PA, Tinsley A, McAdams DA, Stone RB, and Shu LH Exploring the use of
functional models in biomimetic conceptual design. J Mech Des 130(12):121102-1–121102-
13
Sarkar P, Chakrabarti A (2008) The effect of representation of triggers on design outcomes. Artif
Intell Des Anal Manuf 22(02):101–116
Sartori J, Pal U, Chakrabarti A (2010) A methodology for supporting ‘‘transfer’’ in biomimetic
design. AIEDAM 24:483–505
Shu L (2010) A natural-language approach to biomimetic design. AIEDAM 24:507–519
Srinivasan V, Chakrabarti A (2010) An integrated model of designing. ASME JCISE 10(3)
Vattam S, Helms M, Goel A (2008) Compound analogical design: interaction between problem
decomposition and analogical transfer in biologically inspired design. In Proceedings of third
international conference on design computing and cognition, Atlanta, June 2008. Springer,
Berlin, pp 377–396
Education in Bio-Inspired Design
Bruck H, Gershon A, Golden I, Gupta S, Gyger L, Magrab E, Spranklin B (2007) Training

mechanical engineering students to utilize biological inspiration during product development.
Bioinspiration and Biomimetics 2:S198–S209
Glier MW, Tsenn J, Linsey JS, McAdams DA (2011a) Methods for supporting bioinspired
design. In: Proceedings of the ASME 2011 international mechanical engineering congress and
exposition, number IMECE2011-63247, Denver, Colorado
Glier MW, Tsenn J, Linsey JS, McAdams DA (2011b) Concepts in biomimetic design: methods
and tools to incorporate into a biomimetic design course. In: Proceedings of the ASME 2011
international design engineering technical conferences and computers and information in
engineering conference, number DETC2011-48571, Washington, DC
Helms M, Vattam S, Goel K (2010) The effect of functional modeling on understanding
biological systems. In: Proceedings of ASME 2010 international design engineering technical
conferences and computers and information in engineering conference (IDETC/CIE 2010).
August 15–18, 2010, Montreal, Quebec, Canada
Nelson B, Wilson J, Yen J (2009) A study of biologically-inspired design as a context for
enhancing student innovation. In: Proceedings of 39th ASEE/IEEE frontiers in education
conference, 2009
Vakili V, Chiu I, Shu LH, McAdams DA, Stone RB (2007) Including functional models of
biological phenomena as design stimuli. In: Proceedings of the 2007 ASME design
engineering technical conferences and computers and information in engineering conference,
number DETC2007-35776, Las Vegas, Nevada
Weissburg M, Tovey C, Yen J (2010) Enhancing innovation through biologically inspired design.
Adv Nat Sci 3:145–167
Yen J, Weissburg M, Helms M, Goel A (2011) Biologically inspired design: a tool for
interdisciplinary education. In: Bar-Cohen Y (ed) Biomimetics: nature-based innovation,
Taylor & Francis
Yen J, Helms M, Goel A, Vattam S (2010) Evaluating biological systems for their potential in
engineering design. Adv Nat Sci 3(2)
Design Theory for Bio-Inspired Design
Goel A, Bras B, Helms M, Rugaber S, Tovey C Vattam S, Weissburg M, Wiltgen B, Yen J (2011)
Design patterns and cross-domain analogies in biologically inspired sustainable design. In:
Proceedings of AAAI spring Symposium on AI and sustainable design, Stanford University,
Palo Alto, March 2011, pp 45–51
Goel A, Vattam S, Wiltgen B, Helms M (2011b) An information-processing account of creative
analogies in biologically inspired design. In: Proceedings of eighth ACM conference on
creativity and cognition, Atlanta, Nov 2011, pp 71–80
Helms M, Goel A (2012) Analogical problem evolution in biologically inspired design. In:
Proceedings of fifth international conference on design computing and cognition, College
Station, Texas, July 2012, Springer
Nagel JKS, Nagel RL, Stone RB, McAdams DA (2010) Function-based biologically-inspired
concept generation. Artif Intell Eng Des Anal Manuf 24(4):521–553
Vattam S, Helms M, Goel A (2010) A content account of creative analogies in biologically
inspired design. AI Eng Des Anal Manuf Spec Issue Biol Inspired Des 24:467–481
Vincent J, Bogatyreva O, Bogatyrev N, Bowyer A, Pahl A (2006) Biomimetics: its practice and
theory. J R Soc Interface 3:471–482
Chapter 2
The AskNature Database: Enabling
Solutions in Biomimetic Design
Jon-Michael Deldin and Megan Schuknecht
Abstract Practitioners of biomimetic design express one consistent need—access

to relevant biological information. This information is most useful when the
transferable elements have been abstracted from the biological principle, are
organized by design and engineering function, and are supported by contextual
search functions. However, the fine details of how the information is organized and
accessed are critical. In this chapter, we reflect upon AskNature.org, a biomimetic
database created to address these issues, and its key elements, including a Bio-
mimicry Taxonomy, biological strategy pages organized by function, search and
tri-browse features, and bio-inspired product pages as examples of bio-inspired
design successes.
2.1 Introduction
Biomimetic design professionals need access to relevant biological information

expressed in common language (Bar-Cohen 2006). For example, both a biologist
and a civil engineer can talk about ‘‘managing temperature’’ and have a basic
understanding of what the other means, whether one of them is talking about an
elephant seal managing temperature or the other about the need to regulate tem-
perature in a building. However, if the biologist starts the conversation by talking
about the intricate structure of a northern elephant seal’s nasal turbinates, rather
than their function, that information is not going to be of obvious interest or
application to the civil engineer.
J.-M. Deldin M. Schuknecht (&)

Biomimicry 3.8 Institute, Missoula, MT, US
e-mail: megan.schuknecht@biomimicry.net
J.-M. Deldin
e-mail: jon-michael.deldin@biomimicry.net

18 J.-M. Deldin and M. Schuknecht
Table 2.1 AskNature.org Years Page views Unique visitors

website usage from
November 12, 2008, through 2008 125,568 22,386
December 31, 2012 2009 1,066,527 208,661
2010 1,195,928 236,117
2011 1,484,148 295,897
2012 1,790,709 478,486
Source Google analytics
We developed AskNature.org in an effort to translate biological information so

it would be accessible to non-biologists, and to serve as a source of inspiration for
biomimetic design. AskNature is a free, publicly available database of biological
information and abstracts, most of which are drawn from peer-reviewed journals.
The website catalogs biological information by function in ‘‘strategy’’ pages
(described in Sect. 4.1) to bridge the gap between biology and fields like engi-
neering, architecture, industrial design, chemistry, organizational development,
and more. Since its launch at Greenbuild International Conference and Expo in
2008, the site has experienced rapid growth, recording almost 1.8 M page views in
2012 (see Table 2.1). Additionally, AskNature has been lauded by design pro-
fessionals, winning an Earth Award in (2010) and becoming a finalist for an
INDEX Award in 2011 (Biomimicry 3.8 Institute 2011).
Function is one way to interpret biological information in order to create a
bridge of understanding across disciplines, and identifying function can be one key
component to approaching biomimetic design (Helms et al. 2009, 2010; Stone and
Wood 2000; Vattam and Goel 2011). The AskNature team developed the Bio-
mimicry Taxonomy (taxonomy) as a means to organize biological data by function
and present it to a design audience. The taxonomy represents the organizing
schema for all of the biological strategy pages within the database. We will discuss
the taxonomy in greater detail below.
2.2 Data
Trained biologists were responsible for gathering and generating the original
almost 1,300 pages of biological data on AskNature. Researchers read scientific
journals and books and perused scientific news, looking for leads on functional
biology that might be of interest to innovators working to solve human challenges.
Strategies were selected subjectively based on the researchers’ assessment of
whether the strategy held some potential for being useful within the field of bio-
inspired design. AskNature‘s original data set represents a huge amount of human
labor and was only possible due to generous funding from an independent investor.
Individuals continue to generate additional content for AskNature. Most of the
biological data that has been added since 2008 has come from paid staff.
2 The AskNature Database: Enabling Solutions in Biomimetic Design 19
Qualified scientific curators are also able to add biological content. However, due
to the rigors of maintaining the scientific integrity of the data and the need for strict
adherence to the data’s organizational structure, all curated content must be
approved by AskNature‘s content editor.
2.3 The Biomimicry Taxonomy: Organizing Biology

by Function
Once the strategy data were collected, the AskNature team began to look for
patterns and to organize the data according to function. The result was a classi-
fication system we call the Biomimicry Taxonomy, shown in Fig. 2.1. The tax-
onomy categorizes strategies according to three levels: groups (highest level),
subgroups, and function. Overall, the taxonomy includes 8 groups, 30 subgroups,
and 162 functions. Individual strategies represent the next, most detailed level
within the hierarchy.
As a specific example, an insect might face the challenge of protecting itself
from other organisms that want to eat it. Its strategy to meet that challenge might
appear like this within the taxonomy:
Group Maintain physical integrity
Subgroup Protect from biotic factors
Function Protect from animals
Strategy Nanoscale protrusions (AskNature 2008a).
2.3.1 Compared to Other Taxonomies
AskNature staff consulted with external design professionals when creating the
taxonomy, particularly from chemical and materials science disciplines, but this
schema is subjective and represents a small group of biologists attempting to
organize a huge amount of biological data by function. Unlike other schemas that
have been developed to assist bio-inspired design (Glier et al. 2011; Vattam et al.
2010; Vincent et al. 2006; Yen et al. 2011), the taxonomy has not been rigorously
tested to assess its impact on its users.
2.4 A Tour of AskNature.org
AskNature consists of a number of components: biological strategy pages, bio-

mimetic product case studies, a search engine, and a social network.
Fig. 2.1 The biomimicry taxonomy

2.4.1 Strategy Pages
All the biological content on AskNature is captured on strategy pages. As of

January 2013, the AskNature database contains over 1,600 biological strategy
pages. Strategies are solutions a given organism or ecosystem uses to meet
functional challenges.
Each strategy page consists of a carefully crafted title, a short sentence
explaining the essence of the biological strategy, the strategy’s place within the
Biomimicry Taxonomy, a scientific excerpt, biomimetic application ideas and/or
links to biomimetic product pages, and links to scientific references via Google
Scholar or Scirus. In addition, most of the strategy pages include photographs and/
or illustrations of how the strategy works, basic natural history information
(including IUCN Red List (International Union for Conservation of Nature and
Natural Resources 2012) status number, if applicable), links to videos that provide
further context on a given strategy, and links to scientific experts and/or
laboratories.
For example, the title ‘‘Wing scales diffract and scatter light: Morpho butter-
flies,’’ on the strategy page of the same name (AskNature 2008c), tells a user at a
glance what function is being accomplished and by what organism (Fig. 2.2). The
sentence below the illustration of the strategy provides more detail: ‘‘Wings of
Morpho butterflies create color by diffracting and scattering light.’’ On every
strategy page, the structure of this sentence is the same: part of the organism,
organism, what it does, and how it does it. This sentence represents one level of
abstraction regarding the biology of the strategy; that is, it represents one inter-
pretation of function at a given scale. Depending on the content, the scale dis-
cussed may vary.
The level of abstraction depends on a combination of the perspective of the
original biological researcher, the interpretation by AskNature staff, the level of
scientific detail known, and the framework of the database. In this particular case,
much more detail is known about the architecture of the wings and how morpho
butterflies refract and scatter light, that is, the mechanism used to accomplish the
function, but that information and level of abstraction do not fit neatly into the
summary sentence. A user must read further to find out, for example, that ‘‘nano-
sized, transparent, chitin-and-air layered structures that, rather than statically
absorb and reflect certain light wavelengths as pigments and dyes do, selectively
cancel out certain colors through wavelength interference while reflecting others,
depending on the exact structure and interspatial distance between diffracting
layers’’ or to learn any exact dimensions used in this strategy.
While AskNature‘s strategies allow users to see a wide cross section of the
ways organisms have met functional challenges, they generally does not provide a
great level of detail about the mechanism of function. For example, what specif-
ically about the nasal turbinates of northern elephant seals makes counter-current
heat exchange so effective? However, the sheer number of strategy pages, the
Fig. 2.2 Screenshot of the top of AskNature‘s morpho butterfly strategy page (AskNature 2008c)
organizational structure, and the fact that the database is free to the public have
made AskNature a much more widely used tool than others that are accessible only
to select users.
2.5 Biomimetic Products
AskNature features nearly 200 biomimetic product pages for users to see how
others have applied a certain strategy. A product’s page describes the history of the
product, what challenges it solves, how it is different from existing products, and
its bio-inspired principles. For example, Fig. 2.3 shows users how one design
group was inspired by the whitebark pine’s spiral fibers (AskNature 2008b) to
develop a distinctive, 100 % recyclable polyethylene terephthalate (PET) bottle.
As the site grows, we hope to add even more biomimetic products to our database.
Fig. 2.3 Screenshot of an AskNature product page (AskNature 2011) showing the product
details, history, and more
2.6 Searching AskNature
Searching is one of the most vital features on AskNature. Without it, most users
would never discover the wealth of biological strategies available on the site. We
provide two mechanisms for finding content: a traditional search and a taxonomy
browser.
As shown in Table 2.2, visitors use our search engine more than the browsing
option. Every page on AskNature includes a full-text search box with the prompt
‘‘How would Nature….’’ The prompt is meant to encourage users to search by
function; that is, by inserting a verb: How would nature cool? How would nature
create color? How would nature distribute fluids? How would nature build com-
munity? The search function defaults to return strategy pages before any other type
of page within the system, such as user profile pages.
Table 2.2 Search and tri- Page views

browse usage from November
12, 2008, through December Years Browse Search
31, 2012 AskNature‘s search 2008 5,806 10,332
engine is used more often 2009 45,286 130,731
than the tri-browse feature 2010 44,241 173,183
2011 61,602 271,777
2012 62,106 303,861
Source Google analytics
There are numerous possible queries, but they must be plausible and recog-
nizable by the database. For example, if a user types in ‘‘build an airplane,’’ the
search may not return any useful results. After all, Nature does not build airplanes,
but it is a genius at flight. If the user refines his search to look for things such as
‘‘generate lift’’ or ‘‘reduce drag’’—that is, functions he would like his design to
accomplish—chances are he will find much more relevant results.
While finding strategies via search can provide immediate results, it may be
useful for practitioners to view multiple strategies meeting similar functional
challenges. To accommodate this, users can browse strategies and products by
function in our ‘‘tri-browse’’ page shown in Fig. 2.4. The tri-browse page enables
users to explore different strategies solving similar functional challenges, but users
may find inspiration by browsing related subgroups as well.
2.6.1 Query Analysis
We have collected aggregate search queries since AskNature launched in 2008.

Table 2.3 presents the top queries, and Table 2.4 presents a random selection.
These tables reveal a number of observations. First, our suggested queries on each
search result page attract a number of hits, indicating users are following them as a
way of browsing. Second, users are searching without the Biomimicry Taxonomy
and are instead using it as a generic search for organisms, environments, and other
buildings. Third, literal questions are used as queries, which indicates inexperi-
enced Web searchers (we ignore ‘‘how,’’ ‘‘what,’’ ‘‘the,’’ and other common
function words to mitigate this).
In summary, visitors are using our search engine like a traditional keyword Web
search engine and not searching using the Biomimicry Taxonomy. Further
research is needed to determine whether recasting a user’s query in terms of the
Biomimicry Taxonomy would increase user success rates.
Fig. 2.4 Screenshot of the biomimicry taxonomy browser using group = modify, sub-
group = attach, and function = permanently. Matching strategies appear in the rightmost
column
Table 2.3 Top 15 search Hits Query

queries based on aggregate
search logs from September 24964 Adhere to water
14, 2008, through February 19359 How to purify water
20, 2012 18933 Water
18242 Structural color for painting
18201 Sticky berries adhere with strength
and ease
18124 Capture water from fog in an arid environment
15046 Negative pressure used to suck moisture
from soil
15035 Capture water from fogs
14096 Adhere to water
11896 Mutualism in nature
11719 Capture water from fog
11592 Hovering in mid-air
11233 Insects that capture water from fog
11177 How would nature purify water
10634 Structural color
The queries with a leading capital letter are suggested queries on
each search result page
Table 2.4 Randomly Hits Query

selected queries from the
same period as Table 2.3 6172 A leaf capture water
5680 Green
4466 Plants that filter air
3267 How would nature resolve conflicts
2959 Solve green building challenges
2939 Green building
1698 Cool
1547 Improve air flow
1356 Organize an economy
1139 Build buildings
1207 Waterproof
1170 How to purify water, beavers
921 Solar
893 Glue
818 Desert
2.6.2 Social Network
AskNature is the execution of Benyus’ vision (1997, p. 291) of a place where

engineers and biologists can collaborate and share Nature’s solutions to engi-
neering problems. To support this, AskNature enables users to create profiles,
comment on strategies and products, and discuss topics in forums. Users can
connect with others in specific disciplines and countries also via the tri-browse
page. We have not observed the level of collaboration desired, so we need to
conduct a few user studies to determine how AskNature can be more conducive to
collaboration.
2.7 Conclusion
In this paper, we have described AskNature.org and how we categorize its data
according to the Biomimicry Taxonomy. We have learned four preliminary les-
sons from running a bio-inspiration database. First, if one is practicing biomimetic
design, one needs tools to support it. AskNature is one such tool. Second, if a tool
is going to support biomimetic design, it needs to provide a sizable corpus for
inspiration. AskNature provides over 1,600 strategies. Third, it is important to
provide a free and publicly available tool for anyone to use. AskNature‘s content is
free and licensed under a Creative Commons Attribution-Noncommercial 3.0
License. Finally, AskNature is under development. It is an experiment, and we
look forward to adapting AskNature to our user’s needs.
Acknowledgements We thank Ashok Goel and our anonymous reviewers for their invaluable
feedback.
References
AskNature (2008a) Eyes are anti-reflective: elephant hawk-moth. http://www.asknature.org/

strategy/b389c40f5c8660fc082282bac3e00c3f
AskNature (2008b) Spiral fibers strengthen tree trunks: whitebark pine. http://www.
asknature.org/strategy/e70b6b7753773a7177fe8358ec26ed2c
AskNature (2008c) Wing scales diffract and scatter light: Morpho butterflies. http://www.
asknature.org/strategy/1d00d97a206855365c038d57832ebafa
AskNature (2011) Vitalis pet bottle. http://www.asknature.org/product/b5dc4030c48e41b03442d
34aecdf46b4
Bar-Cohen Y (2006) Biomimetics–using nature to inspire human innovation. Bioinspir Biomim
1(1):P1–P12. doi:10.1088/1748-3182/1/1/P01
Benyus JM (1997) Biomimicry: innovation inspired by nature, 1st edn. Morrow, New York
Biomimicry 3.8 Institute (2011) Asknature selected as finalist for prestigious design competition.
http://www.asknature.org/article/view/index2011
Glier MW, Tsenn J, Linsey JS, McAdams DA (2011) Methods for supporting bioinspired design.
ASME Conf Proc 2011(54884):737–744. doi:10.1115/IMECE2011-63247
30(5):606–622
Helms M, Vattam S, Goel A (2010) The effect of functional modeling on understanding complex
biological systems. ASME Conf Proc 2010(44137):107–115. doi:10.1115/DETC2010-28939
International Union for Conservation of Nature and Natural Resources (2012) The IUCN red list
of threatened species. http://www.iucnredlist.org/
Stone RB, Wood KL (2000) Development of a functional basis for design. J Mech Des
122(4):359–370. doi:10.1115/1.1289637
The Earth Awards (2010) Systems winner. http://www.theearthawards.org/site/systems-win/
Vattam SS, Goel AK (2011) Foraging for inspiration: understanding and supporting the online
information seeking practices of biologically inspired designers. ASME Conf Proc
2011(54860):177–186. doi:10.1115/DETC2011-48238
Vattam SS, Helms ME, Goel AK (2010) A content account of creative analogies in biologically
inspired design. Artif Intell Eng Des Anal Manuf 24:467–481. doi:10.1017/S08900604
1000034X
Vincent JFV, Bogatyreva OA, Bogatyrev NR, Bowyer A, Pahl AK (2006) Biomimetics: its
practice and theory. J R Soc Interface 3(9):471–482. doi:10.1098/rsif.2006.0127
interdisciplinary education. In: Bar-Cohen Y (ed) To appear in biomimetics: nature-based
innovation. Newyork, Taylor & Francis
Chapter 3
A Natural Language Approach
to Biomimetic Design
L. H. Shu and Hyunmin Cheong
Abstract Identifying relevant analogies from biology is a significant challenge

for both designers who are interested in applying biological analogies for design
and researchers who are developing general methods to support biomimetic
design. This chapter discusses how a natural language approach facilitates the
identification of biological analogies. We review methods developed to locate
useful biological knowledge in natural language format, for example books and
papers, as well as apply the identified knowledge to design solutions. We then
discuss how these methods can bridge the gap between the increasing amounts of
available information in biology and the goal of capturing computational design
knowledge for biomimetic design. Finally, application examples of the natural
language approach demonstrate that this approach can identify nonobvious anal-
ogies that are based on the transfer of abstract strategies, in addition to the direct
mimicries that are used in most other biologically inspired design solutions.
Keywords Biomimetic design methods

Biomimetic design applications

Natural language processing Computational linguistics Information retrieval

Information extraction Search keywords Search categorization Analogical

transfer Analogical mapping tools Protocol analysis
L. H. Shu (&) H. Cheong

Department of Mechanical and Industrial Engineering, University of Toronto,
Toronto, Canada
e-mail: shu@mie.utoronto.ca
H. Cheong
e-mail: cheong@mie.utoronto.ca

30 L. H. Shu and H. Cheong
3.1 Introduction
Even as increasing examples of biomimetic or biologically inspired designs are

developed and reported, we may have only scratched the surface in identifying and
applying potential biological analogies. Designers who are interested in applying
biological analogies are often limited by their own or others’ bias, knowledge, or
chance observation of relevant biological phenomena. However, much potentially
useful biological information is already widely available but remains underuti-
lized. Increasingly, books, papers, and other natural language resources that
describe biological phenomena are becoming accessible in digital/online format.
To enable a more objective and systematic approach to identifying possible
analogies from the vast amount of such biological information, the lead author has
chosen and pursued the natural language approach to biomimetic design since
2001. Objectives include developing strategies and techniques to first identify
relevant biological analogies from natural language text, and second support
designers in applying these analogies to develop design concepts. Shu (2010) and
Shu et al. (2011) summarized the group’s research work over the previous decade.
Other approaches in biomimetic design include modeling for, and compiling
databases of, biological information. Such approaches may be particularly effec-
tive to aid in transferring knowledge from a chosen biological system to an
engineering solution. However, a challenge that remains is identifying potentially
useful biological knowledge to be modeled and entered into such databases.
This chapter first summarizes various approaches in biomimetic design to place
our work in context. We then present the methods developed for the natural
language approach, how computational techniques can help improve those
methods, and the potential of the approach to support other biomimetic design
research. Application examples are then presented, which demonstrate that the
natural language approach is able to identify nonobvious analogies based on
working strategy rather than superficial similarity.
3.2 Approaches to Support Biomimetic Design
Benyus (1997), a biological sciences writer who popularized the notion that
humans emulate biological phenomena to design sustainable products and pro-
cesses, founded the Biomimicry Institute, a clearinghouse for biomimicry
researchers. A resource supported by the Biomimicry Institute is an online data-
base of biological solutions, available at AskNature.org.
Vincent, a biologist working in engineering, is developing TRIZ (the Russian
system for creative solution of problems) as the main tool of biomimetics. Vincent
and Mann (2002) describe how TRIZ is adapted to support the transfer of knowledge
from the biological to the engineering domain in a method they called ‘BioTRIZ.’
3 A Natural Language Approach to Biomimetic Design 31
In TRIZ, problems are first expressed in terms of contradictions or conflicts. Next,

‘inventive principles’ that typically solve those conflicts are identified and applied.
TRIZ was developed by analyzing over two million Russian patent certificates.
Vincent et al. (2006) created BioTRIZ using the same inductive approach as for
TRIZ, but based on over 500 biological phenomena instead of patent knowledge.
They found that the contradiction chart for BioTRIZ is only 12 % similar to the
original TRIZ, suggesting that different principles are used to solve contradictions in
biology. Notably, biological solutions rely less on energy, and therefore, the use of
biologically inspired design is promising for the development of new and sustain-
able engineering solutions.
While Chakrabarti and Shu (2010) categorized biomimetic research as
descriptive versus prescriptive, here we note that research on methods that support
biomimetic design in general falls under the two high-level categories:
1. Methods to support search, retrieval, and representation of biological phe-
nomena for design
2. Studies to better understand and therefore support the application of biological
analogies to design
Clearly, the two categories are related and how the first category is imple-
mented directly affects the second. Research under the above categories will be
addressed in the order in which they are relevant during the biomimetic design
process.
Helms et al. (2009) identify two directions in biomimetic design:
1. Solution driven, where an interesting biological phenomenon inspires the
search for potential applications and
2. Problem driven, where a given problem motivates the search for biological
analogies that could help solve the problem.
While a number of biologists and design researchers (Benyus 1997; Vincent
and Mann 2002; Bar-Cohen 2006) have suggested potentially useful biological
analogies for engineering applications, few methods have been developed to
support systematic application of the solution-driven approach. The second
direction, the problem-driven approach is more widely practiced, promoted, as
well as studied. Since much of the general methods on biomimetic design support
this approach, this section is devoted to the problem-driven approach.
One recurring message of this chapter is that there are several levels of bio-
logical organization, from the molecular, for example DNA, to the ecosystem/
biosphere that can be exploited for biological analogies. Biological analogies are
often recalled from the organ (e.g., lung, leaf) to organism (e.g., animal, plant)
levels, evidenced by the many instances of existing biomimetic design based on
organs and organisms. However, biomimetic design is more fully exploited by
identifying analogies beyond the obvious ones that come to mind, which motivates
more objective search approaches.
3.2.1 Problem Definition
The problem-driven approach requires that a problem be defined sufficiently to

enable a meaningful search for biological analogies. The majority of work in
biomimetic design suggests that functions are identified in this step. Functional
modeling in engineering typically involves expressing the desired function to be
fulfilled by a design solution. Complex high-level functions are decomposed into
simpler functions, any one, or combination of which can be addressed using
biological analogies. A primary goal in functional description is to avoid limitation
to specific physical solutions for as long as possible.
Functions are typically represented by verbs. The use of nouns, except as
objects of functional verbs, is usually discouraged as this may indicate that a
physical solution is already in mind, and does not give other potential solutions fair
consideration. However, the use of adjectives also has potential, as adjectives may
describe desired qualities of the solution, for example flexible or transparent.
Although to some extent, the use of adjectives also assumes certain physical
entities, for example surfaces that are flexible or transparent, versus the need to
allow for deflection or transmit light.
Other approaches may require redefining the original problem. In TRIZ and
BioTRIZ, the problem is first reformulated as a conflict. Vincent et al. (2006) give
an example where tire chains are needed on icy roads during winter to increase
friction for safe driving. However, chains will damage ice-free road surfaces, and
conventional alternatives such as changing between winter and summer tires are
inconvenient. Therefore, a tire that can instantly adapt to road conditions is
desired. The conflict becomes, ‘How can friction between the tire and road
increase, without increasing the weight (normal force) of the vehicle?’
Helms et al. (2009) and others suggest as a step between problem definition and
solution search, to reframe or ‘biologize’ the problem, that is, by redefining
problems in biological terms, often in the form of a question, for example, ‘How
do biological systems accomplish the desired function?’ The example given re-
frames or biologizes the function of ‘stopping a bullet’ into considering how
biological entities ‘prevent, withstand and heal damage.’ However, this step must
be undertaken with some caution, as it may predispose the designer to specific
analogies or biological phenomena, and levels of biological organization.
Problem definition in biomimetic design is often iterative. Vattam et al. (2008)
describe a complex interplay between problem decomposition and analogy
retrieval and termed this process ‘compound analogical design.’ The authors note
that multiple biological analogies could be used together to form a single design
solution.
3.2.2 Search for Biological Analogies
There are a number of ways to search for biological phenomena that are relevant as
possible analogies to a given problem, each with associated benefits and chal-
lenges. While the final goal is to identify an analogy that provides a working
solution, the initial goal ought to be identifying a large variety of potential bio-
logical analogies in an objective manner.
Lindemann and Gramann (2004) propose a checklist of associations to translate
between technical functions and biology terms. For example, the technical func-
tions of ‘change of the state of aggregation’ and ‘condense a gas’ are associated
with biological terms: nose passages, desert plants/animals, and leaves.
Helms et al. (2009) suggest several solution search heuristics including change
constraints, identify champion adapters, examine variations within a solution
family, and identify biological systems with multi-functionality.
With or without incorporating the above, below are approaches for finding
biological analogies.
3.2.2.1 Ask Biologists Directly
The most obvious way of identifying potential biological analogies is to consult

with biologists. AskNature (www.asknature.org) provides a social network of
biologists specifically for this purpose. The advantage of simply asking biologists
is that the engineer, generally untrained in biology, does not have to search for and
interpret the relevance and potential application of information that may be
unfamiliar. The disadvantage of this approach is that one must have access to
biologists. In addition, biologists, as humans, may be biased toward their areas of
expertise, rather than objectively recall and present a variety of phenomena as
potential analogies.
3.2.2.2 Search Through Database
An obvious approach to address the limitations of asking biologists directly is to

attempt to capture their knowledge in a database. AskNature (www.asknature.org)
supports a database of potential biological analogies compiled by a network of
biologists. One advantage of the database approach is that more focused search
results are often possible. Frequently, the same keywords used to categorize
biological phenomena and often, past engineering solutions developed based on
them, are presented as search keywords. Therefore, the ‘relevance’ of information
found is guaranteed since the same keywords used to categorize and enter infor-
mation are used to then search the database.
Several engineering researchers propose modeling approaches to support the
compilation of databases of biological knowledge for biomimetic design.
Chakrabarti et al. (2005) developed a model to represent causality of natural and

artificial systems and used it to structure information in a database of systems from
both natural and artificial domains. Bruck et al. (2007) created a design repository
of bioinspired products and concepts to support bioinspired robotic projects for
senior mechanical engineering students. Wilson et al. (2009) propose the use of
reverse engineering and ontology to structure a database to support biomimetic
design. Goel et al. (2011) constructed a functionally indexed knowledge base of
structures and behaviors of biological and technological systems. Nagel et al.
(2010) used functional modeling and a design repository to connect biological and
engineering solutions.
The main disadvantage of searching through a database specifically developed
to support biomimetic design is that the search results are limited to what was
entered into the database. Depending on how the database is structured, bias may
be imparted during the categorization of information as it is entered. A simple
example is that while Velcro was developed from burrs, should the biological
entity of a burr be categorized under the engineering function of fastening? If so,
potential functions or strategies other than fastening that can be extracted from the
burr may be lost.
3.2.2.3 Natural Language Based Search
In addition to building a database of biological phenomena categorized for engi-

neering use, another approach takes advantage of the abundant biological infor-
mation already available in natural language format (e.g., texts, papers) by
searching them directly for relevant phenomena. This approach also avoids the
enormous and likely subjective task of cataloging all biological phenomena for
engineering. The natural language approach is supported by Vandevenne et al.
(2011), who propose a scalable approach for the integration of large knowledge
repositories in the biologically inspired design process. While this approach
overcomes the many limitations of other approaches, there are also complex
challenges, as detailed below.
3.3 Methods to Support the Natural Language Approach
This section summarizes methods developed to support the natural language

approach. We also describe how tools from computational linguistics could be
applied to these methods.
3.3.1 Source of Biological Information
The research group at the University of Toronto chose Life, the Science of Biology
by Purves et al. (2001), a reference text for an introductory-level university
biology course, as the initial source of natural language information in 2001. The
text is written at a level that can be easily understood by engineers with little or no
background in biology. In addition to ease of comprehension, introductory-level
texts tend to be general and cover a wide range of organizational levels, from the
molecular and cellular to the ecosystem, such that potential solutions are not
limited to one or two familiar levels. The authors of the textbook kindly provided a
digital searchable version to the University of Toronto to support this work.
Although many other texts are available, this text was particularly well suited to
this purpose.
Our group developed a search tool that identifies matches to keywords in the
text. Other texts can be substituted or added as required for the initial or sub-
sequent search. The challenging task is the initial identification of relevant bio-
logical phenomena. Once relevant phenomena are identified, further details can be
found through more advanced texts, research papers, and traditional research
methods. However, searching through more advanced sources initially will gen-
erate results that are in more technical language and thus more difficult to
understand. Due to the designer frustration this may cause, such results are more
likely to be overlooked even if they are relevant. Therefore, initial results from a
more basic text may be more effective for introducing the subject and confirming
relevance to the design problem. An understanding and confirmed relevance of the
subject then motivates further research and effort to understand the possibly
complex details needed to develop a solution.
3.3.2 Search Keywords
As the same keywords used to categorize phenomena are often presented as

possible search keywords, searching databases specifically created to support
biomimetic design tends to be more straightforward than searching natural lan-
guage sources. Unlike with databases, finding relevant analogies in natural lan-
guage text is subject to the word choices made by authors of the various natural
language sources. That is, multiple terms can describe and thus be required to
locate the same concept.
Vakili and Shu (2001) noted that synonyms are an obvious means to increase
the number of matches for a given functional keyword. Chiu and Shu (2004) noted
that troponyms, verbs that describe specific manners of another verb, for example,
‘ambling’ is a troponym of ‘walking’ as it is a particular manner of walking, also
represent suitable alternative keywords.
Keywords used to search for biological analogies are verbs that describe the
intended effect, or function, of desired solutions. Consistent with functional
description, verbs are strongly preferred over nouns as keywords to initiate sear-
ches. Searching with nouns suggests preconceived analogies or solutions while
searching for verbs that describe the desired action is more likely to objectively
identify a wider range of biological forms that perform that action. For example,
searching with the verb keyword, ‘protect’ will identify several biological phe-
nomena that involve protection. However, searching using the noun keyword,
‘shell’ will only identify information related to shells, thereby limiting potential
protection solutions to those that are based on shells.
WordNet (Miller 1995), an online lexical database, is used to identify syn-
onyms or troponyms for originally formulated keywords. While WordNet uses
semantic information to categorize its lexicon, VerbNet (Karin et al. 2002) uses
both semantic and syntactic information to categorize verbs. Therefore, VerbNet
can identify not only whether a particular verb has a similar meaning as another
verb, but also whether both verbs can be used interchangeably in English without
syntactic error. Therefore, VerbNet may be used to create more coherent groups of
keywords.
3.3.2.1 Biologically Meaningful Keywords
Different lexicons, or vocabularies, used between engineering and biology to

describe the same phenomena, motivated the identification of what Chiu and Shu
(2005) refer to as biologically meaningful keywords. The motivating example
involved the keyword ‘clean,’ which did not locate many useful matches for a
problem involving cleaning. The advice of a domain expert was sought, who
suggested ‘defend’ as an alternative keyword, since some organisms clean as a
defensive mechanism. However, ‘defend’ is not intuitively related to ‘clean’ for
most engineers and has no documented lexical relationship with ‘clean,’ for
example, as a synonym. Therefore, Chiu and Shu (2005, 2007) developed a
bridging method that identifies such highly relevant, but nonobvious, biologically
meaningful keywords objectively and in a repeatable manner, without the need for
domain experts.
The method uses word collocation and frequency analysis and is summarized in
Fig. 3.1 for the ‘clean/defend’ example. First, hypernyms of ‘clean’ are found
using WordNet. A hypernym is a more general form of a word, for example,
‘remove’ is a hypernym of ‘clean.’ WordNet is also used to identify related words,
such as hyponyms/troponyms, of ‘remove,’ for example ‘eliminate’ and ‘kill.’
‘Eliminate’ is a hyponym/troponym of ‘remove,’ as eliminating is a specific
manner of removing. The original and additional keywords are then searched in
the corpus, in this case, the Purves et al. (2001) text. Resulting text excerpts that
contain these words are assessed for relevance using metrics that are described
below.
Fig. 3.1 Flow chart of the keyword bridging method, adapted from Chiu and Shu (2007),
illustrates finding the biologically meaningful keyword ‘defend’ for the engineering keyword
‘clean’
Within the relevant matches, words, typically nouns, for example ‘cells,’
‘plants,’ and ‘disease,’ which frequently occur, are then identified, along with the
verbs that act on them. These verbs are then searched for in a biology dictionary.
Verbs that are defined or part of a defined term are categorized as ‘biologically
significant.’ For example, ‘abscise,’ meaning to naturally separate or fall off, such
as by a dead leaf or ripe fruit, is part of the defined term ‘abscission,’ and therefore a
biologically significant verb. Verbs that appear in the definition of terms, but are not
defined themselves, are categorized as ‘biologically connotative.’ An example of a
biologically connotative word is ‘defend,’ since it occurs in several definitions, but
is not a defined term itself. Biologically significant and biologically connotative
words combined are termed ‘biologically meaningful.’ Biologically meaningful
keywords are then sorted by frequency of occurrence in a biology dictionary. Most
useful bridge words are observed to occur between certain frequency cutoffs.
This method was able to identify ‘defend’ and other biologically meaningful
keywords for ‘clean.’ Chiu and Shu (2007) used the method to also identify
‘survive’ as a biologically meaningful keyword for ‘encapsulate’ and ‘break’ or
‘breakdown’ as a biologically meaningful keyword for ‘release.’ A more obvious
and straightforward approach to identify biologically meaningful keywords was to
simply collect all verbs in the vicinity of the search verb. However, this method
yielded far fewer useful, biologically meaningful words.
Chiu and Shu’s (2007) method involved a number of manual tasks, making it
difficult to scale to larger amounts of text. Computational linguistic tools can be
applied to automatize some of the processes. For example, when identifying fre-
quent nouns, a part-of-speech tagger and stop list could be used to highlight more
meaningful nouns. Also, finding the bridge verbs that modify frequent nouns can
be automatically performed using the typed-dependency parser (de Marneffe et al.
2006), which can identify grammatical relationships between two words in a
sentence, for example, between a verb and its object.
3.3.2.2 Biologically Meaningful Keywords for Verbs of the Functional

Basis
Cheong et al. (2008, 2011) used the bridging method to identify biologically
meaningful keywords for functional terms of the functional basis. The functional
basis contains verb–object pairs intended to comprehensively represent the func-
tionality of mechanical devices (Stone and Wood 2000). During the translation
from functional to biologically meaningful keywords, Cheong et al. (2011) iden-
tified four categories where most biologically meaningful keywords appear. The
four cases were used as criteria for identifying biologically meaningful keywords.
Excerpts by Purves et al. (2001) illustrate these cases.
1. Synonymous pair: Words are used synonymously, often in the same sentence
and adjacent to each other, for example, ‘This information is received and
converted, or transduced, by sensory cells into electric signals….’
2. Implicitly synonymous pair: Words are used synonymously, however, in
separate clauses or sentences, for example, ‘The xylem of tracheophytes con-
ducts water from roots to aboveground plant parts. It contains conducting cells
called tracheary elements, which undergo programmed cell death before they
assume their function of transporting water and dissolved minerals.’
3. Biologically specific form: Biologically meaningful keywords describe a
particular manner of the desired function, for example, ‘Mutations of one of the
homeotic genes, bithorax, transform the third thoracic segment into a second
copy of the second thoracic segment.’
4. Causally related pair (Causal relation): Biologically meaningful keyword is
used to describe the enabling function for the desired function, for example,
‘Humans absorb amino acids by breaking down proteins from food.’
A particularly interesting relation is the causally related pair, where the causal
relation is a higher-order relation between the enabling and desired functions that
may reveal a useful strategy for design-by-analogy. Cheong and Shu (2012)
developed an algorithm, detailed in a later section, to automatically extract
causally related functions from biological text. In addition, we believe that other
semantic relations such as synonymous pairs or biologically specific forms could
also be automatically extracted. For example, Hearst (1998) used lexico-syntactic
patterns to automatically discover hyponym/hypernym relationships between
nouns in a sentence. Perhaps, a similar method could be developed to identify
synonyms/troponyms of verbs in a sentence, which could be applied to automat-
ically identify synonymous pairs or biologically specific forms.
3.3.2.3 Adjectives as Keywords
While verb keywords can locate biological analogies that are functionally related
to desired solutions, adjectives can be used to locate biological analogies that share
similar qualities or working environments with the desired solution. For example,
one can use the keyword ‘dry’ to find solutions that must work well in dry
environments, or the keyword ‘wet’ to search for biological systems that enable
moistness. Ke et al. (2010) used adjectives such as ‘wet,’ ‘humid,’ ‘dry,’ to search
for biological analogies to inspire solutions for novel fuel cell bipolar plate con-
cepts, where proper humidity is crucial to proper operation. Ke et al. (2010)
explored searching for matches where the keyword is used as an adjective. Cheong
and Shu (2012) implemented categorizing adjective search matches by the nouns
modified by the adjectives.
3.3.3 Selection and Categorization of Search Matches
The biologically meaningful keywords described above often lead to a large

number of matches. Not all matches are always useful and designers may find the
large number of matches unmanageable. Therefore, techniques and strategies were
devised to further identify and categorize relevant matches.
3.3.3.1 Abstract/Physical Phenomena
Hacco and Shu (2002) observed that keywords used in conjunction with non-
physical phenomena contribute to many irrelevant matches. For example, in the
following sentence, the verb ‘support’ acts on the abstract entity ‘theory’:
Scientists have now found morphological evidence to support the theory. (Purves et al.
2001).
A designer wishing to find biological phenomena related to ‘supporting’

physical entities, may not find the above search result useful.
Ke et al. (2010) used the WordNet taxonomy to determine whether a particular
noun is abstract or physical. At the highest level of the taxonomy, all nouns are
classified as either an abstract entity or a physical entity. This information can be
tagged on all nouns of the search corpus in advance to distinguish abstract objects.
Creating this metadata is automatic and only needs to be performed once each time
a new corpus is incorporated into the search tool.
3.3.3.2 Sense Disambiguation
Hacco and Shu (2002) noted that matches may contain search keywords used in
different senses. For example, if the keyword ‘seal’ was used to look for biological
analogies to sealing a joint, many matches that contain ‘seal’ in the aquatic
mammal sense are clearly irrelevant.
Ke et al. (2010) explored using the word sense disambiguation algorithm

developed by Pedersen and Kolhatkar (2009) to remove matches with keywords
used in unintended senses. However, word sense disambiguation is still an ongoing
challenge in computational linguistics, with the accuracy of many algorithms at
only about 60 %.
3.3.3.3 Wikipedia Categorization
Ke et al. (2009) also used Wikipedia categories of biology-related entries to sort

search results. This categorization can help designers skim through the search
results based on the levels of biological organization or the types of organisms
identified.
3.3.3.4 Subject/Object Categorization
Cheong and Shu (2012) used the typed-dependency parser (de Marneffe et al.
2006) to automatically identify the subjects and objects of verb keywords. Thus,
search results for verbs can be sorted by both subjects and objects, which enable
designers to more easily identify relevant search results for their design problems.
For example, for the keyword ‘prevent,’ search results that contain ‘water’ as the
object of ‘prevent’ may be especially useful to designers interested in
waterproofing.
3.3.4 Evaluation and Application of Search Matches
Evaluating the relevance of potential analogies across domains is a more complex

process than evaluating the relevance of matches to specific information sought.
For example, in a traditional Web search for ‘map of Toronto,’ it is far more
obvious which matches actually contain the map sought. However, with cross-
domain analogies, the physical entities involved are likely to be different between
analogy and problem. Thus, it is less straightforward to determine the relevance, as
well as usefulness of a given match. Described below are challenges involved from
the evaluation to the application of matches, and known ways of addressing these
challenges.
3.3.4.1 Types of Similarity
Mak and Shu (2004) observed how students applied descriptions of biological
phenomena functionally related to a given problem to solve that problem. Spe-
cifically, students were asked to develop concepts that result in ‘clean clothes,’ by
Fig. 3.2 Types of similarity comparisons between biological phenomena and developed
concepts, adapted from Mak and Shu (2004)
using the following description of a biological phenomenon by Purves et al.

(2001):
Barriers and local agents defend the body—skin is a primary innate defense against
invasion. The bacteria and fungi that normally live and reproduce in great numbers on our
body surfaces without causing disease are referred to as normal flora. These natural
occupants of our bodies compete with pathogens for space and nutrients, so normal flora
are a form of innate defense.
Four similarity types were observed between the description of the biological
phenomenon and the concepts developed using them as stimuli. These similarity
types are shown with examples in Fig. 3.2, whose axes are strategic accuracy and
abstraction of biological entities. Details on the four similarity types are as
follows:
1. Literal implementation: A literal implementation involves using biological
entities, for example bacteria, directly to solve the engineering problem, for
example, by filling clothing pores to prevent dirt from settling. Here, the bio-
logical entities are not abstracted, but rather, used directly, with the same
strategy between source and problem domains.
2. Biological transfer: A biological transfer involves transferring the biological
entities, for example bacteria, into the solution domain, but without applying
the strategy presented in the biological domain. For example, bacteria are used
to provide the solution of clean clothes by eating dirt.
3. Anomaly: An anomalous solution involves neither the entities nor the strategy
from the biological phenomenon. Some anomalous concepts are due to lack of
understanding. Other anomalous concepts are likely due to fixation on a few
words, for example ‘motor’ in motor proteins, in the text description while
disregarding the overall strategy or principle presented.
4. Analogy: The intended analogous solution accurately applies the strategy from
the biological phenomenon to the concept without transferring the biological
entities, for example bacteria, into the solution.
3.3.4.2 Analogical Mapping Tools
Although identifying the undesired categories of similarity reduced the number of

misapplied biomimetic design, two persisting difficulties observed by Mak and
Shu (2008) in students using descriptions of biological phenomena were as fol-
lows: inability to transfer information from biology to engineering and fixation on
specific phrases of the description of biological phenomena. This motivated further
studies where students were provided outlines of strategies to be applied to both
biological and engineering domains to facilitate analogical mapping. Compared to
results without the mapping tool, the quality of generated concepts improved.
However, although participants who used the mapping tool extracted strategies
consistent with the biological phenomena presented, they continued to apply
strategies to specific attributes of the given problem. Because generated concepts
varied with selected attributes, a possible method to reduce fixation and increase
the variety of solutions developed is to instruct the designer to list the problem
attributes, for example ‘clothing,’ ‘dirt,’ ‘detergent,’ and then develop ways in
which the biological strategies can be applied to each item in the list of attributes.
Further work by Cheong and Shu (2009) and Cheong et al. (2010) support that a
template helps designers to recognize causal relationships in descriptions of bio-
logical phenomena and transfer this relationship across domains. Figure 3.3 shows
an example of the template, which can be used to transfer functions of the relevant
strategy, while abstracting and mapping the associated entities from the biological
domain to the solution domain.
Fig. 3.3 Template for abstracting and mapping causally related functions from descriptions of
biological phenomena, adapted from Cheong et al. (2010)
3.3.4.3 Understanding Analogical Reasoning in Biomimetic Design
The challenges observed in novice designers while identifying and applying bio-
logical analogies motivated us to study analogical reasoning from a different
approach. In particular, we were interested in understanding why designers tend to
fixate on certain features of biological phenomena and fail to identify the relevant
analogy, as observed by Mak and Shu (2008), Cheong and Shu (2009), and Helms
et al. (2009).
Therefore, Cheong et al. (2012) studied verbal protocols of fourth-year
mechanical design students working in groups to solve one of three biomimetic
design problems. Each problem had a corresponding description of a biological
phenomenon to be used as a source of inspiration. The discourse of all group
members was recorded. Both qualitative and quantitative analyses were performed
on the verbal protocol.
We noted that participants made three types of similarity comparisons, based on
Gentner’s (1983) work on analogical reasoning:
• Entity: A comparison to superficial characteristics of entities of the biological
phenomenon.
• Function: A comparison to functions of the biological phenomenon.
• Strategy: A comparison involving a higher-order relation (strategy) from the
biological phenomenon.
Qualitatively, we noted that:

• Similarity between analogous elements at lower levels of comparison, for
example superficial (entity) and functional, prevented novice designers from
detecting the higher, overall strategy of the analogy.
• Domain knowledge can provide readily available associations at low levels of
comparison and induce fixation.
• Novice designers focus on mapping multiple features of the source analog,
instead of projecting multiple inferences (solutions) from the identified analogy,
perhaps due to lack of confidence in the analogy.
Quantitative analysis was performed to evaluate the types of similarity com-

parisons made over time. Figure 3.4 shows the trend of similarity comparisons
over time for one of the groups. This representation enables an objective assess-
ment of the analogical reasoning process occurring during biomimetic design
sessions. Such visualization of design protocols would be useful not only for
researchers studying the protocols, but also for designers to better understand their
own performance.
Applying computational linguistic techniques such as word frequency/co-
occurrence count and lexical chain analysis may enable more fine-grained and
automatic analysis of the analogical reasoning process. The application of natural
language processing techniques to analyze verbal protocols has been widely used
in other design research. Similar studies of verbal protocols in the context of
Fig. 3.4 Graph depicting the types of similarity comparison made over time, adapted from
Cheong et al. (2012). The similarity comparison index on the y-axis is the rolling average of
instances of similarity comparisons over five-second time segments
biomimetic design, which requires complex analogical reasoning, may warrant

more focus in the future.
3.3.5 Automatic Extraction of Causally Related Functions
This section summarizes an algorithm developed and applied by Cheong and Shu
(2012) to automatically extract causally related functions from text. As motivation,
Cheong and Shu (2009) and Cheong et al. (2010) had noted that causally related
functions in descriptions of biological phenomena may help designers identify and
apply relevant analogies. Gentner (1983) puts forth that higher-order relations,
such as causal relations, provide a framework for correct analogical transfer. Other
researchers in biomimetic design, including Goel et al. (2009) and Chakrabarti
et al. (2005), have modeled biological systems using a causality framework.
Computational linguistics may be applied to extract causal information from
text. Some of the most promising algorithms (Khoo et al. 2000; Girju 2003) use a
set of explicit causative words, for example ‘causes’ or ‘because,’ and lexico-
syntactic patterns to identify cause and effect from text. While these algorithms are
able to extract explicit causal relations between concepts, our work focused on
extracting causally related functions that are implicit and may be more relevant for
biomimetic design.
3.3.5.1 Extraction Method
A set of linguistic patterns that represent causally related functions was first
manually identified from seven chapters, covering a variety of topics, from our
search corpus, Life (Purves et al. 2001). Table 3.1 shows the linguistic patterns
identified.
For each pattern, a particular grammatical relation is held between two causally
related verbs that can be automatically detected from text parsed with Stanford
typed-dependency parser v1.6.7 (de Marneffe et al. 2006). For example, in
‘Lysozymes destroy bacteria to protect animals,’ the parser is able to identify the
grammatical relation open clausal complement between ‘destroy’ and ‘protect,’
that is, xcomp(destroy-2, protect-5). The numbers within the parentheses indicate
the position of the tagged word in the sentence parsed. The definitions of the
identified grammatical relations are shown in Table 3.1.
The extraction algorithm compares the part-of-speech tags and dependency
relations of the corpus to rules defined based on the linguistic patterns of
Table 3.1. The algorithm mainly looks for types of dependency relations, for
example xcomp, prepc_by, or purpcl, and uses the part-of-speech tags to test
additional rules, such as the exception listed under Pattern #1 in Table 3.1.
The extraction algorithm was implemented in our search tool, which takes a
verb keyword as input and automatically retrieves enabling functions for the
keyword verb. In other words, the search tool returns causal relations that are
formed between the search keywords (desired functions) and the identified
enabling functions. Figure 3.5 shows the results of an example search.
Some matches that contained nonmeaningful verbs were removed. These verbs
include light verbs, simple causative verbs, and other frequently appearing verbs in
the corpus. A light verb has little semantic meaning on its own but becomes more
meaningful when combined with an object, for example ‘use’ versus ‘use heat.’
Simple causative verbs, including ‘cause,’ ‘lead to,’ ‘allow,’ are used to identify
explicit causality, but on their own, do not tend to provide functional information.
Manning and Schutze (1999) note that the most frequently appearing words in a
corpus are likely to be semantically weak, that is, not meaningful. Cheong and Shu
Table 3.1 Typed-dependency relations and corresponding syntactic rules used to extract caus-
ally related functions, adapted from Cheong and Shu (2012). ‘DR’ stands for dependency relation
1. Lysozymes destroy bacteria to protect animals
The verb ‘protect’ is an open clausal complement (DR: xcomp) to the verb ‘destroy,’ that is,
‘protect’ does not have its own subject, but has the same subject as ‘destroy’
Exception: When the first verb is intransitive, that is, does not have an object, the verbs are
usually not causally related. For example, ‘I like to swim’ does not express any causality
although ‘swim’ is defined as an open clausal complement to ‘like’
2. Bacteria are destroyed to protect animals
Similar to Pattern #1, the verb ‘protect’ is an open clausal complement (DR: xcomp) to the
verb ‘destroy.’ In this case, however, the main verb ‘destroy’ is in the passive voice and the
exception rule for Pattern #1 is ignored
3. Lysozymes destroy bacteria, protecting animals
The verb ‘protect’ is an open clausal complement (DR: xcomp) to the verb ‘destroy’
4. By destroying invading bacteria, lysozyme protects animals
The gerund ‘destroying’ is a prepositional clausal modifier of the verb ‘protect,’ linked with
the preposition ‘by’ (DR: prepc_by)
5. To protect animals, lysozymes destroy bacteria
The verb ‘protect’ is part of a purpose clause modifier ‘To protect animals,’ which specifies
the purpose of the following clause ‘lysozymes destroy bacteria’ (DR: purpcl)
6. Destroying bacteria protects animals
The gerund ‘destroying’ acts as a clausal subject for the verb ‘protect’ (DR: csubj)
Fig. 3.5 Enabling functions for the keyword ‘move’ extracted and categorized in search tool
using Purves et al. (2001), adapted from Cheong and Shu (2012)
(2012) confirmed that the most frequent verbs in the corpus included many light
and simple causative verbs and therefore created a procedure to remove matches
with such semantically weak verbs.
3.3.5.2 Algorithm Accuracy
Development testing was performed to examine the accuracy of the extraction

algorithm and identify sources of errors to be removed. The performance of the
algorithm was compared against the manual coding of causal relations by a single
coder. The test was performed on three randomly selected chapters that were not
used for algorithm development. Precision, recall, and F-measure, which are
defined below, were used to assess the accuracy.
• Precision = (# of correctly retrieved causal relations)/(# of causal relations
retrieved)
• Recall = (# of correctly retrieved causal relations)/(# of causal relations in text)
• F-measure = 2 * (Precision) * (Recall)/[(Precision) ? (Recall)].
The extraction algorithm scored precisions of 0.800–0.933, recalls of

0.762–0.778, and F-measures of 0.780–0.848. For all three chapters, high precision
and moderate recall were obtained. The narrow range of performance measures
suggests the consistency of the extraction algorithm over different topics of text.
The majority, that is, 69 %, of the sources of errors involved parsing errors.
These include the parser incorrectly identifying the part of speech of a relevant
verb, not being able to find dependency relations between causally related verbs,
and incorrectly identifying the dependency relation between causally related verbs.
The most frequent nonparsing error was when the algorithm was unable to identify
a complex causal relation.
3.3.5.3 Challenges
Most of the processing tasks required for the causal relation extraction are auto-
mated. Also, the extraction rules for the algorithm use domain-independent lin-
guistic patterns. Therefore, the process is scalable to extracting information from a
large amount of text. On the other hand, the algorithm does not completely capture
causally related functions in biological text, as described below.
Causally related functions from multiple sentences. Currently, the parser only
identifies grammatical relations between words within a single sentence, which
limits the extraction algorithm to find causally related functions from a single
sentence. Identifying causally related functions across multiple sentences would
require anaphora resolution. Anaphora resolution refers to determining the ante-
cedent or referent of an anaphor, such as the pronouns ‘them’ or ‘it.’ Hobbs (1978)
first tackled the problem by analyzing a parse tree, while Soon et al. (2001)
recently used a set of syntactic/semantic conditionals to evaluate the candidate
referent of anaphors. However, the performance of these algorithms still requires
improvement for practical applications.
Causally related functions involving light verbs. The current algorithm removes
matches with instances of light verbs. A technique called light verb construction
reduces a light verb and its following object into a ‘heavy’ verb that has more
semantic context on its own, for example, by reducing ‘to take a walk’ into ‘to
walk.’ The simplest technique identifies whether the complement noun itself can
be used in a verb form. Stevenson et al. (2004), however, observed various lim-
itations in light verb construction. For instance, original nuances may be lost, for
example, the meaning of ‘to use heat’ differs from the meaning of ‘to heat.’
3.4 Examples Applying the Natural Language Approach
The natural language approach enables designers to retrieve biological analogies

that are not limited to those compiled in databases. Our research group applied the
natural language approach to a number of case studies, including design for
remanufacture (Shu and Flowers 1999; Hacco and Shu 2002), authorized disas-
sembly (Saitou et al. 2007), microassembly (Shu et al. 2003, 2006), sensing (Lenau
et al. 2008), redesign of fuel cells (Currie et al. 2009; Ke et al. 2009, 2010),
protection during hobbies (Cheong et al. 2008), and protection from lunar regolith
(Davidson et al. 2009). Below, three of these case studies, incorporating insights as
they were gained over the course of a decade, are presented.
3.4.1 Design for Remanufacture
Snap fits are often used as a fastening method due to their ease of assembly.
However, snap fits frequently break and are difficult to repair during remanufac-
ture. For example, a toner cartridge with failed snap fits would likely be discarded
rather than repaired or reused. Hacco and Shu (2002), therefore, searched for
biological analogies that could inspire solutions to redesign snap fits for easier
remanufacture.
3.4.1.1 Keywords and Biological Phenomena Identified
Initial keywords ‘remanufacture’ and ‘refurbish’ did not retrieve any matches in
the biological text. The authors therefore used direct and related synonyms of the
initial keywords for the search, which included keywords such as ‘renew,’
‘restore,’ ‘repair,’ ‘correct.’
The keywords ‘repair’ and ‘correct’ located useful biological phenomena. The
first phenomenon is based on the strategy used by plants to replace damaged parts:
The defense systems of plants and animals differ. Animals generally repair tissues that
have been infected. Plants, on the other hand, do not make repairs. Instead, they seal off
and sacrifice the damaged tissue so that the rest of the plant does not become infected. This
approach works because most plants, unlike most animals, can replace damaged parts by
growing new stems, leaves, and roots (Purves et al. 2001).
Relevant to the design problem, plants do not expend resources to repair

damaged parts but instead grow new parts to replace damaged ones. This strategy
can be implemented as a product component designed such that the failed feature
can be easily removed and replaced.
Another useful phenomenon located was based on fainting:
Blood must be returned from the veins to the heart so that circulation can continue. If the
veins are above the level of the heart, gravity helps blood flow, but below the level of the
heart, blood must be moved against the pull of gravity. If too much blood remains in
the veins, then too little blood returns to the heart, and thus too little blood is pumped to
the brain; a person may faint as a result. Fainting is self-correcting: A fainting person
falls, thereby moving out of the position in which gravity caused blood to accumulate in
the lower body (Purves et al. 2001).
The biological phenomenon of fainting suggests that a form of defensive failure

can be an effective strategy in preventing more serious failure.
3.4.1.2 Solution Developed
Incorporating the preemptive failure strategy to the redesign of snap fits, Hacco and
Shu (2002) specified predetermined breakpoints, shown in Fig. 3.6a, that may cause
earlier failure, but the part containing the snap fit feature can be more easily reused.
Fig. 3.6 Snap fit redesigned to facilitate easy refurbishment, adapted from Hacco and Shu
(2002). a Re-designed snap fit. b Replacement snap fit installed
Incorporating the strategy to sacrifice and replace parts, a possible planned

replacement part is shown in Fig. 3.6b that can be used once the sacrificial feature in
Fig. 3.6a fails.
3.4.2 Microassembly
Size effects complicate the handling and assembly of micromechanical parts.

Specifically, surface-related forces, for example, electrostatic, van der Waals and
surface tension forces, dominate gravitational forces. As a result, sticking between
the micropart and the gripping device during release hinders the automation of
microassembly operations. Shu et al. (2006) developed biologically inspired
solutions to overcome sticking in microassembly.
Keywords used include ‘remove’ and ‘release.’ In addition, ‘defend’ was selected
as a keyword, because biological systems often ‘remove’ parts as a defensive
mechanism.
The keyword ‘defend’ led to a match with DNA transcription as the basis for
how proteins are selectively synthesized to defend against infections, which led to
a concept where features with different geometries would be used to maximize or
minimize surface contact with the microobject as needed. However, this complex
interaction mapped into a relatively complex solution compared to one that was
developed based on the abscission principle, as described below.
Multiple keywords, including ‘defend,’ ‘remove,’ and ‘release’ led to the
biological phenomenon of abscission. Abscission is the process by which leaves,
petals, and fruits separate from a plant. A hormone called auxin is strategically
released in plants to direct growth. When auxin is no longer produced, other
hormones are released. The combined effect of these hormones breaks down parts
of plants, for example, stalks of leaves damaged through infection or are no longer
needed, as in the winter season, such that they become completely detached from
the plant. The base of some leaves contains the abscission zone, which is a special
layer of cells. Without auxin, these cells swell and form a cork-like material, which
cuts off nutrients to the leaf, forms a seal between the leaf and the plant, and
protects the plant once the leaf separates.
3.4.2.2 Solution Developed
The abscission principle was applied abstractly to overcome difficulties associated

with ‘sticking’ as follows. The microobject is released together with a part of the
tool designated as sacrificial, which can be of significant mass to take advantage of
gravity. The object can thus be easily released, and the sacrificial part of the tool
can then either remain with the microobject or be subsequently removed.
For the specific application of inserting a 0.6 mm metallic microscrew into a
plastic counterpart, the abscission zone is physically implemented as a polypro-
pylene rod of 4 mm diameter that is easily gripped and positioned by a small
industrial robot with six degrees of freedom and a specified repeatability
of ±0.02 mm (see Fig. 3.7). The tip of the polypropylene rod is locally melted by
heating and then pressed over the head of the screw. Upon solidification of the
polypropylene, a solid bond between the rod and the screw is formed, and a robot
can then manipulate the screw into the plastic counterpart. Once the screw is
tightened into its final position, the resulting increased torque will break the bond
between screw and rod.
Fig. 3.7 Implementation of the abscission strategy for micropart release, adapted from Shu et al.
(2006)
3.4.3 Protection from Lunar Regolith
The Light Detection and Ranging (LIDAR) device is an optical instrument that can
detect particle concentrations kilometers above the instrument itself. A LIDAR
consists of a high-powered laser that points upward and an optical receiver. A
cover, when closed, protects both the laser beam canister and the receiver lens.
Improved protection is sought for the lens both during operation (when the cover is
open) and while idle. Lunar regolith, or lunar dust, causes significant problems in
lunar exploration because the dust is pervasive, abrasive, and has been known to
damage and cause premature failure of lunar exploration equipment.
Two components of the lunar environment limit the choice of materials. First,
the thin lunar atmosphere approximates a hard vacuum and eliminates the speci-
fication of materials such as polymers, which will outgas in a vacuum and result in
severe physical degradation. Second, in the most common lunar exploration areas,
temperatures range from 120 C during the day to -150 C at night. Solutions
therefore must accommodate large and rapid temperature swings that occur during
the change from day to night and vice versa, further limiting the selection and
arrangement of materials.
Both mechanical and electrostatic aspects contribute to lunar regolith adhering
strongly to all surfaces. Without an atmosphere, no wind rounds regolith particles.
The sharp, jagged edges of regolith cause the mechanical aspect of adhesion, as
well as abrasion of mechanical seals. Furthermore, the small size of most regolith
particles (below 70 lm) leads to infiltration of almost all mechanical systems.
Positive charging of particles by solar wind during the lunar day and negative
charging by plasma electron currents at night enable particles to cling to
ungrounded surfaces. Finer regolith grains will levitate under this electrostatic
charging and thus, dust is present at the instrument level, even without mechanical
disturbance of regolith.
‘Protect’ and its corresponding biologically meaningful keywords ‘cover,’ ‘sur-

round,’ ‘inhibit,’ ‘destroy,’ ‘change shape,’ ‘bind,’ ‘repel,’ etc., are used as the
search keywords.
Several biological phenomena were found to be relevant for potential design
solutions. The keyword ‘protect’ identified bivalve shells that deflect sand away
from mating surfaces and have a ligament joint that is more resistant to particle
fouling than a rotational hinge. The keyword ‘repel’ identified phenomena such as
waxy coating on hair that repels water and the use of like charges to repel. The
keyword ‘destroy,’ although seemingly antonymous to ‘protect,’ identified phe-
nomena describing proteins destroying invading bacteria to protect animals.
3.4.3.2 Solutions Developed
Strategies from the first two biological phenomena, bivalve shells and repelling
charges, were implemented in the chosen two-piece cover system concept, shown
in Fig. 3.8 that mimics the curvature and ligament joint of the bivalve shell. The
curvature of the lids reduces the amount of lunar dust falling onto the inside lens
area during opening, and the ligament hinge avoids the relative motion found in
rotating joints, increasing resistance to invasive dust particles. When the cover is
open during operation of the LIDAR, high-voltage DC electromagnetic fields repel
dust particles away from the lens surface.
3.4.4 Conclusions from Examples
Below are retrospective observations from first-hand experience in the selection of

biological analogies, extraction of biological strategies and their application in
engineered systems described in our examples.
3.4.4.1 Identify Other Preventive Solutions in Biology
Many of the biological analogies discussed in our examples are based on pre-
ventive strategies in biology. For the remanufacture and LIDAR protection
example, this is not surprising because of the nature of the problems and the
keywords used for the searches. On the other hand, it is interesting that a pre-
ventive strategy, that is, abscission in plants, was also identified and finally chosen
to address the microassembly release problem.
In biology, the consequence of failure or damage may be dire. Therefore,
strategies such as abscission and fainting are used before more critical damage to
Fig. 3.8 CAD model of the bivalve-inspired lid assembly, adapted from Davidson et al. (2009)
the organism occurs. There is likely an abundance of preventive strategies in

biology that are highly relevant to engineering, particularly the development of
more sustainable products and processes.
3.4.4.2 Implement Strategies Passively to Enable/Replace Active

Solutions
Although not detailed in this chapter, earlier reports of both the remanufacture and
microassembly examples include biological analogies that were not as obviously
or successfully implemented in the engineering solution. For example, identified
using the keyword ‘repair’ for the remanufacture problem are DNA repair
mechanisms. Specifically, excision repair targets damaged sections of a DNA
molecule, including that which occurs during the life of the cell. Chemically
damaged abnormal bases are excised and replaced with functional bases. Clearly,
this phenomenon is highly analogous to the repair of damaged parts during
remanufacture. The text on excision repair by Purves et al. (2001) follows:
For example, in excision repair, certain enzymes ‘inspect’ the cell’s DNA. When they find
mispaired bases, chemically modified bases, or points at which one strand has more bases
than the other (with the result that one or more bases of one strand form an unpaired loop),
these enzymes cut the defective strand. Another enzyme cuts away the bases adjacent to
and including the offending base, and DNA polymerase and DNA ligase synthesize and
seal up a new (usually correct) piece to replace the excised one.
While the above description confirms relevance, it is difficult to apply the active
strategies given. For instance, who or what can be used in place of enzymes in an
engineering solution to ‘inspect, find, cut, synthesize, seal up, and replace’ the
defective parts, activities needed in excision repair? Relegating these duties to a
human repair technician confers no improvement over the current situation in
remanufacture. Further details on excision repair from a more advanced source
mention conformation changes that result from the interaction of enzymes and
contribute to the above activities in the repair sequence. The conformation change
strategy was then matched to the engineering strategy where failure induces a
conformation change that helps release the part for replacement, for example, self-
removal during failure, somewhat implemented already in the chosen solution
described.
Another biological analogy identified for the microassembly example also
involves complex molecular interactions. This phenomenon was developed into a
potentially feasible physical concept, but was far more complex to implement than
the polypropylene rods used to represent the abscission zone in the chosen solution
described. However, as engineering operates at increasingly smaller scales, it is
possible that these molecular strategies can be better implemented directly.
In the lunar regolith example, the charge-based solution, although also ‘active’
in that energy is required, can be implemented without the need for a physical
actor, for example, an astronaut, a robot, or a wiper to clear the lens of regolith
while the LIDAR is being used, all of which would be more complex as well as
potentially unsafe.
3.4.4.3 Consider Analogies by Environment in Addition to Function
In addition to searching for and selecting biological analogies based on functional

similarity, as suggested by many biomimetic design researchers, the lunar regolith
example highlights the role of environmental similarity. Specifically, while the
clam was identified more by superficial than functional characteristics, it turned
out to provide a wealth of transferable strategies. The clam’s physical character-
istics that are suited to its particulate-laden environment, for example, shell shape
that sheds sand and ligament hinge less susceptible to particle fouling, are also
well suited to protection from lunar dust. Conversely, strategies from a more
obvious eye protection analogy, though more functionally similar, turned out to be
much less directly transferable to the lunar environment. For example, tears/fluids
are unsuitable in the lunar environment, and wiping (by the eyelid) is unsuitable
for abrasive particles. Therefore, we now support the use of adjective keywords
and sort matches by the objects the adjectives describe, in our search tool.
Helms et al. (2009) indirectly refer to the identification of biological analogies
by environment as well as function though champion adapters—organisms that
survive in the most extreme of the environment of interest.
3.4.4.4 Explore Transferring Strategic in Addition to Superficial

Similarity from Biological Analogies
While many of our example solutions, for example, for remanufacture and mi-
croassembly, apply strategies from biological phenomena that share little super-
ficial similarity, many biomimetic design examples mimic biological phenomena
that share both functional and superficial similarity with the design solution. For
example, many legged robots are based on, and appear superficially similar to,
insects; Velcro shares both functional and physical similarity to the burrs on which
it is based; even lotus-effect-based solutions share physical and functional simi-
larities at the micro- and possibly nano-level with the lotus surface.
Several design researchers (Bonnardel 2000; Hon and Zeiner 2004; Tseng et al.
2008) report that cross-domain analogies, which often involve functional but not
superficial similarities, can evoke creative design solutions. Cross-domain analo-
gies are frequently more challenging to apply than within-domain analogies
because designers must identify similarities at the functional level, while there
may be little similarity at the superficial level. However, because it is easier for
humans to recall and observe information based on superficial similarity, designers
left to identify biological analogies based on their own or others’ observation or
recollection of knowledge may be more likely to develop solutions that directly
mimic biological systems/phenomena based on superficial similarities.
Our natural language search approach has focused on using keywords, for
example verbs and adjectives, to more objectively identify potential analogies
based on functional and environmental similarity. Our examples have demon-
strated the ability of this approach to identify analogies that seem unrelated, but
are in fact functionally or environmentally relevant to the design problems. Our
approach is consistent with artificial intelligence systems that make comparisons
between designs at functional, behavioral, or causal levels to support analogical
design (Goel 1997).
3.4.5 Conclusions on the Natural Language Approach
Our work on the natural language approach to support biomimetic design has
progressed to incorporate more computational linguistic tools and to automatize
the processes involved. Our latest work on the automatic extraction of causally
related functions in particular creates the potential to extract useful information
from a large amount of text sources. At the same time, it is essential to question
whether the information extraction techniques are capturing the appropriate type of
information, and the completeness of the captured information.
In addition, difficulties experienced by designers when applying analogies
found in natural language sources must be considered. Natural language text, or
language in general, is a flexible yet rich medium to represent and communicate
knowledge. How designers interpret text descriptions of biological knowledge and
form appropriate analogies to design solutions should be studied in more detail.
Interestingly, computational linguistic tools can also assist in such cognitive
studies.
Described below are other possible benefits of the natural language approach,
including its potential to bridge the gap between the increasing amount of bio-
logical information available and capturing meaningful information to support
biomimetic design.
3.4.5.1 Translate Natural Language Text to Other Formats

that Support Biomimetic Design
An extensive amount of information in biology is becoming available in digital/

online format. For example, Rebholz-Schuhmann et al. (2005) observed that the
number of articles available from Medline (Medical Literature Analysis and
Retrieval System Online) is exponentially increasing.
Although indexing biological information for the purpose of biomimetic design
is extremely useful, we cannot expect either biologists or designers to compile all
the available information (Bar-Cohen 2006). AskNature (www.asknature.org), a
popular database of biological information for biologically inspired design, has
been able to index about 1,400 strategies over the past five years, with most
strategies compiled before 2009. Vandevenne et al. (2011) noted that manually
compiling information would not efficiently integrate the vast majority of bio-
logical knowledge into design.
Importantly, most of the growing digital information is in the form of natural
language text. The field of natural language processing has existed since the 1950s
(Turing 1950). Therefore, plenty of tools exist that can be used to capture
meaningful information from the biological domain. This chapter referenced a
number of these tools that support the natural language approach to biomimetic
design.
One promising direction is to automatically translate natural language text into
a form that more directly supports other approaches to biomimetic design. Many
modeling frameworks developed for biomimetic design, for example, SBF models
by Goel et al. (2009) and SAPPhIRE by Chakrabarti et al. (2005), may benefit
from techniques developed to automatically translate text information to the
specific representation format used in modeling. Databases to support biomimetic
design become increasingly useful according to the amount of knowledge they
contain.
Limitations of this approach exist because English, for one, is a highly
ambiguous language. Many areas of natural language processing, such as word
sense disambiguation, anaphora resolution, and speech recognition, are actively
being researched. The accuracy of the information extraction and translation
efforts may be limited by the performance of the state-of-the-art natural language
processing techniques. On the other hand, increased application of natural lan-
guage processing in research domains other than computational linguistics will
likely support the advancement of existing techniques.
3.4.5.2 Determine Linguistic Relations Relevant to Biomimetic Design
An important step in the translation of natural language text to other useful formats
is identifying which syntactic or semantic information must be captured. For
example, the causal relation extraction algorithm used syntactic relations, such as
open clausal complement, prepositional clause modifier, etc., shown in Table 3.1,
to identify causally related verbs. Most computational models in biomimetic
design are based on formal frameworks to index and reason with biological
information. Chakrabarti et al. (2006) defined different relations of verbs, nouns,
and adjectives to express each SAPPhIRE construct. Goel et al. (2009) applied
well-defined syntax and semantics to specify SBF models. Nagel et al. (2010) used
functional basis lexicons to form verb–object pairs to represent biological systems.
Because these approaches have already defined specific frameworks to index
relevant biological information, researchers could then determine whether lin-
guistic relations relevant to the corresponding frameworks can be identified in
natural language text.
Since patterns of linguistic relations are present in all natural language text,
once techniques and strategies are developed to identify specific sets of useful
linguistic relations, they could be easily shared or extended to other information

extraction tasks. In a sense, the linguistic relations found in natural language text
could be considered an abstract framework that can be developed into specific
frameworks suitable for specific approaches of modeling biological information.
One challenge in any computing task is ensuring completeness. For instance,
the syntactic relations listed in Table 3.1 may not capture all causally related verbs
in a corpus. In fact, another syntactic relation that could be used to identify
causally related verbs was located while testing the relations. Therefore, the
framework required for information extraction may be learned while examining
the source of the information sought.
While many researchers in biomimetic design are working toward developing
frameworks for indexing/reasoning, fewer are working on locating appropriate
information. The examination of linguistic relations and corresponding computa-
tional linguistic techniques may thus advance biomimetic design.
3.4.5.3 Identify Patterns in Biological Phenomena
One benefit that comes with the ability to process a large amount of information is
the chance to identify patterns in information. TRIZ was developed based on
manually observing patterns of innovation in over 40,000 patents and has been
used as an effective problem-solving tool. Such a procedure is obviously difficult
to replicate, given the significant undertaking required. BioTRIZ developed by
Vincent et al. (2006) was based on about 500 biological phenomena. Son et al.
(2012) have recently investigated the patterns of biological transformations in 113
biological systems/organisms.
With highly automated information extraction techniques, such as the causal
relation extraction algorithm by Cheong and Shu (2012), patterns of strategies
used by biological systems could be identified. These patterns may reveal ubiq-
uitous solutions employed in nature, which could be mapped to solve a variety of
problems in engineering. Hoeller et al. (2007) propose that capturing recurring
solutions in biology would inspire effective designs of sustainable products or
services. Yen and Weissburg (2007) observe that the generality and robustness of a
particular biological strategy may be determined by whether that strategy is
implemented across many organisms. In addition to advancing biomimetic design,
significant findings on patterns of biological strategies may also advance our
understanding of biology.
3.5 Summary
This chapter summarized previous work on the natural language approach to

biomimetic design and discussed some possible directions for future work. Three
key points are as follows:
1. The natural language approach can enable designers to identify nonobvious,

previously unused analogies as demonstrated by our group’s application
examples.
2. Various computational linguistic tools can be used to develop and automate the
natural language approach.
3. An automated natural language approach could be applied to translate the
enormous amount of biological information in natural language format to
knowledge that more directly informs design, thereby supporting other work in
biomimetic design.
References
Bar-Cohen Y (2006) Biomimetics—Using nature to inspire human innovation. Bioinspiration

Biomimetics 1(1):1–12
Benyus JM (1997) Biomimicry: Innovation inspired by nature. William Morrow & Co, New York
Bonnardel N (2000) Towards understanding and supporting creativity in design: Analogies in a
constrained cognitive environment. Knowl-Based Syst 13(7–8):505–513
Bruck H, Gershon A, Golden I, Gupta S, Gyger L, Magrab E (2007) Training mechanical
engineering students to utilize biological inspiration during product development. Bioinspi-
ration Biomimetics 2(4):S198–S209
Chakrabarti A, Shu LH (2010) Biologically inspired design. Spec Issue Biologically Inspired Des
AIEDAM 24(4):453–454
Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju BS (2005) A functional representation for
aiding biomimetic and artificial inspiration of new ideas. Artif Intell Eng Des Anal Manuf
19(2):113–132
Cheong H, Shu LH (2009) Effective analogical transfer using biological descriptions retrieved
with functional and biologically meaningful keywords. Proceedings ASME International
design engineering technical conference, San Diego, CA, 30–2 Aug–Sept 2009, DETC2009-
86680
Cheong H, Shu LH (2012) Automatic extraction of causally related functions from natural-
language text for biomimetic design. Proc. ASME Int. Design engineering technical
conference, Chicago, IL, 12–15 Aug 2012, DETC2012-70732
Cheong H, Shu LH, Stone RB, McAdams DA (2008) Translating terms of the functional basis
into biologically meaningful keywords. Proceedings ASME international design engineering
technical conference, NYC, NY, 3–6 Aug 2008, DETC2008-49363
Cheong H, Chiu I, Shu LH (2010) Extraction and transfer of biological analogies for creative
concept generation. Proceedings ASME international design engineering technical confer-
ence, Montreal, Canada, 15–18 Aug 2010, DETC2010-29006
for functional terms of the functional basis. J Mech Des 133:021007
Cheong H, Hallihan G, Shu LH (2012) Understanding analogical reasoning in biomimetic design:
an inductive approach. Design computing and cognition DCC’12, College Station, TX, 7–9
June 2012
Chiu I, Shu LH (2004) Natural language analysis for biomimetic design. Proceedings of ASME
design engineering technical conference, Salt Lake City, UT, 28–2 Sept–Oct 2004,
DETC2004-57250
Chiu I, Shu LH (2005) Bridging cross-domain terminology for biomimetic design. Proceedings
ASME international design engineering technical conference, Long Beach, CA, 24–28 Sept
2005, DETC2005-84908
Chiu I, Shu LH (2007) Biomimetic design through natural language analysis to facilitate cross-
domain information retrieval. Artif Intell Eng Des Anal Manuf 21(1):45–59
Currie J, Fung K, Mazza AG, Wallace JS, Shu LH (2009) A comparison of biomimetic design
and TRIZ applied to the design of a proton exchange membrane fuel cell. 6th International
conference on innovation and practices in engineering design and engineering education,
Hamilton, Ontario, 27–29 July 2009
Davidson M, Bligh D, Maloney N, McKnight C, Young W, Shu LH, Potvin M-J, Warkentin A
(2009) Biomimetic design of a multi-layered dust protection system for optical instruments
operating in the lunar environment. 6th International Conference on Innovation and practices
in engineering design and engineering education, Hamilton, Ontario, 27–29 July 2009
de Marneffe M-C, MacCartney B, Manning CD (2006) Generating typed dependency parses from
phrase structure parses. Proceedings of the international conference on language resources and
evaluation, pp 449–454
Gentner D (1983) Structure-mapping: a theoretical framework for analogy. Cogn Sci
7(2):155–170
Girju R (2003) Automatic detection of causal relations for question answering. Proceedings of the
ACL 2003 workshop on multilingual summarization and question answering, vol 12.
pp 76–83
Goel AK (1997) Design, analogy, and creativity. IEEE Expert 12(3):62–70
Goel AK, Rugaber S, Vattam S (2009) Structure, behavior, and function of complex systems: The
structure, behavior, and function modeling language. Artif Intell Eng Des Anal Manuf
23(1):23–35
Goel AK, Vattam S, Wiltgen B, Helms M (2011) Cognitive, collaborative, conceptual and
creative—Four characteristics of the next generation of knowledge-based CAD systems: a
Hacco E, Shu LH (2002) Biomimetic concept generation applied to design for remanufacture.
Proceedings ASME international design engineering technical Conference, Montreal, Canada,
29–2 Sept–Oct, DETC2002-34177
Hearst M (1998) Automated discovery of WordNet relations. In: Fellbaum C (ed) WordNet: An
electronic lexical database. MIT Press, Cambridge
Helms M, Vattam SS, Goel AK (2009) Biologically inspired design: process and products. Des
Stud 30(5):606–622
Hobbs JR (1978) Resolving pronoun references. Lingua 44(4):311–338
Hoeller N, Salustri F, DeLuca D, Pedersen Zari M, Love M, McKeag T, Stephens E, Reap J,
Sopchak L (2007) Patterns from Nature. Proc. of the 2007 Society for Experimental
Mechanics Annual Conference and Exposition on Experimental and Applied Mechanics,
Springfield, MA, 4–6 June 2007
Hon KKB, Zeiner J (2004) Knowledge brokering for assisting the generation of automotive
product design. CIRP Ann 53(1):159–162
Karin K, Martha P, Owen R (2002) Extending PropBank with VerbNet semantic predicates.
Workshop on applied interlinguas, held in conjunction with AMTA-2002, Tiburon, CA, 8-12
Oct 2002
Ke J, Wallace JS, Shu LH (2009) Supporting biomimetic design through categorization of
natural-language keyword-search results. Proceedings ASME international design engineering
technical conference, San Diego, CA, 30–2 Aug–Sept, DETC2009-86681
Ke J, Chiu I, Wallace JS, Shu LH (2010) Supporting biomimetic design by embedding metadata
in natural-language corpora. Proceedings ASME international design engineering technical
conference, Montreal, Canada, 15–18 Aug 2010, DETC2010-29057
Khoo C, Chan S, Niu Y (2000) Extracting causal knowledge from a medical database using
graphical patterns. Proc. of the 38th Annual Meeting of the Association for Computational
Lingustics
Lenau T, Cheong H, Shu L (2008) Sensing in nature: using biomimetics for design of sensors.
Sensor Rev 28(4):311–316
Lindemann U, Gramann J (2004) Engineering design using biological principles. Proceedings 8th
international design conference design 2004, pp 355–360
Mak TW, Shu LH (2004) Abstraction of biological analogies for design. CIRP Ann
53(1):117–120
Mak TW, Shu LH (2008) Using descriptions of biological phenomena for idea generation. Res
Eng Design 19(1):21–28
Manning C, Schutze H (1999) Foundations of Statistical Natural Language Processing. MIT
Press, Cambridge
Miller GA (1995) WordNet: a lexical database for English. Commun ACM 35(11):39–41
Nagel JKS, Nagel RL, Stone RB, McAdams DA (2010) Function-based, biologically inspired
concept generation. Special Issue Biologically Inspired Des AIEDAM 24(4):521–535
Pedersen T, Kolhatkar V (2009) WordNet::SenseRelate::AllWords—A broad coverage word
sense tagger that maximizes semantic relatedness. Proceedings of the North American chapter
of the association for computational linguistics—Human language technologies conference,
Boulder, CO, 1-3 June 2009, pp 17–20
Purves WK, Sadava D, Orians GH, Heller HC (2001) Life, the science of biology, 6/e. Sinauer
Associates, Sunderland
Rebholz-Schuhmann D, Kirsch H, Couto F (2005) Facts from text—Is text mining ready to
deliver? PLoS Biol 3(2):e65. doi:10.1371/journal.pbio.0030065
Saitou K, Shalaby M, Shu LH (2007) Bioanalogous mechanical joints for authorized disassembly.
CIRP Ann 56(1):33–36
Shu LH (2010) A natural-language approach to biomimetic design. Special Issue Biologically
Inspired Design J AIEDAM 24(4):483–505
Shu LH, Flowers WC (1999) Application of a design-for-remanufacture framework to the
selection of product life-cycle fastening and joining methods. Int J Robot Comput Integr
Manuf (Special Issue Remanufacturing) 15(3):179–190
Shu LH, Lenau TA, Hansen HN, Alting L (2003) Biomimetics applied to centering in
microassembly. CIRP Ann 52(1):101–104
Shu L, Hansen H, Gegeckaite A, Moon J, Chan C (2006) Case study in biomimetic design:
handling and assembly of microparts. ASME international design engineering technical
conferences & computers and information in engineering conference, Philadelphia, PA, 10–13
September 2006, DETC2006-99398
Shu LH, Ueda K, Chiu I, Cheong H (2011) Biologically inspired design. CIRP Ann
60(2):673–693
Son JJ, Raulf C, Cheong H, Shu LH (2012) Applying patterns of transformation in biology to
design. Proceedings ASME international design engineering technical conference, Chicago,
IL, 12–15 Aug 2012, DETC2012-71296
Soon WM, Ng HT, Lim DCY (2001) A machine learning approach to conference resolution of
noun phrases. Comput Linguist 27(4):521–544
Stevenson S, Fazly A, North R (2004) Statistical measures of the semi-productivity of light verb
constructions. Proceedings of the ACL 2004 workshop on multiword expressions: Integrating
processing, pp 1–8
Stone RB, Wood KL (2000) Development of a functional basis for design. J Mech Des
122(4):359–369
Tseng I, Moss J, Cagan J, Kotovsky K (2008) The role of timing and analogical similarity in the
stimulation of idea generation in design. Des Stud 29(3):203–221
Turing AM (1950) Computing machinery and intelligence. Mind 59:433–460
Vakili V, Shu LH (2001) Towards biomimetic concept generation. Proceedings ASME design
technical conferences, Pittsburgh, PA, 9–12 Sept 2001, DETC2001-21715 (DTM)
Vandevenne D, Verhaegen P-A, Dewulf S, Duflou JR (2011) A scalable approach for the
integration of large knowledge repositories in the biologically-inspired design process.
International conference on engineering design, Technical University of Denmark, 15–18 Aug

2011
Vattam SS, Helms ME, Goel AK (2008) Compound analogical design: Interaction between
problem decomposition and analogical transfer in biologically inspired design. Design
computing and cognition DCC’08, Atlanta, GA, pp 377–396
Vincent JFV, Mann DL (2002) Systematic technology transfer from biology to engineering.
Philos Trans R Soc Phys Sci 360(1791):159–173
Vincent JFV, Bogatyreva OA, Bogatyreva NR, Bowyer A, Pahl A-K (2006) Biomimetics: its
practice and theory. J R Soc Interface 3(9):471–482
Wilson J, Chang P, Yim S, Rosen DW (2009) Developing a bio-inspired design repository using
ontologies. Proceedings ASME international design engineering technical conference, San
Diego, CA, DETC2009-87272
collected contributions. Bioinspiration Biomimetics doi: 10.1088/1748-3190/2/4/E01
Chapter 4
A Thesaurus for Bioinspired Engineering
Design
Jacquelyn K. S. Nagel
Abstract Biological systems provide insight into sustainable and adaptable

design, which often leads to designs that are more elegant, efficient, and sustain-
able. There are, however, significant hurdles to performing bioinspired design.
This chapter presents a design tool, the engineering-to-biology thesaurus, that
addresses several challenges engineers may encounter when performing bioin-
spired design, allowing engineers without advanced biological knowledge to
leverage nature’s ingenuity during engineering design. Along with the thesaurus
tables, detailed information on the thesaurus model, structure, population, term
placement, term placement review, and limitations is provided. Applications of the
design tool are discussed. Examples are provided to demonstrate the goals and
applications of the design tool followed by a review of integration with compu-
tational design tools.
Keywords Thesaurus
Function-based design
Design tools
Functional

modeling Functional basis Translation Analogies Concept generation

Identification Inspiration search Brainstorming Dialogue facilitation
4.1 Introduction
Engineering design is considered a creative endeavor involving many activities,

which encourages the use of engineering principles, imagination, prior knowledge,
stored knowledge, and a designer’s intuition to create engineering solutions. The
resulting solution may or may not be innovative, novel, or what some would call
creative; however, the design should fulfill a purpose or answer a need (Hyman 1998;
J. K. S. Nagel (&)
Department of Engineering, James Madison University Harrisonburg,
Harrisonburg, VA, USA
e-mail: nageljk@jmu.edu

64 J. K. S. Nagel
Otto and Wood 2001; Dym and Little 2004; Ulrich and Eppinger 2004; Voland 2004;
Ullman 2009; Dieter and Schmidt 2009; Cross 2008). To arrive at a solution, it is not
uncommon for engineers to use design methods and tools, make analogies among
different engineering disciplines (i.e., an electrical resistor and mechanical damper
are mathematically analogous) during ideation to find solutions, or use metaphors to
frame or assist in defining the design problem (Hey et al. 2008). The use of design
tools and methods is recognized as standard practice in both industry and academia.
These tools and methods play a pivotal role in many of the design activities from
concept generation to detailed design. The leap made between engineering disci-
plines using analogies is to be expected as one gains more experience. Making a leap
between domains, however, is less likely to occur without an impetus. Take for
instance Velcro, if it were not for the curiosity of George de Mestral that caused
him to investigate how the tiny burrs he and his dog accumulated from walking
through wooded areas, modern day hook and loop may never have been invented or it
may not be as effective. George de Mestral’s chance observation of a biological
phenomenon resulted in a very simple, reusable material (de Mestral 1955) that has
been used for securing everyday items such as shoes to mission-critical items needed
for exploring space.
The natural world provides numerous cases for inspiration in engineering
design. Though biological organisms, phenomena, and strategies, herein referred
to as biological systems, provide a wealth of elegant and ingenious approaches to
problem solving, there are challenges that prevent designers from leveraging the
full insight into the biological domain. The leap from engineering to biological
science and back has posed a challenge. Engineers often struggle with how to best
use the vast amount of biological information available from the natural world
around them. Often, it is because there is a knowledge gap or terminology is
difficult or terminology takes different meanings. Moreover, the time required to
learn and become fluent in biology poses too large a hurdle. This reveals a fun-
damental problem of working across the engineering and biology domains. The
effort and time required to become a competent engineering designer creates
significant obstacles to becoming sufficiently knowledgeable about biological
systems (the converse can also be said).
This chapter presents a design tool, the engineering-to-biology thesaurus, that
addresses the main barriers to bioinspired design and aims to lower the hurdle,
allowing engineers without advanced biological knowledge to leverage nature’s
ingenuity during engineering design. The three key goals of this thesaurus are to
(1) lessen the burden when working with knowledge from the biological domain
by providing a link between engineering and biological terminologies; (2) assist
designers by establishing connections between the two domains; and (3) to
facilitate bioinspired design.
From the perspective of engineering design, biological information that can
interface with existing design methods will further lessen the burden on the
designer and increase the probability of looking to nature for inspiration. One
approach to interfacing biological information with engineering design is through
terminology or, more specifically, through grouping of synonyms and related
4 A Thesaurus for Bioinspired Engineering Design 65
Fig. 4.1 Graphical representation of the engineering-to-biology thesaurus
concepts in a classified form as a thesaurus. Linking biology terms to engineering

modeling terms (used for functional representation and abstraction) assists in re-
framing the biological terminology in an engineering context. Thus, the biological
information is accessible to engineering designers with varying biological
knowledge, but a common understanding of engineering design methodologies.
The engineering-to-biology thesaurus design tool, described in this chapter and
graphically depicted in Fig. 4.1, addresses several challenges engineers may
encounter when performing bioinspired design, which include:
• Terminology differences.
• Understanding how a biological system works or functions.
• Discovering relevant biological systems.
• Abstracting biological principles/solutions for inspiration.
Acknowledging the difficulties of utilizing biological systems in engineering
design, the engineering-to-biology thesaurus serves as a versatile design tool that
affords design engineers, with a limited biological background, a means for
developing connections between nature and engineering. The primary applications
include assisting engineers with translation of biological information, communi-
cation with biologists, and concept generation with biological information.
Building upon multiple design language research efforts, the engineering-to-biol-
ogy thesaurus design tool has the potential to play a key role in several bioinspired
design activities.
66 J. K. S. Nagel
4.2 Background and Related Work
It is evident that nature can inspire innovative engineering designs (Toko 2000;
Bar-Cohen 2006a, b; Benyus 1997; Stroble et al. 2009a; Vincent et al. 2006;
Brebbia 2006, 2008; Brebbia and Carpi 2010; Brebbia and Collins 2004; Brebbia
et al. 2002; Forbes 2006; The Biomimicry Institute 2009). Biological systems
provide insight into sustainable and adaptable design, which often leads to designs
that are more elegant, efficient, and sustainable. For engineering designers, how-
ever, to adopt such a practice, design tools, techniques, and methods are needed.
Utilizing biological information during the engineering design process has taken
many forms. Inspiration for solving or finding direct solutions to engineering
problems has been obtained through chance observances (Hill 1995; Nachtigall
1989, 2000, 2002), functional keyword searches (Chiu and Shu 2007a, b; Vakili
and Shu 2001; Bruck et al. 2007), systematic reverse engineering (Wilson and
Rosen 2007; Lindemann and Gramann 2004), use of function–structure–behavior
terms to search a database (Sarkar et al. 2008; Wilson et al. 2009; Design and
Intelligence Laboratory 2010), TRIZ (Vincent et al. 2006), analogical reasoning
(Tsujimoto et al. 2008; Mak and Shu 2004, 2008; Helms et al. 2009; Vattam et al.
2010a), and functional representation through functional models (Nagel et al.
2008, 2010a, b; Vakili and Shu 2007; Shu et al. 2007; Stroble et al. 2009b; Vattam
et al. 2010b). Although each method has a different procedure, they all share one
thing in common; the promising biological system or phenomena must be
abstracted to capture the functional principle.
The research presented in this chapter explores the structure and purpose of an
engineering design thesaurus and how it enhances an existing design lexicon for
bioinspired design. Researchers at many universities are working on the knowl-
edge transfer problem between the engineering and biological domains by further
developing function- or function–behavior–structure-based design languages.
Research into the engineering-to-biology thesaurus builds upon multiple design
language research efforts.
The formal idea of a function-based design language, a standard set of engi-
neering function and flow terms for systematically creating function structures,
was originally proposed by Pahl and Beitz (2007). A function represents an
operation performed on a flow of material, signal, or energy. Numerous
researchers further evolved this set of generally valid functions and flows. Hundal
proposed a further refined set of function and flow classes in (Hundal 1990);
however, flows were excluded. Little et al. (1997) developed a set of function and
flow terms, which classified both functions and flows at class and basic levels.
Szykman et al. (1999) created a standardized taxonomy of function and flow terms,
separated into classes down to the fourth level, for the purpose of computer-based
design. Separately, but at the same time, Stone and Wood developed a well-
defined standardized modeling lexicon comprised of defined function and flow sets
with definitions and examples, entitled the Functional Basis (Stone and Wood
2000). Hirtz et al. (2002) later reconciled the Functional Basis into its most current
set of terms, with research efforts from the National Institute of Standards and
Technology (NIST), two universities, and their industrial partners. Within the
Functional Basis, there exist eight classes of functions and three classes of flows,
both having an increase in specification at the secondary and tertiary levels. There
are 21 secondary and 24 tertiary functions, accompanied by correspondent terms to
aid the designer in choosing the correct function. Similarly, there are 20 secondary
and 22 tertiary flows accompanied by correspondent terms.
In a similar vein, function–behavior–structure languages were being explored
and developed to provide designers a way to uniformly represent system behavior.
Umeda et al. (Umeda et al. 1990) postulated that structure correlates with a sys-
tem’s state, or with the physical description of an entity in a design, while behavior
was defined as the change in the state, and function was defined as the realization
of the behavior through the use of the design. The work by Goel et al. uses
structure to represent a physical description for components, function is the pre-
and post-conditions for the behavior of the system, and behavior is the transition
between states (Goel et al. 2009; Goel and Chandrasekaran 1992). To capture
environmental interactions with a system, Gero developed the situated function–
behavior–structure framework (Gero and Kannengiesser 2002). Building on this
prior work, Chakrabarti et al. (2005) developed an approach to describe natural
and artificial systems and their functionality. The representation is implemented in
a software package entitled Idea-Inspire that allows one to search a database
comprised of natural and artificial complex mechanical systems with a function–
behavior–structure, or verb–noun–adjective, set. Each entry’s motion or process is
described functionally by behavioral language in the form of a function–behavior–
structure model, which the user chooses from a predefined list of terms. The Idea-
Inspire software yields seven behavioral constructs following the SAPPhIRE
model—state change, action, parts, phenomenon, input, organ, and effect—for
each search result that adequately fits the chosen function–behavior–structure set
(Srinivasan and Chakrabarti 2009).
Researchers at the University of Toronto worked to provide designers with bio-
logically meaningful words that correspond to engineering functions. To identify
biologically meaningful words, the strategy developed by Chiu and Shu for searching
biological literature using functional keywords for design inspiration (Chiu and Shu
2007a, b) was employed. Cheong et al. (2008) used the search strategy in conjunction
with the terms of the Functional Basis to create the listing of biologically meaningful
words. Only the Functional Basis functions in the secondary, tertiary, and corre-
spondent levels were analyzed. Based on semantic relationships, the engineering
function terms of the Functional Basis were used to generate a list of biologically
significant and connotative functional keywords. In a similar vein, Stroble et al.
(2009c) worked to provide designers with biological terms that correspond to
engineering flows of the Functional Basis. Flow-type biological correspondent terms
were collected utilizing an organized verb–noun search that extracts collocated
words from a biological text. The Functional Basis flows in the class, secondary, and
tertiary levels were analyzed. The macrorelevant biological flow terms identified
were mapped to engineering flow terms through dictionary cross-referencing.
68 J. K. S. Nagel
The work by Cheong et al. concentrated on finding specific connections

between a subset of engineering functions and biological words, resulting in only a
partial thesaurus. Stroble et al. concentrated on finding specific connections
between engineering flows and biological terms, which also resulted in a partial
thesaurus that was later expanded to include biological function terms by Nagel
et al. (2010c). Both works share a common starting point, but diverged in
approaches to filling out a thesaurus that maps engineering terms to biology terms
as well as to what depth the thesaurus was populated. Additionally, the work of
Nagel et al. (2010c) combined multiple design language efforts by (Chakrabarti
et al. 2005; Srinivasan and Chakrabarti 2009; Cheong et al. 2008; Stroble et al.
2009c; Nagel et al. 2010c) into a single thesaurus. The result of that work is
presented in this chapter.
4.3 Engineering-to-Biology Thesaurus
The engineering-to-biology thesaurus design tool (Nagel et al. 2010c) aims to

facilitate collaboration between biologists and engineers and the discovery and
creation of bioinspired engineering solutions. Applications of the engineering-to-
biology thesaurus include, but are not limited to, (1) translation of biological
information to increase comprehension or develop connections, (2) concept gen-
eration with biological inspiration through function-based approaches, and (3)
dialogue facilitation between the engineering and biology communities. To
achieve the stated applications, the structure of the thesaurus was molded to fit the
knowledge and purpose of an engineering design perspective; biological synonyms
and related concepts to the engineering terms of a well-established modeling
lexicon are grouped together. The engineering-to-biology thesaurus, provided in
Tables 4.1, 4.2, 4.3 and 4.4, does more than arranging terminology of one domain
side by side with terminology of another; it serves as the intermediary between the
biology and engineering domains. Furthermore, this design tool increases the
interaction between the users and the knowledge resource (Lopez-Huertas 1997)
and aims to increase a designer’s efficiency when working across the engineering
and biology domains. Efficient information retrieval through the engineering-to-
biology thesaurus allows an engineering designer to cross into the biology domain
and gain functional knowledge without becoming overwhelmed by unfamiliar
biological systems and terminology. This is, and is envisioned to be, a work in
progress that slowly and steadily bridges the terminology gap between the two
domains based on contributions of a community of researchers and practitioners.
In this section, the thesaurus model, population methods of the biological flows
and functions, term placement, term placement review, and limitations are
explained.
Table 4.1 Engineering-to-biology thesaurus function terms
Functional basis terms Biological function correspondent terms
Class Secondary Tertiary
Branch Separate Bleaching, meiosis, abscission, mitosis, segment, electrophoresis, dialysis, denature, free, detach, release
Divide Divide, prophase, metaphase, anaphase, cleave, cytokinesis
Remove Deoxygenate, filtrate, liberate, expulsion, evacuate
Distribute Circulate, diffusion, exchange, disperse, scatter, spread, spray
Channel Import Absorb, attract, consume, inhale, intake
Export Bind, block, breakdown, excrete, inactivate, repel
Transfer Migrate, transfer
Transport Circulate, conduct, diffuse, pump, shift, displace, fly, swim, jump, bounce
Transmit Communicate, transduce
Guide Orient, position, slide, tunnel
Translate Synthesize, transcribe
Rotate Oscillate, spin, turn, swivel, roll
Allow Articulate
4 A Thesaurus for Bioinspired Engineering Design
DOF
Connect Couple Recombination, mate, build, phosphorylate, bond, synthesis, latch, lock, extend, link, overlap
Join Bind, adhere, bond, fuse
Link Clamp, activate, bind, project
Mix Blend, contract, exchange, fragment
Convert Convert Polymerize, synthesize, burn, gluconeogenesis, metabolize, grow, transduction, transduce, fermentation, glycolysis,
hydrolyze, hydrolysis, respiration, ionize, decompose, degrade, develop, mutate, photosynthesize
Control Actuate Activate, induce, trigger
Magnitude Regulate Electrophoresis, gate, organogenesis, respire, sustain, preserve, remain, stabilize, maintain, regulate
Increase Hyperpolarize, pinocytosis, grow, expand, multiply, replicate
Decrease Compress, coil, divide, fold, shorten, wrap
Change Pinocytosis, degrade, alter, bind, catalyze, contract, hydrolysis, twist, mutate, radiate, charged, slip, acclimatize,
alternate, fluctuate
Decrement Decarboxylation, constrict
69
(continued)
Table 4.1 (continued)
70

Shape Elongate, stretch, attach, spread
Condition Osmosis, constrict
Stop Clog, extinguish, halt, interphase, seal, suspend
Prevent Constrain, obstruct
Inhibit Cover, destroy, inhibit, repress, surround
Provision Store Conserve, hold, convert, deposit, photosynthesize
Contain Absorb
Collect Absorb, catch, breakdown, concentrate, digest, reduce
Supply Feed, lactate
Signal Sense Detect Detect, locate, see, smell, transduce
Measure Observe, monitor, gauge, watch
Indicate Fluoresce, communicate, react, mark
Process Learn
Support Develop, wrap
Stabilize Homeostasis, cling, hold, bind, connect
Secure Surround, envelope
J. K. S. Nagel
Table 4.2 Engineering-to-biology thesaurus material flow terms

Material Human Being, body
Gas Oxygen, nitrogen, chlorine
Liquid Acid, chemical, water, blood, solution, base, buffer, fluid,
plasma
Solid Object Fiber, body, substrate, microfilament, microtubules, structure,
chain, organ, nucleus, tissue, muscle, cilia, flagella, tube,
vein, heart, plant, somite, apoplast, stem, kidney, egg,
ovary, leaf, embryo, bacteria, chloroplast, carbon, adipose,
angiosperm, meristems, mineral, stoma, shoot, seed,
capillary, receptors, hair, bone, tendon, neuron,
sporangium, photoreceptor, mechanoreceptor,
chromosome, petiole, lysosome, archaea, cone, strand,
centriole, spore, zygote, sulfur, lipoprotein, nephron,
hyphae, plasmodesma, conifer, plasmid, plastid, xylem,
pigment, sperm, hippocampus, phloem
Particulate Cytokinin, pyruvate, nicotine, opium, glycerol, carotenoid,
GTP, ATP, urea, RNA, tRNA, mRNA, DNA, glucagon,
parathormone, cryptochromes, ligand, promoter, gene,
exon, intron, molecule, enzyme, lipid, hormone
Composite Enzyme, virus, ribosome, prokaryote, macromolecule,
polymerase, nucleotide, polypeptide, organelle, symplast,
mesophyll, brood, codon, messenger, DNA, RNA,
cytoplasm, organ, tissue
Mixture Gas–gas Air
Liquid– Hormone, melatonin, thyroxine, calcitonin, thyrotropin,
liquid estrogen, somatostatin, cortisol, glucagon,
adrenocorticotropin, testosterone, auxin, insulin,
intracellular fluid, extracellular fluid, spinal fluid, poison,
urine, peptide, solution, steroid
Solid– Adenosine, glomerulus, blastula, monosaccharide, membrane,
solid phosphate, ribosome, centrosomes
Solid– Algae, synapse, cell, glia, phytochrome, retina, protein,
Liquid receptor site, repressor, hemoglobin, blood, membrane,
bacterium
4.3.1 Thesaurus Model
The purpose of a thesaurus is to represent information in a classified form to group

synonyms and related concepts. A thesaurus of the English language has classes
and categories with an index of terms directing the user to the correct instance (i.e.,
noun, verb, adjective) of the term under examination. The engineering-to-biology
thesaurus design tool has a unique structure and classification; it is merged with the
reconciled Functional Basis as a set of biological correspondent terms. In the
Functional Basis lexicon, a function represents an action or transformation (verb)
being carried out, and a flow represents the type (noun), material, signal, or energy,
72 J. K. S. Nagel
Table 4.3 Engineering-to-biology thesaurus signal flow terms

Functional basis terms Biological flow correspondent terms
Signal Status Change, variation, lateral, swelling, catalyzed, translation,
exposed, active, separated, cycle, formation, reaction, redox,
deficient, saturated, diffusion, broken, hybridization,
orientation, resting, cue, magnetic, volume, under, organized,
fruiting, fatty, anaphase, metaphase, prophase, conjugation,
osmolarity, senescence, signal
Auditory Sound
Olfactory Smell
Tactile Pain
Taste Gustation
Visual Length, shortened, long, dark, full, double
Control Place, inhibit, release, excrete, develop, match, induce, digest,
integrate, translation, transduction, equilibrium, grow, splice,
capture, distribute, phosphorylation
Analog Binding, center, synthesis, photosynthesis
Discrete Flower, translocation
Table 4.4 Engineering-to-biology thesaurus energy flow terms

Functional basis terms Biological flow correspondent terms
Energy Acoustic Echolocation, sound wave
Chemical Calorie, metabolism, glucose, glycogen, ligand, nutrient,
starch, fuel, sugar, mitochondria, lipid, gibberellin
Electrical Electron, potential, feedback, charge, field
Electromagnetic Optical Light, infrared radiation
Solar Light, sunlight, ultraviolet light
Hydraulic Pressure, osmosis, osmoregulation
Magnetic Gravity, field, wave
Mechanical Muscle contraction, pressure, tension, stretch, depress
Pneumatic Pressure
Thermal Temperature, heat, infrared radiation, cold
passing through the functions of the system. There exist eight classes of functions
and three classes of flows, both having an increase in specification at the secondary
and tertiary levels. Both functions and flows have a set of engineering corre-
spondent terms that aid the designer in choosing Functional Basis terms during
model creation. The complete function and flow lexicon can be found in (Hirtz
et al. 2002).
The engineering-to-biology thesaurus provides biological correspondent terms
for engineering functions and flows at the class, secondary, and tertiary levels,
which follows the structure and classification of the Functional Basis. This
grouping highlights the synonyms and related concepts across the two domains.
The thesaurus does not include an exhaustive set of engineering and biology terms;
rather, it contains a representative set of engineering and biology terms. Biological
correspondent terms to the Functional Basis functions and flows are provided in
place of the original engineering correspondent terms, as shown in Tables 4.1, 4.2,
4.3 and 4.4. Only biological verbs and nouns that are synonymous to terms of the
Functional Basis are considered. The thesaurus does not include adjectives nor
does it include an index. The Functional Basis class-level terms, however, do
emulate the classes of a traditional thesaurus. Furthermore, the secondary- and
tertiary-level Functional Basis terms emulate the categories of a traditional the-
saurus. Thus, the classification is predetermined according to the Functional Basis
model; however, it remains the intermediary between the biology and engineering
domains.
A tool such as the engineering-to-biology thesaurus increases the interaction
between the users and the knowledge resource (Lopez-Huertas 1997) by presenting
the information as a lookup table. This simple format fosters one to make asso-
ciations between the engineering and biological lexicons, thus strengthening the
designer’s ability to utilize biological information.
4.3.2 Biological Functions
The majority of biological information is written in such a way that correlating

biological verbs with Functional Basis functions is relatively straightforward.
However, there are always exceptions. Well-known functional terms that appear in
a biological corpus may not have the meaning an engineer would typically know.
For instance, the term bleaching outside of the biological domain means to clean,
sterilize, or whiten, as most know. The biological meaning, however, refers to the
process of separation between the retina and opsin in vertebrate eyes and causes
the retinal molecule to lose its photosensitivity (Campbell and Reece 2003). It is
these types of exceptions that researchers were cognizant of when compiling the
set of biological correspondent function terms for the engineering-to-biology
thesaurus. Keyword searches of a biology textbook using the automated infor-
mation retrieval tool (Stroble et al. 2009d) and Functional Basis functions were
performed to gather a list of collocated verbs that occur within the same sentence
as the search word. To signify which function terms are utilized in both domains,
the Functional Basis term is repeated in the biological correspondent list. It should
be noted that some of the biological correspondent function terms are nouns that
name a process corresponding to a Functional Basis function. Identified biological
functions were cross-referenced in the Oxford American dictionary (McKean
2005), Henderson’s dictionary of biological terms (Lawrence and Holmes 1989),
and the Oxford Dictionary of Biology (Matrin and Hine 2000) before placement in
the thesaurus, which was at the discretion of the author. All other function terms
were obtained from research performed at the Indian Institute of Science and
University of Toronto, which are made explicit in (Nagel et al. 2010c).
74 J. K. S. Nagel
Functional terms from the Indian Institute of Science were collected from the
Idea-Inspire software. Every natural system entered into the software’s database
was indexed using the predetermined list of verbs, nouns, and adjectives. Ana-
lyzing the list of verbs by cluster (Srinivasan and Chakrabarti 2009) revealed
scientific terms applicable to biological systems grouped with engineering terms
exactly matching those of the Functional Basis. Utilizing multiple dictionaries as
in the prior analysis, the verbs of Idea-Inspire were mapped to the Functional Basis
functions as biological correspondent function terms.
Functional terms from the University of Toronto were collected from the work
of Cheong et al. (2008). As background work was already performed on the
semantic relationships of the biologically meaningful words to Functional Basis
functions, further investigation was not performed. Rather, the terms were directly
added to the thesaurus as biological correspondent function terms following the
classification in Cheong et al. (2008).
4.3.3 Biological Flows
In the authors’ experience, understanding biological terms that were considered

flows (material, signal, and energy) when utilizing biological systems for idea
generation or design inspiration posed the most difficulty. Determining whether a
biological material is liquid, solid, or a mixture by its name typically requires
domain knowledge that most engineers do not have, which can cause biological
information to be perplexing. Similarly, needing a reference to look up biological
terms each time, a potential biological system was found made the design process
tedious and disrupted thought patterns leading to decreased efficiency. Thus, the
inclusion of biological correspondent flow terms strengthens the ability of the
design tool to address the challenges engineers may encounter when performing
bioinspired design.
Identification of engineering-to-biology thesaurus biological correspondent
flow terms was achieved through keyword searches of a biology textbook using the
automated information retrieval tool (Stroble et al. 2009c, d). Functional Basis
functions were used for the keyword searches to extract biological nouns that an
engineering designer interested in function-based design might encounter. The
nouns that were collocated within the sentence to the search word were counted
and sorted by frequency of appearance. All nouns that appeared more than two
times were considered macrorelevent. Each macrorelevent term was determined if
it was of signal, material, or energy type through cross-referencing in the new
Oxford American dictionary (McKean 2005) and Henderson’s dictionary of bio-
logical terms (Lawrence and Holmes 1989) before being placed, which was at the
discretion of the author.
4.3.4 Term Placement Review
Mapping engineering terms to the biological domain, which the author is not an
expert in, requires a review of the relationships by a biologist. A biologist in two
instances reviewed the thesaurus: (1) when the biological correspondent flow
listing was generated and (2) when the biological correspondent function listing
was generated. The terms of the thesaurus in Tables 4.1, 4.2, 4.3 and 4.4 represent
the complete set of terms that are the result of the biologist’s reviews. Review of
the placement, type, and structure of the thesaurus terms was initially performed
by a biology student at Missouri University of Science & Technology after flows
were placed (Linsey 2008a, b). A large group of tertiary terms under discrete
control signal were moved to the secondary level as they could be either discrete or
analog control signals. Other misplaced biological correspondent terms were
moved between the tertiary-level material classifications. After the first-term
placement review, 32 (10.9 %) of the flow terms were moved to a more appro-
priate mapping, 20 (6.8 %) were removed completely, and 11 new terms were
added, thus changing the total biological correspondent flow term count to 285
from 294.
A second review of the placement, type, and structure of the thesaurus terms
was performed by a professor of zoology at Oregon State University after func-
tions were placed (Brownell 2010a, b). The professor of zoology reviewed both the
function and flow biological correspondent terms and offered his insight. To better
map the terminology to the engineering domain, biological terms of a similar type
or related concept that were scattered were moved to the same classification and
the multiple meanings of terms are emphasized through repetition of the terms
across classifications. For example, all the terms representing molecules across
multiple material tertiary terms were collected and placed into the particulates’
classification. A similar change was made for terms representing hormones, which
were placed in the composites’ classification. The term organ is found under
multiple classifications as it can be thought of as an object or a composite of
tissues. Other changes include removing terms due to ambiguity and changing
term tense. After the second-term placement review, 34 (11.9 %) of the flow terms
were moved, 24 (8.4 %) were corrected to the proper tense, and 22 (7.7 %) were
removed. Also, 7 new terms were added and 5 existing terms were repeated, thus
changing the total biological correspondent flow term count from 285 to 275. For
the biological correspondent function terms, the changes consisted of moving 2
(0.01 %) of the terms, correcting 1 (0.005 %) to the proper tense, and removing 5
(2.4 %) of the terms. Also, 1 new term was added and 3 existing terms were
repeated, thus changing the total biological correspondent function term count to
206 from 207.
76 J. K. S. Nagel
4.3.5 Thesaurus Details
Key challenges to the approach for populating the thesaurus were (1) the time
required searching each term to generate a listing of collocated terms and (2)
understanding the definitions provided by the three dictionaries used in the anal-
ysis. To determine the material, energy, or signal type of the flow term in question,
generally multiple biological dictionary entries were referenced. Considering
biological processes that perform a specific function within the system revealed
many macrorelevant terms that would have been overlooked if only verbs were
analyzed.
The Functional Basis offers a definition and example for each class, secondary,
and tertiary terms. Definitions of the correspondent terms are, however, not pro-
vided. Rather, the correspondent terms are synonyms to the Functional Basis
terms. This is also true for the biological correspondent terms. Biological terms
that correspond to multiple functions or flows are repeated and are italicized to
designate the special case of those terms. This treatment is similar to the repeated
words of the engineering correspondent terms.
4.3.6 Limitations
As with all engineering design tools, limitations exist. For the engineering-to-
biology thesaurus, the limitations are the terms available and the focus on mapping
biological terms to the functions and flows of an existing engineering modeling
language. Setting the boundaries on the small but representative set of engineering
terms of the Functional Basis is a major limitation. This directly affects the design
tool and limits the biological terms that can be correlated with the engineering
domain. For example, the biological function of protect loosely aligns with the
Functional Basis tertiary term prevent, which is under the secondary-level term
stop. This relationship, however, is not intuitive because a shell of an animal
protects the animal from harm or death, but it does not prevent or stop the attack
from occurring. Therefore, there are relevant biological function terms that are
often performed by biological systems that are currently not included. While the
thesaurus does not include a comprehensive list of all biology and engineering
terms, it does, however, contain a representative set that can guide the designer to
make informed judgments on terms that are not included. Depending on the level
of confidence by the designer, this may or may not be achievable. Both the
engineering and biology term sets could be expanded to include other relevant
terms, as they are made available.
Translating biological information to the engineering domain is possible, as
shown in Sect. 4.4.2; however, it is not as intuitive to an engineer as translating
information from one engineering subdomain to another (i.e., fluid resistance to
electrical resistance). Designers must make creative leaps through reliance on prior
knowledge, a knowledge base, a design tool, or communication with a biologist to

perform translation across domains. The current state of the art, which includes the
engineering-to-biology thesaurus, allows engineers to translate a function in nature
to an equivalent, or analogous, engineering function. What the current state of the
art does not do is automatically identify what the biological system performing the
function is or translate the fundamental mechanism of the biological system to the
engineering domain. Therefore, the engineering-to-biology thesaurus requires that
the user has a good understanding of engineering function, which will assist in
making the connections between the domains to correlate biological systems with
engineered systems.
An additional limitation of the design tool is that it is a manual design tool. A
designer must manually look up terms in the thesaurus tables. To assist the manual
process, an alternate form of the thesaurus, organized and alphabetized by bio-
logical term, was created to better assist the manual translation from biology to
engineering. It can be found at www.designengineeringlab.org. Although two
versions of the thesaurus have been compiled, a computational version that would
take input terms from either domain and automatically look up the corresponding
term or, if the term does not exist, suggest a relevant corresponding term would be
beneficial. Development of a biology-to-engineering translator for automatic
machine translation of biological information into an engineering context is
underway.
4.4 Design Tool Applications
Mapping terms between the biology and engineering domains not only reduces the
terminology barrier and addresses other challenges engineers face when working
across the two domains, but also supports several engineering design applications.
Applications of the engineering-to-biology thesaurus include, but are not limited
to, (1) translation of biological information to increase comprehension or develop
connections, (2) concept generation with biological inspiration through function-
based approaches, and (3) dialogue facilitation between the engineering and
biology communities. The concept generation application activities supported by
the thesaurus are the identification of relevant biological terms to use during
brainstorming or searching for biological inspiration, functional modeling of
biological systems, and identification of analogies between the domains. Fig-
ure 4.2 depicts how the applications of the engineering-to-biology thesaurus
design tool fit within a typical function-based engineering design process. Com-
munication and translation applications can occur at many points during the
engineering design process, affording versatility of the design tool. Additional
versatility is provided during the early phases of design through the multiple
concept generation application activities. Integration and prescriptive use of the
engineering-to-biology thesaurus design tool into an engineering design process to
facilitate bioinspired design are described in Chap. 5.
78 J. K. S. Nagel
Fig. 4.2 Thesaurus applications mapped to a typical function-based design process (adapted
from Pahl et al. 2007)
All the design activities lead to, or can be incorporated into, concept devel-
opment, which is an overarching application of this design tool. Additionally, the
majority of these applications capture the biological system through an abstraction,
which is very valuable in solving design problems and can assist designers with
learning design principles from nature. The following subsections describe the
inherent and plausible applications of the design tool. Application examples are
also provided along with references to in-depth examples.
4.4.1 Communication
The design tool presented in this chapter can aid the communication and collab-
oration between engineers and biologists. There are distinct relationships that the
two domains share. For example, function in engineering is analogous to physi-
ology in biology, and structure in engineering is analogous to morphology in
biology. Similar relationships exist among the subject-domain-specific terms of
biology and engineering. Many are captured in the thesaurus tables. Recognizing
those relationships can help an engineer or biologist understand the other’s
viewpoint, which, subsequently, can aid communication across the domains.
Specifically, engineers get glimpse of how biological terms have multiple mean-
ings and levels of interpretation. Being aware of the terminology differences and
levels of abstraction can ease communication with biologists.
The engineering-to-biology thesaurus can assist engineers with asking better
questions of biologists. For example, asking ‘‘how does a biological system
transmit information?’’ is very open-ended and may not lead to a useful answer. A
better question to ask would be ‘‘what role does transduction play in communi-
cating information?’’ By replacing transmit with transduction and rewording, the
question to be more specific is likely to reduce the time clarifying what is meant by
certain terms or phrases, and more time is spent discussing topics of interest,
resulting in an intellectually stimulating conversation versus a frustrating discus-
sion of the differences between the domains. Using the terms of the thesaurus can
similarly aid a biologist to reciprocate and give an answer in a context that an
engineer is more likely to understand.
4.4.2 Translation
Lopez-Huertas wrote that a thesaurus ‘‘… is thought of as a way of easing com-

munication between texts and users in order to increase the interaction in infor-
mation retrieval, and thus facilitate information transfer’’ ( Lopez-Huertas 1997).
The engineering-to-biology thesaurus can aid engineering designers with the
translation process that changes biological information into engineering ‘‘speak’’
to facilitate information transfer. Translation aims to effectively change the context
of the information from biology to engineering. The fundamental activity of
translation is to increase comprehension of biological information. Empirically,
translation has led to a greater understanding of the inspiring biological system and
impacts the time required to develop a bioinspired concept. Consequently,
increased comprehension positively affects other engineering-to-biology thesaurus
applications as well as the bioinspired design process. Other reasons for translation
include exploration of analogies, assistance with concept generation, and com-
munication across the domains.
Translation is achieved by manually substituting biological terms that appear in
the thesaurus with their corresponding Functional Basis terms. Substitution of
terms can be approached incrementally or all at once. Essentially, this will rewrite
the biological information in engineering ‘‘speak’’ and increase the likelihood of a
designer making connections between the two sets of information. Describing a
biological system in engineering terms of the Functional Basis is advantageous.
Not only does it increase the likelihood of a designer understanding the biological
system, but it also lends itself to formulating connections between the biological
and engineering domains, easy comparison to other abstractions, and easy inte-
gration with function-based design methods. Efficient information retrieval
through the engineering-to-biology thesaurus allows an engineering designer to
cross into the biological domain and gain functional knowledge without becoming
overwhelmed by unfamiliar biological systems and terminology.
The following example serves as a qualitative measure of the engineering-to-
biology thesaurus to show that the design tool can assist in translating biological
information into engineering ‘‘speak’’ without requiring the designer to learn deep
80 J. K. S. Nagel
biological knowledge. Consider the text excerpt describing insect olfaction

through antennae taken from the section Chemoreception in The Encyclopedia of
Insects (Mitchell 2003). The contextual difference between the original and
translated forms aims to, ‘‘eas[e] communication between texts and users in order
to increase the interaction in information retrieval, and thus facilitate information
transfer’’ (Nagel et al. 2010c). The original text from Mitchell is as follows.
In insects, odor molecules first contact the cuticular surface, and because it is waxy, they
easily dissolve. From here they move in two dimensions, and some find their way into the
opening of a pore canal. …. Eventually, however, before it arrives at the receptor surface
of a dendrite, the hydrophobic odor molecule will encounter water. …. The other type
binds less specifically [to] a variety of nonpheromone molecules (e.g., food odors) and are
called general odor binding proteins (GOBP). The odorant binding proteins (OBP) act as
shuttles and carry odor molecules through the aqueous medium to the surface of the
dendrite. In the membrane of the sensory cell are receptors for various odors, depending
on the specificity of the cell.
Mappings used to translate the text on insect chemoreception are with function-
and flow-type terms. Many biological terms in the text excerpt were found in the
thesaurus tables and manually swapped with the corresponding engineering terms.
Examples are molecule, protein, receptor, dendrite, shuttles, and cell. While the
biological process of olfaction through antennae described might be clear, the
types of materials and signals involved in the process are less clear. Translation of
the flow-type terms assists in determining the types of materials and signals
involved. Biological terms to translate that were not found within the thesaurus,
such as dendrite, were addressed through translation of the definition of the bio-
logical term. By manually identifying unclear biological terms and substituting
Functional Basis terms, the translated text presents the information in a more
generalized context. The translated insect olfaction text is as follows:
In insects, odor particulates first contact the solid material surface, and because it is waxy,
they easily dissolve. From here the odor particulate moves in two dimensions, and some
find their way into the opening of a pore canal. … Eventually, however, before it arrives at
the solid object surface of a an object that transfers electrical energy, the hydrophobic odor
particulate will encounter a liquid material. … The other type joins less specifically [to] a
variety of nonpheromone particulates (e.g., food odors) and are called general odor joining
solid-liquid mixtures. The odorant joining solid-liquid mixtures act as guides and transfer
odor particulates through the liquid material to the surface of the object that transfers
electrical energy. In the solid-solid material of the sensory solid-liquid material are solid
objects for various odors, depending on the specificity of the solid-liquid material.
The biological information is now presented in a more generalized, engineering

context, which can be used to facilitate a range of bioinspired design activities.
One could further translate the text if the biological system is still unclear. Each
engineering designer will have different background knowledge of biology;
therefore, iterations are recommended until the designer comprehends the bio-
logical information.
Fig. 4.3 Typical concept generation processes utilizing the thesaurus
4.4.3 Concept Generation
Concept generation—manual or computational—aims to generate several con-

ceptual design variants. During this process, engineers engage in many activities
and draw on their prior knowledge or search knowledge bases such as specialized
repositories, design catalogs, patents, and technical literature (Voland 2004; Otto
and Wood 2001; Ulrich and Eppinger 2004; Cross 2008). Nature is another
resource available to engineers during concept generation, and the subject of
biology serves as the body of knowledge engineers can interface with to gain
inspiration. By mapping biological terms to the Functional Basis modeling lan-
guage terms, the engineering-to-biology thesaurus integrates with existing func-
tion-based concept generation activities to facilitate bioinspired design. The
concept generation application activities supported by the thesaurus are identifi-
cation of relevant biological terms to use during brainstorming or searching for
biological inspiration, functional modeling of biological systems, and identifica-
tion of analogies between the domains. Figure 4.3 depicts the process for the
concept generation application activities that fit within the conceptual phase of the
design process given in Fig. 4.2.
The process flow in Fig. 4.3 with a dashed line relies on the use of biological
functional models to derive bioinspired conceptual designs. In this concept gen-
eration approach, the biological system is known or chosen prior to concept
generation. The biological functional model is used to query a knowledge base
indexed by engineering function, such as the Design Repository1 discussed in Sect.
4.5.1, to retrieve engineered components and subsystems that perform the same
functions as the biological system. The MEMIC tool described in Sect. 4.5.3
supports an automated morphological matrix approach that interfaces with the
Design Repository. A manual morphological matrix approach could also be taken
by manually searching patents, design catalogs, and literature for solutions to the
biological functions. Once engineered solutions are determined, the designer must
draw analogies or connections to make the leap to a bioinspired concept.
1
Design Repository www.designengineeringlab.org.
82 J. K. S. Nagel
Fig. 4.4 Relationship of biological functional models to engineering functional models
The process flow in Fig. 4.3 with a solid line relies on the use of engineering
functional models to derive bioinspired conceptual designs. In this concept gen-
eration approach, the inspiring biological systems are not known prior to concept
generation and are identified through searching a biological knowledge base.
Biological knowledge bases that integrate with the thesaurus include the Design
Repository and the biological corpora that are made available to the organized
search tool discussed in Sect. 4.5.2. Once biological solutions are identified, the
designer must learn more about the inspiring systems to draw analogies or con-
nections to make the leap to a bioinspired concept. Translation of biological
information or communication with biologists may be necessary.
Considering biological systems through generalized engineering terms allows
connections to be made between the domains, which facilitates knowledge transfer,
and allows the biological information to be used during the engineering design
process. Prior knowledge of a broad range of engineered systems and processes is
not required for concept generation of bioinspired designs; however, that knowl-
edge provides the impetus for readily recognizing the connections between systems
of two dissimilar domains. Concepts are formulated directly and indirectly from
biological inspiration. Using the thesaurus could result in conceptual designs that
partially (i.e., one or two components) or completely (i.e., entire design) mimic a
biological system. Although the engineering-to-biology thesaurus assists in making
the leap from biology to engineering, to arrive at the final concept, the designer is
still required to make the leap within the engineering domain.
The following subsections describe the concept generation application activities
in detail.
4.4.3.1 Functional Modeling of Biological Systems
The engineering-to-biology thesaurus provides direction when choosing the best-

suited function or flow term to objectively model a biological system. A wide
Fig. 4.5 Biological functional model of chemoreception
range of biological terms have been collected and placed into the thesaurus, which
can accommodate a designer when developing functional models of well-known to
lesser-known biological systems. Functional modeling of biological systems
allows representation of solutions to specific engineering functions and direct
knowledge discovery of the similarities and differences between biological and
engineered systems, as viewed from a functional perspective (Nagel et al. 2010d).
Figure 4.4 summarizes the relationships between biological and engineering
functional models. Detailed instructions and examples of creating biological
functional models can be found in (Nagel et al. 2010b, d, 2011).
A biological functional model could be thought of as an additional form of
translation, which allows a designer to qualitatively model a system to compre-
hend and compare functionality. Representing biological, functionally using the
lexicon of the Functional Basis, also allows biological solutions to be stored in an
engineering design knowledge base for future reuse, such as for concept generation
or for educational purposes. The archived biological solutions can then be recalled
and adapted to engineered systems. The creation of engineered systems that
implement strategies or principles of their biological counterparts without repro-
ducing physical biological entities is an additional benefit to biological functional
models.
Continuing the example from Sect. 4.4.2, a biological functional model of
insect olfaction through antennae is developed and provided in Fig. 4.5. The
information contained within the prior text excerpt is further expanded below to
assist in understanding details of insect olfaction or chemoreception.
Antennae are made of a chitin–protein complex referred to as cuticle, which are
porous, and covered in a waxy layer to prevent desiccation (Mitchell 2003).
Multiple parts of the insect body, particularly the antennae, are covered in
84 J. K. S. Nagel
cuticular protrusions in the form of sensilla (e.g., hairs, pegs) that house the
chemically sensitive cells for olfaction (Mitchell 2003; Møller 2003; Eguchi and
Tominaga 1999; Klowden 2008). In order to detect the chemical stimulus, the odor
molecules must make contact with the waxy layer of a sensillum and travel
through the porous cuticle. Once inside, odor molecules encounter an aqueous
medium containing odor-binding proteins and receptor sites on the dendrite sur-
face (Mitchell 2003; Møller 2003; Eguchi and Tominaga 1999). As the name
implies, the odor-binding proteins bind to the odor molecules and essentially
shuttle one odor molecule at a time to a receptor site. The dendrite is connected to
a sensory cell that, in most cases, is activated by specific odor types (e.g., food,
pheromones) through the receptor sites at the dendrite surface. Regardless, once an
odor molecule comes into contact with a receptor site, a signal is generated, the
signal is amplified, and the odor-binding protein then causes hydrolysis to separate
the odor molecule from the receptor site and the protein itself (Mitchell 2003;
Møller 2003; Eguchi and Tominaga 1999). The odor-binding protein is responsible
and required for receptor site activation and deactivation (Klowden 2008). Binding
to a receptor site causes activation and conformational change and leads to the
generation of an action potential (electrical signal), which is summarized as signal
transduction. This is achieved through second messengers, typically cyclic aden-
osine monophosphate, which increases the sensory cell’s permeability to sodium
ions and alters the electrical potential of the cell membrane (Mitchell 2003; Møller
2003; Eguchi and Tominaga 1999; Klowden 2008). After the signal has been
generated and separation by hydrolysis is complete, esterase enzymes breakdown
the odor molecule and the odor-binding protein is recycled.
Using the terms within Tables 4.1, 4.2, 4.3 and 4.4, the biological information
is mapped to engineering terms, similar to the translation process described in
Sect. 4.4.2. Where the translation example of Sect. 4.4.2 had an emphasis on
biological flows, the biological functional model has an emphasis on biological
functions. Combined, the two create a comprehensive translation of biological
information into an engineering context. The following paragraph demonstrates
the distilled biological information with the substituted engineering terms in italics
used to create the biological functional model in Fig. 4.5.
When the chemical energy, odorant, enters the insect cuticle, the odor-binding
proteins immediately sense their presence and begin the detection process. The
function of join represents the protein binding to the chemical stimulus, which is
then carried to the receptor site noted by the function of transport. The couple
function denotes binding of the odor molecule and odorant-binding protein to the
receptor site. Change represents the activation, conformational change in the
receptor site, and generation of an action potential and is why the flows of
chemical energy and mixture materials are all present for that function. Signifying
the receptor site deactivation in parallel with the electrical signal that is sent to the
nervous system to be identified is the function of separate and actuate, respec-
tively. The final portion of the chemoreception process is transmission of the
electrical signal to the brain to produce a response.
Following the process outlined in Fig. 4.3, the biological functional model of
chemoreception could be used to inspire an innovative chemical sensor. Addi-
tionally, the model could be entered into a specialized knowledge base for future
bioinspired design reuse. An example of modeling the symbiotic biological system
lichen with the help of the thesaurus can be found in (Nagel et al. 2010b, 2011).
The lichen biological functional model resulted in a concept for an innovative
solar energy system that is adaptable to different climates to improve efficiency of
electricity generation.
4.4.3.2 Inspiration Brainstorming and Search
Identification or discovery of relevant biological systems to take inspiration from

when addressing an engineering problem can be a challenge. The biological terms
of the engineering-to-biology thesaurus can assist in discovering biological sys-
tems to use for design inspiration by offering more functions for brainstorming and
relevant keywords to use when searching for inspiration. It is often advantageous
to reframe the design problem to relate it to biology (Helms et al. 2009). Searching
a biology corpus, such as a textbook, for biological inspiration based on engi-
neering terms typically produces results that are mixed. Results containing the
search word often use the search word out of context, not at all or in a different
sense than the designer intended. By utilizing the biological correspondent terms
of the thesaurus when searching for specific functions or flows that solve the
engineering problem, search results improve (Nagel and Stone 2012) and become
more focused on the desired biological systems, consequently, aiding concept
generation.
Inspiration brainstorming and searching integrates with the concept generation
processes depicted in Fig. 4.3 and can be applied with manual or computational
function-based methods. Two applications include identification of biological
systems that address customer needs and identification of a biological system for
functional modeling.
Applying the thesaurus during brainstorming biological solutions or searching for
biological systems that offer solutions to engineering problems can assist in
achieving relevant results. For example, consider the design of a braking system for a
pedaled vehicle. A simple way to relate the problem to biology is to consider ‘‘How
do biological systems brake?’’ This approach, however, will not yield results that
relate to the given design problem. To increase relevant results, the designer can
develop a black box or detailed functional model to determine engineering functions
that closely relate to the braking system design. One result is the Functional Basis
engineering term stop. Looking for the engineering term stop in the engineering-to-
biology thesaurus (Table 4.1) leads to the biological correspondent terms of
extinguish, halt, clog, seal, and suspend. Using the identified set of biological cor-
respondent function terms provides a wide range of terms to assist in brainstorming
or searching to identify biological systems that address the design problem.
86 J. K. S. Nagel
Identified biological systems include spider webs, fainting goats, giraffe throat
valves, blood clots, scabs, clam shells, and puffer fish.
With several biological systems identified, a designer, depending on the level of
understanding, can begin to develop bioinspired conceptual designs. If compre-
hension is low, the designer could look up information on each system to perform
translation or functional modeling or analogy discovery to become inspired.
Considering fainting goats and puffer fish leads to two interesting braking system
concept variants. Learning that fainting goats’ muscles seize when they become
over-excited leads to a concept of suspending bike movement to prevent input
from the rider and slow down. Essentially, the entire bike acts as the braking
system. Learning that puffer fish expand to halt predators leads to a concept that
has tires that act as brakes by expanding to create more friction and stop the bike
from moving.
4.4.3.3 Discovery of Analogies
The established terminology relationships of the engineering-to-biology thesaurus

provide a foundation for discovering analogies between the engineering and
biology domains. Abstractions of biological functions, processes, materials, sig-
nals, and energies, as presented in the engineering-to-biology thesaurus, allow a
designer to make connections with engineering abstractions, principles, compo-
nents, and systems. Connections can be formulated directly and indirectly from
biological inspiration. Thus, connections and analogies do not need to be one to
one, as in the case of directly copying a biological system. Discovering analogies
between biology and engineering can happen at any time during exploration of a
biological system for inspiration; however, after the translation of biological
information into an engineering context, the connection may become clearer.
Considering biological systems in a generalized engineering context allows
connections to be made between the domains, which facilitates analogy discovery.
For example, consider the southern three-banded armadillo, which has three bands
along its back that spread and allow it to roll up into a ball and cover itself from
predators in the event of an attack. The armor surrounds the animal’s tail, head,
feet, and back. Similar to the development of biological functional models, the
biological information is manually mapped to engineering terms using the the-
saurus tables to assist the discovery of analogies. Looking for the biological terms
of surround, cover, and spread in the thesaurus tables leads to the corresponding
engineering terms of inhibit and shape. Inhibit is a tertiary term under stop, and
shape is a tertiary term under change, with both falling under the class of control
magnitude. Using the identified engineering terminology, the designer can identify
analogous engineered systems to the armadillo that involve in changing shape to
inhibit an external input. Engineered systems that perform the functions corre-
sponding to surround, cover, and spread are stadiums with retractable roofs and
convertible automobiles. Both engineered systems change shape and inhibit the
‘‘enemy’’ of bad weather.
Querying the Design Repository facilitates a computational approach to dis-

covering analogies. With both biological and engineered systems in the knowledge
base, one could retrieve biological and engineered components when searching for
solutions, thus allowing for comparison and analogy discovery through function.
Both the manual and computational approaches to analogy discovery can lead to
an ‘‘Aha’’ moment that assists in concept generation.
4.5 Integration with Computational Design Tools
Computational design tools promise engineers a faster realization of potential

design solutions based on previously known products and implementations, or
databases of design information. The engineering-to-biology thesaurus design tool,
which is a stand-alone tool, has been integrated with established function-based
computational design tools to support bioinspired engineering design. The fol-
lowing subsections describe the integration efforts.
4.5.1 Design Repository
The Design Repository is an engineering design knowledge base that contains

descriptive product information such as functionality, component physical
parameters, manufacturing processes, failure, and component compatibility of over
130 consumer products. Each consumer product was decomposed and functionally
modeled using the Functional Basis. Each repository entry is designated as an
artifact or assembly of artifacts, whether it performs a supporting function (sec-
ondary to the product’s operation) and the class of the artifact when entered into
the repository database. Additionally, several artifact attributes are captured and
stored in a relational database.
The integration of the thesaurus with the Design Repository has been with the
population of a biomimetic design repository, which enables the storage of bio-
logical knowledge indexed by engineering function. Storing the biological infor-
mation based on the function the biological system solves allows quick access to
principle solutions. There are a total of 30 biological entries in the Design
Repository that comprise the biomimetic design repository. The Design Repository
facilitates computational concept generation and comparison of biological and
engineered components. The designer chooses from resulting computational
concept generator suggestions, engineered and biological, to develop a complete
conceptual design.
88 J. K. S. Nagel
4.5.2 Organized Search Tool
The organized search tool (Stroble et al. 2009d) was developed for retrieving
relevant biological systems that perform functions of interest. Specifically, the
organized search tool is designed to work with non-engineering subject-domain-
specific information, such as biology. The majority of biological information is
written in natural language format, which prompted the investigation of using both
a Functional Basis function and flow term when searching for solutions. Realizing
how the topic of the text is treated increases the extensibility of the organized
verb–noun search algorithm.
The verb–noun combination search strategy incorporates the terms of the
engineering-to-biology thesaurus into the search algorithm and provides two levels
of results: (1) associated with verb only, of which the user can choose to utilize or
ignore, and (2) the narrowed results associated with the verb–noun pair. This
search strategy requires the designer to first form an abstraction (e.g., functional
model) of the unsolved problem using the Functional Basis lexicon. The verbs
(functions) of the abstraction are input as keywords in the organized search tool to
generate a list of biological matches. The search algorithm swaps the engineering
function term for the corresponding biological function terms in the engineering-
to-biology thesaurus. The biological corpora are then searched for the biological
function, and all sentences containing the function are extracted for further pro-
cessing. Each match is stored for display to the user. When multiple biological
correspondent function terms are present, the search is executed recursively until
all corresponding biological functions have been searched. The noun listing is then
used in combination with the search verb results for a second, more detailed search
to locate specific text excerpts from the biological corpora that describe how the
biological systems perform the functionality with the desired flows.
This search strategy is embodied in an automated retrieval tool that allows an
engineering designer to selectively choose which corpora to search and to upload
additional searchable information as it is made available. The user interface ini-
tially presents the designer with a function (verb) entry field and search options.
Search options prompt the designer to choose from exact word, derivatives of the
word, and partial word. Once the biological corpora are searched for the function
term, the designer is presented with a listing of flows (nouns) that occur in
proximity to the searched verb for each corpus searched followed by a group of
sentences that include the function and listed flows. The resultant biological
information is more relevant and focused due to the integration of the thesaurus in
the search algorithm.
4.5.3 MEMIC
Computational concept generation is an efficient way to generate several con-

ceptual design variants. Also, it adds the benefit of providing lists of engineering
components that may be used to solve a particular function. The morphological
evaluation machine and interactive conceptualizer (MEMIC) was created for use
during the early stages of design to produce design solutions for an engineering
design from a functional model using knowledge of existing engineered products
(Bryant Arnold et al. 2008; Bryant et al. 2007). The concept generator software
MEMIC accepts an input functional model and uses functionality and compati-
bility information stored in the Design Repository to generate, filter, and rank full
concept variants. The software returns a listing of engineering component solu-
tions for each function–flow pair of the input functional model, which allows a
designer to easily choose between multiple solutions for a given function and
interactively build a complete conceptual design.
To extend the MEMIC computational concept generation approach to support
bioinspired design, the organized search tool algorithm, which utilizes the engi-
neering-to-biology thesaurus, is merged (Nagel and Stone 2012). Integrating
biological information with an established, computational method for concept
generation enables designers to consider taking inspiration from biology without
having to expend extra effort to learn a new method. Computational concept
generation of bioinspired designs through this approach requires the designer to
input desired functionality, in the form of a functional model. Functionality is a
useful metric for defining a conceptual idea, as functional representation has been
shown to reduce fixation of how a product or device would look and operate (Otto
and Wood 2001; Pahl et al. 2007). Each function/flow pair of the model is then
searched in the engineering and biological knowledge bases, the Design Reposi-
tory and the biological corpora made available to the organized search tool,
respectively, to identify solutions to the function/flow pairs.
This computational approach expands the biological knowledge base beyond
that of the biomimetic design repository to provide a greater range of biological
solutions for engineering inspiration. Multiple solutions from both domains, to
each function/flow pair, are returned and presented to the designer. The biological
solutions are not indented for physical use, but are intended for spurring creative
ideas or connections to the engineering domain that could be implemented in an
engineered system that partially (i.e., one or two components) to completely (i.e.,
entire design) mimic a biological system. This computational approach assists in
identifying biological solutions to engineering functions, analogies, and engi-
neering components that map to biological system attributes. To arrive at the final
concept, however, the designer is required to identify principles, components,
materials, and/or systems within the engineering domain that support what the
biological solution suggests. Therefore, this approach lends itself more toward
innovative design problems where novel solutions tend to dominate.
90 J. K. S. Nagel
4.6 Conclusion
The engineering-to-biology thesaurus presented in this chapter affords engineers,

with limited biological background, a tool for leveraging nature’s ingenuity during
many steps of the design process. An engineering-to-biology thesaurus (1) lessens
the burden when working with knowledge from the biological domain by pro-
viding a link between engineering and biological terminologies; (2) assists
designers in establishing connections between the two domains; and (3) facilitates
bioinspired design through many applications. The thesaurus is presented as a
versatile design tool that can be used during many points in the bioinspired design
process and is customizable to the designer based on needs and prior biological
knowledge.
Biological terms in the thesaurus are correlated with the engineering domain
through pairing with a synonymous function or flow term of the Functional Basis
lexicon. The engineering-to-biology thesaurus is a work in progress and is not a
comprehensive list of all biological and engineering terms. Rather than encom-
passing all engineering terminology, the confined set of generalized engineering
terms of the Functional Basis was chosen as an initial starting point. Through this
research, biological correspondent function and flow (material, signal, and energy)
terms were mapped to the engineering domain. The mapping groups related terms
and concepts in a classification structure that supports bioinspired design activities
in the engineering domain. Additionally, the terminology mappings of the engi-
neering-to-biology thesaurus facilitate recognition of related terms and concepts
between engineering and biology by providing a structured framework. The
engineering-to-biology thesaurus increases the interaction between the users and
the knowledge resource and fosters one to make associations between the engi-
neering and biological lexicons, thus strengthening the designer’s ability to utilize
biological information.
Listing biological correspondent terms that an engineering designer interested
in function-based design might encounter is among the first step to bridging the
terminology gap between the biology and engineering domains. Furthermore, the
engineering-to-biology thesaurus is a subject-domain-oriented, intermediary
structure, which can be updated as needs are identified.
The work to develop and formulate the engineering-to-biology design tool
makes a fundamental contribution to the fields of engineering design and bio-
mimicry. The thesaurus is envisioned to enable the engineering and biology
communities to better collaborate, create, and discover. Furthermore, it facilitates
many bioinspired design activities as a stand-alone tool, but also supports and
enables computational bioinspired design activities.
Future work for improving the engineering-to-biology thesaurus includes
examining potential terms through clustering and analyzing terms contained within
the glossary of a collegiate entry-level biological textbook. While collocated terms
provide an indication for macrorelevant terms, clustering analysis could be utilized
to find less obvious, but equally important, biological terms for thesaurus
population. Additionally, biological texts that focus on a topic of interest (i.e.,

insects, fungi) should be analyzed for relevant biological terms that an introduc-
tory text does not include. Further work also includes adopting a hierarchy for the
mapped biological terms. Many of the repeated biological flows currently in the
thesaurus indicate that the biological flow has different scales. For example, an
organ is a composite of tissues, and tissue is a composite of cells. A hierarchy
could help make the terminology mappings between the domains more clear.
Beyond the terms within the thesaurus, there are future work opportunities for
the implementation of the tool as well. The terms have successfully been inte-
grated with other computational design tools; however, a stand-alone computa-
tional version of the thesaurus would still be useful. Development has begun on
biology-to-engineering translator for automatic machine translation of biological
information into an engineering context. A long-term goal is implementation of the
thesaurus as a dynamic Web entity that would allow collaboration where those
with different expertise can contribute to the terms as well as support dissemi-
nation of the research in real time.
References
Bar-Cohen Y (2006a) Biomimetics biologically inspired technologies. CRC/Taylor & Francis,

Boca Raton
Bar-Cohen Y (2006b) Biomimetics—using nature to inspire human innovation. J Bioinspiration
Biomimetics 1:P1–P12
Benyus JM (1997) Biomimicry innovation inspired by nature. Morrow, New York
Brebbia CA (2006) Design and nature III: comparing design in nature with science and
engineering, vol 3. WIT, Southampton
Brebbia CA (2008) Design & nature IV: comparing design in nature with science and
engineering. WIT, Southampton
Brebbia CA, Carpi A (2010) Design & nature V: comparing design in nature with science and
engineering. WIT, Southampton
Brebbia CA, Collins MW (2004) Design and nature II: comparing design in nature with science
and engineering, vol 3. WIT, Southampton
Brebbia CA, Sucharov LJ, Pascolo P (2002) Design and nature: comparing design in nature with
science and engineering, vol 1. WIT, Southampton
Brownell P (2010a) Engineering-to-biology thesaurus term check meeting. Nagel JKS. 8 Apr
2010, Corvallis
Brownell P (2010b) Engineering-to-biology thesaurus term check and validation meeting. Nagel
JKS. 4 May 2010, Corvallis
Bruck HA, Gershon AL, Golden I, Gupta SK, Gyger LS Jr, Magrab EB, Spranklin BW (2007)
Training mechanical engineering students to utilize biological inspiration during product
development. Bioinspiration Biomimetics 2:S198–S209
Bryant Arnold CR, Stone RB, McAdams DA (2008) MEMIC: an interactive morphological
matrix tool for automated concept generation. In: Mason JFaS (ed) Industrial engineering
research conference
Bryant C, Bohm M, McAdams D, Stone R (2007) An interactive morphological matrix
computational design tool: a hybrid of two methods. Paper presented at the ASME 2007
IDETC/CIE, Las Vegas, NV
92 J. K. S. Nagel
Campbell NA, Reece JB (2003) Biology. Pearson Benjamin Cummings, San Francisco
Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju BS (2005) A functional representation for
aiding biomimetic and artificial inspiration of new ideas. Artif Intell Eng Des Anal Manuf
19:113–132
Cheong H, Shu LH, Stone RB, McAdams DA (2008) Translating terms of the functional basis
into biologically meaningful words. In: 2008 ASME IDETC/CIE, New York City, NY
Chiu I, Shu LH (2007a) Biomimetic design through natural language analysis to facilitate cross-
domain information retrieval. AIEDAM 21(1):45–59
Chiu I, Shu LH (2007b) Using language as related stimuli for concept generation. AIEDAM
21(2):103–121
Cross N (2008) Engineering design methods: strategies for product design. Wiley, Chichester
de Mestral G (1955) Velvet type fabric and method of producing same. USA Patent
Design & Intelligence Laboratory (2010) Biologue. http://home.cc.gatech.edu/dil/336. Accessed
10 Dec 2012
Dieter GE, Schmidt LC (2009) Engineering design, 4th edn. McGraw-Hill, New York
Dym CL, Little P (2004) Engineering design: a project-based introduction. Wiley, New York
Eguchi E, Tominaga Y (1999) Atlas of arthropod sensory receptors: dynamic morphology in
relation to function. Springer, Tokyo
Forbes P (2006) The gecko’s foot: bio-inspiration: engineering new materials from nature. W. W.
Norton & Co, New York
Gero JS, Kannengiesser U (2002) The situated function—behaviour—structure framework. Artif
Intell Des, pp 89–104
Goel AK, Chandrasekaran B (1992) Case-based design: a task analysis. Artif Intell Approaches
Eng Des 2:165–184
Goel AK, Rugaber S, Vattam S (2009) Structure, behavior, and function of complex systems: the
structure, behavior, and function modeling language. Artif Intell Eng Des Anal Manuf
23(1):23–35
Helms M, Vattam SS, Goel AK (2009) Biologically inspired design: products and processes. Des
Stud 30(5):606–622
Hey J, Linsey J, Agogino AM, Wood KL (2008) Analogies and metaphors in creative design. Int
J Eng Educ 24(2):283–294
Hill B (1995) Bionic—element for fixing the aim and finding the solution in the technical
problem solving process. In: IDATER 1995 conference, Loughborough University, Lough-
borough. http://hdl.handle.net/2134/1509
Hirtz J, Stone R, McAdams D, Szykman S, Wood K (2002) A functional basis for engineering
design: reconciling and evolving previous efforts. Res Eng Design 13(2):65–82
Hundal M (1990) A systematic method for developing function structures, solutions and concept
variants. Mech Mach Theory 25(3):243–256
Hyman B (1998) Engineering design. Prentice-Hall, New Jersey
Klowden MJ (2008) Physiological systems in insects. Academic Press, Oxford
Lawrence E, Holmes S (1989) Henderson’s dictionary of biological terms. Wiley, New York
Lindemann U, Gramann J (2004) Engineering design using biological principles. In: International
design conference—DESIGN 2004, Dubrovnik
Linsey F (2008a) Biological flow correspondent term check meeting. Stroble JK. 20 June 2008,
Rolla
Linsey F (2008b) Biological flow correspondent term check and validation meeting. Stroble JK.
10 July 2008, Rolla
Little A, Wood K, McAdams D (1997) Functional analysis: a fundamental empirical study for
reverse engineering, benchmarking and redesign. In: 1997 ASME IDETC/CIE, Sacramento, CA
Lopez-Huertas MJ (1997) Thesaurus structure design: a conceptual approach for improved
interaction. J Documentation 53(2):139–177
531(1):117–120
Mak TW, Shu LH (2008) Using descriptions of biological phenomena for idea generation. Res
Matrin E, Hine RS (2000) Oxford dictionary of biology. Oxford University Press, Oxford
McKean E (2005) The new Oxford American dictionary. Oxford University Press, New York
Mitchell BK (2003) Chemoreception. Encyclopedia of insects. Academic Press, Amsterdam,
pp 169–174
Møller AR (2003) Sensory systems: anatomy and physiology. Academic Press, Amsterdam
Nachtigall W (1989) Konstructionen: Biologie und Technik. VDI, Dùsseldorf
Nachtigall W (2000) The big book of bionics: new technologies, following the example of nature.
German Verlags-Anstalt
Nachtigall W (2002) Bionics: principles and examples for engineers and scientists. 2nd edn.
Springer
Nagel JKS, Stone RB (2012) A computational approach to biologically-inspired design.
AIEDAM 26(2):0
Nagel R, Tinsley A, Midha P, McAdams D, Stone R, Shu L (2008) Exploring the use of
functional models in biomimetic design. J Mech Des 130(12):11–23
Nagel JKS, Stone RB, McAdams DA (2010a) Exploring the use of category and scale to scope a
biological functional model. In: 2010 ASME IDETC/CIE, Montreal, Quebec, Canada
Nagel JKS, Nagel RL, Stone RB, McAdams DA (2010b) Function-based, biologically inspired
concept generation. AIEDAM 24(4):521–535
Nagel JKS, Stone RB, McAdams DA (2010c) An engineering-to-biology thesaurus for
engineering design. In: ASME IDETC/CIE 2010 DTM-28233, Montreal, Quebec, Canada
Nagel JKS, Stone RB, McAdams DA (2010d) Exploring the use of category and scale to scope a
biological functional model. In: ASME IDETC/CIE 2010, DTM-28873, Montreal, Quebec,
Canada
Nagel JKS, Nagel RL, Stone RB (2011) Abstracting biology in engineering design. Int J Des Eng
4(1):23–40
Otto KN, Wood KL (2001) product design: techniques in reverse engineering and new product
development. Prentice-Hall, Upper Saddle River
Pahl G, Beitz W, Feldhusen J, Grote KH (2007) Engineering design: a systematic approach, 3rd
edn. Springer
Sarkar P, Phaneendra S, Chakrabarti A (2008) Developing engineering products using inspiration
from nature. J Comput Inf Sci Eng 8(3):1–9
Shu LH, Stone RB, McAdams DA, Greer JL (2007) Integrating function-based and biomimetic
design for automatic concept generation. In: International conference on engineering design,
Paris, France
Srinivasan V, Chakrabarti A (2009) SAPPhIRE—an approach to analysis and synthesis. In:
International conference on engineering design, Stanford, USA
Srinivasan V, Chakrabarti A (2009) SAPPhIRE—an approach to analysis and synthesis. In: 3rd
Symposium on research in product design, CPDM, IISc, Bangalore, India. doi:2-417
Stone R, Wood K (2000) Development of a functional basis for design. J Mech Des
122(4):359–370
Stroble JK, Stone RB, Watkins SE (2009a) An overview of biomimetic sensor technology. Sens
Rev 28(2):112–119
Stroble JK, Stone RB, McAdams DA (2009b) Conceptualization of biomimetic sensors through
functional representation of natural sensing solutions. In: International conference of
engineering design, Stanford, California
Stroble JK, Stone RB, McAdams DA, Watkins SE (2009c) An engineering-to-biology thesaurus
to promote better collaboration, creativity and discovery. In: CIRP design conference 2009,
Cranfield, Bedfordshire, UK, pp 353–368
Stroble JK, Stone RB, McAdams DA, Goeke MS, Watkins SE (2009d) Automated retrieval of
non-engineering domain solutions to engineering problems. In: CIRP design conference 2009,
Cranfield, Bedfordshire, UK, pp 78–85
94 J. K. S. Nagel
Szykman S, Racz J, Sriram R (1999) The representation of function in computer-based design. In:
Proceedings of the ASME design theory and methodology conference, Las Vegas, NV
The Biomimicry Institute (2009) Biomimicry: a tool for innovation. http://www.bio
mimicryinstitute.org/about-us/biomimicry-a-tool-for-innovation.html. Accessed 10 Oct 2009
Toko K (2000) Biomimetic sensor technology. Cambridge University Press, Cambridge
Tsujimoto K, Miura S, Tsumaya A, Nagai Y, Chakrabarti A, Taura T (2008) A method for
creative behavioral design based on analogy and blending from natural things. In: 2008
ASME IDETC/CIE, New York, USA. DETC2008-49389
Ullman DG (2009) The mechanical design process, 4th edn. McGraw-Hill, Inc., New York
Ulrich KT, Eppinger SD (2004) Product design and development. McGraw-Hill/Irwin, Boston
Umeda Y, Takeda H, Tomiyama T, Yoshikawa H (1990) Function, behaviour, and structure.
AIENG’90 Applications of AI in Engineering, pp 177–193
Vakili V, Shu LH (2001) Towards biomimetic concept generation. In: 2001 ASME IDETC/CIE,
Pittsburgh, Pennsylvania
Vakili V, Shu LH (2007) Including functional models of biological phenomena as design stimuli.
In: 2007 ASME IDETC/CIE, Las Vegas, NV
Vattam S, Helms M, Goel A (2010a) A content account of creative analogies in biologically
inspired design. AIEDAM 24:467–481
Vattam S, Wiltgen B, Helms M, Goel A, Yen J (2010b) DANE: fostering creativity in and
through biologically inspired design. In: Proceedings of first international conference on
design creativity, Kobe, Japan, pp 127–132
Vincent JFV, Bogatyreva OA, Bogatyrev NR, Bowyer A, Pahl A-K (2006) Biomimetics: its
practice and theory. J R Soc Interface 3:471–482
Voland G (2004) Engineering by design, 2nd edn. Pearson Prentice Hall, Upper Saddle River
Wilson JO, Rosen D (2007) Systematic reverse engineering of biological systems. In: 2007
ASME IDETC/CIE, Las Vegas, Nevada
Wilson J, Chang P, Yim S, Rosen D (2009) Developing a bio-inspired design repository using
ontologies. In: 2009 ASME IDETC/CIE, California, USA
Chapter 5
Function-Based Biologically Inspired
Design
Jacquelyn K. S. Nagel, Robert B. Stone and Daniel A. McAdams
Abstract A ‘‘big picture’’ approach to a systematic, function-based (drawing

from a Pahl and Beitz approach) biologically inspired design is presented in this
chapter. The approach supports two different starting, or perhaps motivating,
points: a customer need motivated product design and a biological system moti-
vated product opportunity. Both approaches rely on a designer’s ability to create a
functional model that either captures customer needs or represents the biological
system of interest. This methodology relies directly on the designer’s ability to
make connections between dissimilar domain information. Following presentation
of the methodology are two validation approaches. One examines current bio-
logically inspired products either in production or presented in the literature to
demonstrate that the systematic design methodology for biologically inspired
design can reproduce the existing design. The second validation exercise inves-
tigates three needs–based design problems that lead to plausible biologically
inspired solutions.
Keywords Design methodology

Framework
Function-based design

Systematic design Design tools Functional modeling Connections Analogies

Sensor design Problem-driven approach Biology-driven approach Concept
development
J. K. S. Nagel (&)
Department of Engineering, James Madison University, Harrisonburg, VA, USA
e-mail: nageljk@jmu.edu
R. B. Stone
School of Mechanical Industrial and Manufacturing Engineering, Oregon State University,
Corvallis, OR, USA
e-mail: robert.stone@oregon.state.edu
D. A. McAdams
Department of Mechanical Engineering, Texas A&M University, College Station, TX, USA
e-mail: dmcadams@tamu.edu

DOI: 10.1007/978-1-4471-5248-4_5, Ó Springer-Verlag London 2014
96 J. K. S. Nagel et al.
5.1 Introduction
Biological organisms, phenomena, and strategies, herein referred to as biological

systems, provide a rich set of examples that can be used to inspire engineering
innovation. Bioinspired, or biomimetic, designs are publicly viewed as creative
and novel solutions to human problems. Moreover, some biomimetic designs have
become so commonplace that it is hard to imagine life without them (e.g., Velcro,
airplanes). Although the bioinspired solutions are innovative and useful, the
majority of inspiration taken from nature has happened by chance observation or
dedicated study of a specific biological entity (e.g., gecko). This reveals a fun-
damental problem of working across the engineering and biological domains. The
effort and time required to become a competent engineering designer creates
significant obstacles to becoming sufficiently knowledgeable about biological
systems (the converse can also be said). This research aims to go beyond the
element of chance, reduce the amount of time and effort required to develop
bioinspired solutions, and bridge the seemingly immense disconnect between the
engineering and biological domains.
This chapter presents a comprehensive design approach, including a method-
ology and supporting tools, that integrates with Pahl- and Beitz-based (Pahl et al.
2007) or Otto- and Wood-based (Otto and Wood 2001) functional design tech-
niques to facilitate biologically inspired design. As a function-based method, it
offers several advantages: archival and transmittal of design information; reduces
fixation on esthetic features or a particular physical solution; allows one to define
the scope or boundary of the design problem as broad or narrow as necessary; and
encourages one to draw upon experience and knowledge stored in a database or
through creative methods during concept generation. Function, as used in sys-
tematic design, is recognized as a way to connect nature and engineering through a
commonality. This approach addresses the knowledge requirement problem of
working across the biology and engineering domains and, through functional
modeling, offers the aforementioned advantages.
A design methodology that facilitates systematic biologically inspired design
from a problem-driven or biology-driven perspective is given. Following a tradi-
tional design approach, the problem-driven perspective begins from customer
needs or a given problem to solve. Following curiosity, the biology-driven
approach encourages a designer to explore a biological system of interest to learn
more about it or develop an innovative solution that could potentially be applied to
an existing problem. The methodology also prescribes the tools of the framework
to use while guiding the designer through the steps. Although systematic, the
design methodology is also versatile by providing a designer multiple avenues that
lead to a biologically inspired design.
Biologically inspired design is a young field within engineering design. Con-
sequently, validation of the design methodology is challenging. Thus, two
approaches to validation are pursued. One examines current biologically inspired
products either in production or presented in the literature to demonstrate that the
5 Function-Based Biologically Inspired Design 97
systematic design methodology for biologically inspired design can reproduce the
existing design. The second validation exercise investigates three needs–based
design case studies that lead to plausible biologically inspired solutions. The
needs-based design case studies demonstrate the systematic design approach to
biologically inspired design from the problem-driven perspective. A reference to a
case study from the biology-driven perspective is also provided. The thought
processes, model iterations, and connections that are the leaps that enable the
ingenuity of the nature to be discovered and adapted for use in engineered systems
are also presented.
Using functional representation and abstraction to describe biological systems
presents the natural designs in an engineering context and allows designers to
make connections between biological and engineered systems. Thus, the biological
information is accessible to engineering designers with varying biological
knowledge, but a common understanding of engineering design methodologies. By
creating a bridge between the two domains through the perspective of function,
engineers can leverage the elegant designs found in the world around them. This
work demonstrates the feasibility of using systematic design for the discovery of
innovative engineering designs without requiring expert-level knowledge, but
rather broad knowledge of many fields.
5.2 Framework to Support Method
Various tools and techniques combine or interact to support the overall systematic
biologically inspired design methodology. The engineering-to-biology thesaurus
(see Chap. 4) is the backbone of this framework, as it assists in modeling bio-
logical systems and searching for inspiration or solutions. Consequently, the
thesaurus also assists in concept generation, both directly and indirectly. Indirect
assistance is through the modeling method and organized search tool, and direct
assistance is through designer knowledge of a biological process (e.g., the con-
version of sunlight to sugars) that could solve a set of design needs of a product.
The following section makes this indirect assistance more explicit.
A framework, by definition, is an arrangement of parts that provides a system or
concept a basic form (McKean 2005). In the framework developed here, the parts
are identify, translate, represent, and conceptualize. The system or concept being
supported is the systematic biologically inspired design methodology. What makes
this framework particularly useful for design is the flexibility a designer is afforded
when working toward a biologically inspired solution. Each tool and technique can
be used individually and in multiple combinations. Prior work has shown how the
engineering-to-biology thesaurus integrates with and improves organized search
tools (Nagel et al. 2010; Nagel et al. 2010; Nagel et al. 2010), supports biological
functional modeling method (Nagel et al. 2010; Nagel et al. 2010; Nagel et al.
2010), and integrates with concept generation approaches (Nagel 2010). Specific
interaction benefits are pointed out in the Venn diagram of Fig. 5.1. Using all four
parts results in systematic biologically inspired design.
Fig. 5.1 Framework details
Table 5.1 Framework integration with existing design tools

Framework tool or technique Existing tool leveraged by framework
Engineering-to-biology thesaurus Functional basis
Biological functional modeling method Functional basis
Organized search tool Functional basis, design repository
Concept generation approaches MEMIC, automated morphological matrix,
design repository
Knowledge requirements for working between domains can be alleviated with

design methods that integrate the existing engineering design tools. This frame-
work alleviates knowledge requirements by integrating the function-based design
methodologies, in particular, the systematic design process defined by Pahl et al.
(2007). The framework also leverages existing design tools to further integrate
with function-based design methodologies, as shown in Table 5.1. The existing
design tools include a functional modeling lexicon, a repository of design infor-
mation, and automated concept generation methods.
5.3 Biologically Inspired Design Methodology
In this section, the overall design methodology is given. In support of the sys-
tematic biologically inspired design methodology, several separate techniques
coalesce to guide the designer from initial curiosity or customer needs to complete
concept (Nagel and Stone 2011). The framework for making connections between
biology and engineering is presented first followed by the design methodology.
Two holistic approaches by the Biomimicry Institute (2010) and Helms et al.
(2009) offer a process for designers to follow, from a biology perspective and an
engineering perspective, respectively, and a database of biological information that
supports biologically inspired design. Additionally, the work by Helms et al.
(2009) and Vattam et al. (2008) occurred in parallel with this work unbeknownst to
the authors at the time. Both bodies of research made several similar observations
that guided their work along distinct, but related paths. For example, Helms et al.
(2009) first formally identified the two approaches of problem-based and solution-
based bioinspired design that emerged from an earlier study of design students
utilizing biological inspiration to solve engineering design problems (Vattam et al.
2008). The authors of this chapter made a similar observation while exploring how
to use functional models as a means of finding related biological inspiration for
design (Nagel et al. 2008) and identified two approaches to handle both problem-
based and solution-based starting points for biologically inspired design. These
approaches, however, do not offer a complete framework of design tools that
integrate with the functional model-driven design process to support the designer
throughout biologically inspired design. This chapter addresses that gap by
offering a systematic methodology for both problem- and solution-based entry
points along with a framework of supporting design tools that integrate with
established functional model-based, engineering design methodologies.
The aim of the systematic design methodology of this chapter is to provide
enough structure to assist in the bioinspired design process without hindering the
creativity and inventiveness of the designer. This method is intended to foster and
guide the abilities of the designer and encourage objective evaluation of the
results. The systematic approaches of this method are further intended to render
designing based on biological inspiration comprehensible and steer the efforts of
designers down purposeful paths.
5.3.1 Making Connections
This research relies on the designer’s ability to identify and formulate connections
between biological and engineering domains. Just as there are different learning
styles, there are multiple ways to make connections. Analogies (Casakin 2006;
Gentner 1988; Gentner 1983; Gick and Holyoak 1980; Goel 1997; Hofstadter
1995; Tsujimoto et al. 2008; Linsey et al. 2008; Balazs and Brown 2001; Mak and
Shu 2004; Bhatta and Goel 1997; Smith 1998; Nagai and Taura 2006) are the most
widely used and have multiple forms. Direct, indirect, and compound analogies
have all been used to connect a biological system to an engineering solution. A
direct analogy mimics the biological system one to one. An indirect analogy uses
the biological system to spur analogies for inspiration but does not mimic every
aspect of the biological system. A compound analogy is the combination of
multiple biological system attributes that lead to analogous engineered systems.

The level of difficulty in accessing and transferring an analogy is largely dependent
on how remote or close the distance between the domains is Johnson-Laird (1989).
Because much of nature exhibits functionality and behavior in a comparable
context to engineering, analogies with engineering are possible. To exemplify the
connection-making process for analogies, consider a few textual examples.
A microflow detection sensor directly mimics the physiology and morphology
of hair cells that make up the lateral line system in a fish. The connection is
through the principle of a bending moment that is created from perpendicular flow
against vertical hair cells. Direct analogy mimicry is achieved through fabrication
of ‘‘hair-like’’ vertical structures on the end of a horizontal cantilever beam (Fan
et al. 2002; Motamed and Yan 2005).
An indirect analogy bioinspired design was created between the common strain
gage and the physiology of the campaniform sensillum or flexible exocuticle that
many insects possess. An elliptical opening in the insect’s cuticle, which is cov-
ered by a thin membrane layer, senses deformation because of the stress con-
centration (Gnatzy et al. 1987; Grunert and Gnatzy 1987). The connection for this
system is that the opening causes mechanical coupling and global amplification to
occur. Mimicry is achieved by optically measuring the stress concentration at a
circular or elliptical hole in a rigid material when pressure is applied, resulting in a
novel sensor that can sense strain in all directions (360°) (Wicaksono et al. 2004).
In the case of designing an electronic display that can be viewed in bright
sunlight, a bioinspired compound analogy was used to solve the problem. For the
display problem, hummingbird feather and morpho-butterfly wing attributes were
combined to develop a solution (Vattam et al. 2008). Hummingbird feathers
contain a series of alternating layers of thin films with different thicknesses instead
of the intricate christmas-tree-like structures within an air gap that butterfly wings
possess. The connections here are the air gap and ‘‘thin-film-like’’ structures,
which are readily used in electronics processing today. Adding an air gap between
thin films of varying thicknesses provided the right inspiration to develop the
BrightView project (Vattam et al. 2008).
A designer must also be aware of analogies that hurt the design or ones that are
overly complicated. Consider the biological phenomenon of abscission. When a
leaf of a plant is damaged, it stops the flow of auxin and allows abscisic acid to
dominate, thus forming a seal around the base of the leaf stem and over a period of
time the leaf falls off (Campbell and Reece 2003). This biological system was used
to inspire a solution to the problem of tiny parts sticking to a robot gripper in a
microassembly process (Shu et al. 2006).
Considering indirect analogy for this case allowed the researchers to develop a
sacrificial tool assembly. To highlight the analogy, consider the gripper as anal-
ogous to the plant, the sacrificial part of the tool as analogous to the abscission
zone, and the tiny screw as analogous to the leaf that is released. Separation is
achieved through the breakdown of the sacrificial part of the tool (abscission
zone). Notice, liquids that are analogous to auxin and abscisic acid are not present
in the final design.
Developing a direct analogy of abscission would require a flowing chemical

that secures and releases the tiny screw. Release of the tiny screw would occur
some time after the chemical flow stops and a chemical reaction takes place to
loosen the part from the gripper. The indirect analogy that disregards liquids is the
preferred solution concept. Therefore, another analogy form should be considered
if the results lead to a bad design.
Two other approaches to formulating connections are through first principles
(Hubka and Eder 1984; Lindemann and Gramann 2004; Otto and Wood 2001;
Vincent and Mann 2002) and metaphors (Casakin 2007, 2006; Hey et al. 2008;
Forty 1989). Analysis of physiology, structure, or behavior can lead to a con-
nection made through first principles. Physical laws and concepts, such as the
conservation of energy, that govern science as we know it also apply to natural
systems.
Identifying a first principle shared by both domains leads to a connection and
possibly innovation. Consider how ducks and other birds regulate their tempera-
ture during the winter to stay alive. The principle of heat exchange between the
body and the legs is carried out to reduce the amount of heat lost through blood
that is circulated through the legs (Lindemann and Gramann 2004). To date,
metaphors for biologically inspired design have only been documented for
architectural structures (Dollens 2009). The multiple approaches to formulating
connections allow a designer to discover and become inspired in a manner that
best suits him or her.
5.3.2 The Methodology
A pictorial representation of the method is given in Fig. 5.2. The flower is used to
show that the methodology is an organic process that has systematic design roots.
Each of the steps is discussed in greater detail below. The majority of, if not all,
design processes are iterative, and this methodology follows the same convention.
Cues for when to iterate are provided. Furthermore, the design methodology here
should not be viewed as a rigid sequence in which one must follow each minute
detailed step. Rather, it should be viewed as a starting point or a set of guidelines
that aim to arrive at a biologically inspired design.
Figure 5.3 summarizes the avenues of the problem-driven approach that closely
follows traditional systematic design, and the avenues of the problem-driven
approach that starts from a known biological solution are summarized in Fig. 5.4.
Figure 5.5 is a flowchart of the biology-driven approach that starts from curiosity.
Also, the parts of the framework and existing design tools used in each step of the
method are made explicit in Table 5.2.
Step 1: Needs or Curiosity

Initially, a designer can choose to start from a traditional set of customer needs or
explore a curiosity. These two approaches are identified as problem-driven and
Fig. 5.2 An overview of the five steps of the systematic biologically inspired design
methodology
biology-driven, respectively. Taking the problem-driven approach means the

designer must gather a set of needs, requirements, and constrains. Many sources
exist to aid the designer with proper needs gathering (Otto and Wood 2001;
Fig. 5.3 Flowchart of the problem-driven approach that closely follows traditional systematic
design (a) basic systematic approach to biologically inspired design using a function-based
framework (b) systematic approach to biologically inspired design using a function-based
framework with biological system identification guided by search heuristics
Fig. 5.4 Flowchart of the problem-driven approach that starts from a known biological solution
(a) basic systematic approach to biologically inspired design using a function-based framework
when an inspiring biological system is already identified (b) systematic approach to biologically
inspired design using a function-based framework when translation of a biological process into
functional language is needed
Fig. 5.5 Flowchart of the

biology-driven approach
Table 5.2 Framework parts and existing design tools used in each step of the methodology
Tools Steps Needs or Decompose Query Make Create
curiosity connections concepts
Framework E2B thesaurus X X X X
Search tool X X
Biological functional X X X X
model
Biological concept X X
generation
Existing Design repository X X X
Functional basis X X X
MEMIC X
Automated X
morphological
matrix
Ullman 2009; Ulrich and Eppinger 2004; Voland 2004; Hyman 1998; Dym and
Little 2004). Identifying customer needs is the most critical part of the design
process as they form the basis for device functionality and specifications. Taking
the biology-driven approach means a designer already knows of an interesting
biological system in which he or she would like to investigate. Thus, the focus is
on finding an application for the principles and concepts exhibited in that bio-
logical system rather than on finding principles to solve an engineering need.
However, the organized search tool can be used to find a biological system to
investigate initially.
Step 2: Decompose
The second step involves decomposing the needs or interesting biological system
into, first, a black box model and, second, a functional model. All models created
with this method use the Functional Basis modeling lexicon. With regard to tra-
ditional systematic design, the black box aims to abstract the overall function of the
device that is to be designed. Whereas for a biological system, the black box model
describes an interesting function, structure, behavior, or strategy of the system.
Next, the input and output flows to the black box are determined. These flows are
prompted by the customer needs from the first step or the needs/attributes of the
biological system needed to achieve the black box functionality. The next task is to
create the functional model. Decompose the black box description into subfunc-
tions connected by flows of energy, material, or signal (Pahl et al. 2007; Otto and
Wood 2001; Stone 1997; Stone and Wood 2000; Nagel 2010). Functional model
creation is often an iterative task. Before moving on to the next step, check to see
that all customer needs have been met by identifying the flows and subfunction
chains that address them (Pahl et al. 2007; Otto and Wood 2001; Stone 1997; Stone
and Wood 2000; Nagel 2010). When following the biology-driven approach, the
designer can refer to the general biological modeling methodology presented in
(Nagel et al. 2010) for assistance with creating a biological functional model.
Step 3: Query
Step 3 involves querying a knowledge base to identify solutions to each function/
flow pair of the functional model. Two knowledge bases are required: one con-
taining successful engineered systems and the other containing biological systems.
To integrate with this method, both are indexed by engineering function and flow.
The Design Repository1 containing descriptive product information serves as the
engineered systems’ body of knowledge. It also includes product information such
as functionality, component parameters, manufacturing processes, failure, and
component connectivity. The Design Repository now contains design knowledge
of over 113 consumer products and 30 biological systems. Instead of creating a
large knowledge base containing functionally decomposed biological systems,
similar to the design repository, an introductory biology textbook serves as the
biological systems’ body of knowledge. Although it is not indexed by engineering
function, the engineering-to-biology thesaurus provides a starting point to find
inspiration with engineering function and flow.
The tasks that comprise Step 3 begin with using the MEMIC software, or
automated morphological matrix search tool, to query the Design Repository and
the organized search software to query the biological corpus.
Based on the number of results returned for engineered and biological solutions,
the search may need to be repeated. For engineered solutions, it can be helpful to
abstract terms to the next level in the hierarchy. For example, if transport, a
tertiary-level term, does not return any repository entries, then the secondary-level
term transfer should be used. The same approach applies to flows.
Step 4: Make Connections

Step 4 involves making connections. Connections through analogies, metaphors,
and first principles assist in bridging the biology and engineering domains. Sec-
tion 5.3.1 contains a thorough discussion of how connections are formulated.
Step 5: Concept Generation

The fifth step involves performing concept generation and creating biologically
inspired conceptual solutions. Concept synthesis involves analysis, reflection, and
synthesis. Analysis is on the returned engineered and biological solutions from
Step 3. Reflection is on the connections to the engineering domain formulated in
Step 4. Synthesis is of the existing engineering solutions, engineering solutions
inspired by biology and inventive solutions inspired by biology to derive a new
idea. Once synthesis takes place, the result will be at least one concept. Depending
on the number of solutions returned during Step 3 and the connections made
during Step 4, multiple concepts may result. Evaluation of concepts follows sys-
tematic design; Pugh charts or decision matrices are used to rank concepts and
narrow the selection to the superior few concepts (Otto and Wood 2001; Ullman
2009; Ulrich and Eppinger 2004; Voland 2004; Hyman 1998; Dym and Little 2004).
1
www.designengineeringlab.org
Once a final concept has been reached, the next phase of systematic design,
Detailed Design (Pahl et al. 2007), can initiate.
5.3.3 Comparison of Methodology Approaches
Notice that the major differences between the avenues to biologically inspired
design as shown in Figs. 5.3 and 5.4 are within Step 2. When a designer follows
traditional systematic design, flowchart A of Fig. 5.3, and decomposes a functional
model from a black box model, the resultant model is referred to as a conceptual
functional model. The conceptual functional model is used to query the engi-
neering and biology knowledge bases. In the event that no connections can be
formalized, then the designer should return to the query step and try different
levels of functions and flows. The same holds true for when no concepts are
synthesized.
Following flowchart B of Fig. 5.3 instructs the designer to use the organized
search tool heuristics for the initial query. Once biological solutions are gathered,
then the Design Repository is queried to supplement the biological solutions with
engineering solutions. One difference here is that the results of the search tool
should be screened first before moving on. In the event that no connections can be
formalized, then the designer should return to the query step and try a different
heuristic. The same holds true for when no concepts are synthesized.
If a designer knows of a biological system that can solve the black box func-
tionality, then a biological functional model of this system can be created to drive
the methodology as shown in flowchart A of Fig. 5.4. In the event that no con-
nections can be formalized, then the designer should return to the decompose step
and either modify the biological functional model or choose a different biological
system for exploration. The same holds true for when no concepts are synthesized.
Consider the scenario when a designer can describe a known biological process
that solves the black box functionality, as shown in flowchart B of Fig. 5.4. Using
the biological process as a starting point, the designer can search for biological
systems that perform that process using the organized search tool. If the search
results are too narrow or uninspiring, the designer can use the engineering-to-
biology thesaurus to identify new query terms. Once a biological system is chosen,
then the designer can define the biological system with a functional model and use
it to discover engineered solutions that perform the same functions to make
connections back to the engineering domain, unless connections are readily
facilitated. In the event that no connections can be formalized, then the designer
should return to Step 2 and either choose a different biological system for
exploration or, if no model was created, develop a biological functional. The same
holds true for when no concepts are synthesized. In both process flows of Fig. 5.4,
if the Design Repository results are limited, then the designer should try different
levels of functions and flows to broaden the search.
The fifth approach under this methodology, termed the biology-driven

approach, is driven by curiosity and the possibility of creating an innovative
solution based on the inspiring biological system that may or may not be fit to a
problem. The main use for this approach, however, is the creation of archivable
biological knowledge that is in an engineering context. This type of knowledge can
be reused in future design activities when stored in a knowledge base. The main
tasks of this approach, as shown in Fig. 5.5, follow the steps of the systematic
biologically inspired design method. In the event that no connections can be
formalized, then the designer should return to the decompose step and modify the
biological functional. The same holds true for when no concepts are synthesized.
Another option is to return to Step 1, choose a different biological system for
exploration, and start over. Also, if the Design Repository results are meager or
non-existent, then the designer should try different levels of functions and flows to
broaden the search.
5.4 Validation of Method
Validation of the systematic biologically inspired design methodology is achieved

through the application of the methodology to (1) check whether it reproduces
existing biomimetic products and (2) identify a closely related development ver-
sion of a concept through the literature review. Analysis and reproduction of
existing biomimetic products through primary function allow the verification of
the methodology as the result is known. Validating the methodology for non-
existing biomimetic products requires a review of literature to quantify whether
the concept variants are realistic in some near-term form. Finding a closely related
development version of the biologically inspired conceptual design indicates that
the concept is feasible. Similarity is based on functionality and components chosen
to achieve functionality.
Six existing biomimetic products are analyzed through the application of the
method to demonstrate that the method can reproduce what is known. To further
demonstrate the validity of this method, three needs–based design problems are
presented. The methodology is considered successful when a designer can analyze
a biological system and identify connections between biology and engineering
through function that lead to inspiration of a concept. Further, more detailed
validation cases of how the methodology can result in innovative solutions fol-
lowing the problem-driven approach are given in (Nagel 2010), and how the
methodology can result in an innovative solution following the biology-driven
approach is given in Nagel et al. (2010), Nagel (2010).
5.4.1 Proof Through Existing Biomimetic Technology
The first validation exercise is to analyze existing biomimetic products, apply the
systematic design methodology, and verify that the biological system used to
Table 5.3 Analysis of existing biomimetic products to validate methodology

Existing Mimicked Primary function/flow Source Can
biomimetic biological pair(s) method
products system reproduce
design?
Walking stick for Echolocation Detect solid Biological Yes
visually of bats corpus
impaired that
uses sonar
Passive heating and Termite Regulate thermal energy, Asknature.org Yes
cooling mounds distribute thermal energy,
buildings distribute gas, remove gas
Self-cleaning Lotus Inhibit solid, inhibit liquid, Asknature.org Yes
surfaces decrease solid
Motion detector Compound Detect solid, Sense solid Biological Yes
vision corpus
Color changing Morpho- Change visual signal Asknature.org Yes
material without butterfly
harmful
chemicals
Microassembly Abscission of Separate solid Biological Yes
with sacrificial plants corpus
gripper
inspire the original design is used in the results in such a way that could lead to a
reproduction.
Table 5.3 lists six existing biomimetic products. These technologies represent
the fields of electrical, civil, and mechanical engineering and material science. The
validation studies here relied on technical descriptions of these products found in
the literature. From the descriptions, primary function/flow pairs were identified
and represented with Functional Basis terminology. The primary function/flow
pairs are then used to query the biological knowledge base.
Both representation and querying utilize the engineering-to-biology thesaurus.
If the mimicked biological system is within the query results and described in a
way that makes a connection and results in a similar concept to the existing
biomimetic technology, then it follows that the method can reproduce the design.
During this validation exercise, AskNature was added to the biological corpus
knowledge base to augment topical limitations in the corpus. AskNature2 is an
online database that biologists, engineers, designers, chemists, etc., can contribute
focused on biologically inspired design.
Following the five steps of the methodology, all six existing biomimetic
technologies were reproduced. The three found within the organized search tool
required the substitution of biological function and flow terms of the thesaurus,
while the three found within the AskNature database needed substitution of only
2
www.asknature.org
the flow term. Additionally, other biological systems were identified that also solve
the function/flow pair, which, if a redesign was undertaken, could result in com-
pound analogical design.
5.4.2 Smart Flooring Case Study
Validation through exploration of non-existing biomimetic products is performed

here on a smart flooring application. Consider the following scenario. A customer
wants to create a security/surveillance product that looks like ordinary carpet,
mats, rugs, etc., to detect intruders, a presence of something in a room or move-
ment. Requirements for the smart flooring include a detection mechanism unseen
by human eye, durability, composition of common materials, and a quick detection
response. Also, the system needs to be autonomous.
The new design should offer advantages over current systems. Current tagged-
based systems require the user to carry a badge or other device to be tracked or
monitored, and simply removing the trackable item can defeat the system. Radar
or similar systems require calibration and an area map to be created. Each time the
area layout is changed, the map needs to be updated. Video surveillance and heat
signature systems can be very expensive and often require a person to watch the
real-time video feed.
These needs and constraints are mapped to flows as shown in Table 5.4 to
complete the first step of the methodology. Next, the black box model is created to
guide the decomposition of the flows into a functional model. Figure 5.6 provides
the black box model. Figure 5.7 provides the conceptual functional model. The
model of Fig. 5.7 was created with a boundary of the flooring in place, electrical
energy is supplied to the detection mechanism, and when an object or human
interacts with the flooring, a signal is generated.
Table 5.4 Needs of smart Needs/Constraints Functional basis flow

flooring device mapped to
flows Object/human to detect Solid material
Quick detection response Status signal
System power Electrical energy
Fig. 5.6 Smart flooring

black box model
Fig. 5.7 Smart flooring conceptual functional model
From here the search heuristics were applied, following flowchart B of Fig. 5.3.
Using the general inspiration heuristic, several interesting biological systems were
found to perform the function of detect. An organized search of a biological corpus
resulted in the following query results (Purves et al. 2001):
1. The hair cell
2. Electroreceptors found in electric fish
3. Epithelial cells
4. Genes that mark recombinant DNA
5. DNA
6. Why birds flock in large groups
7. Echolocation
8. Carotid and aortic stretch receptors
9. Membrane receptor proteins
10. Graded action potentials.
Next, following Step 4 of the methodology, the biological systems returned for
the function of detect are analyzed to formalize connections.
Hair cells are analogous to cantilevers and would detect a presence when dis-
turbed. In a similar manner, the carotid and aortic stretch receptors are analogous
to flexible materials such as polymers. A polymer would detect a disturbance when
pressure is applied.
Echolocation is analogous to radar. Radar is already used to detect objects;
however, it is not a distributed system, as would be needed for a smart flooring
concept. The final connection made from the above list is with the electroreceptor
of fish. Electroreceptors generate an electric field for navigation of the environ-
ment, to locate objects, which is also analogous to radar. Echolocation uses sound
waves where electrolocation uses electric waves.
Now that the query results have been analyzed and reflected upon to get to
establish connections between biology and engineering, and the next task, fol-
lowing flowchart B of Fig. 5.3, is to query the engineering knowledge base and
supplement the biologically inspired solutions with engineering solutions to
complete the design. The returned engineering solutions, shown in Table 5.5, are
Table 5.5 Engineering solutions mapped to function/flow pairs

Function/Flow Engineering solution
Import/electrical energy Battery, circuit board, electric motor, electric wire, electric
switch
Regulate/electrical energy Actuation lever, capacitor, circuit board, automobile distributor,
electric switch, heating element, transistor, transformer,
thermostat, regulator, volume knob
Transfer/electrical energy Battery, circuit board, electric wire, electric motor, electric
socket, electric plate, electric switch, heating element, USB
cable, light fixture, speaker
Transmit/electrical energy Electrical wire, battery contacts, motor controller
Detect/solid material Read head, line guide
Indicate/status signal Light, tube, displacement gauge, LCD screen
Export/electrical energy Circuit board, electric wire, electric switch
then synthesized. Functional representation enables a thorough understanding of

the requirements while decreasing the tendency of designers to fixate on some
particular physical solution for a problem. Therefore, this approach is form
independent and relies on the designer’s ability to develop a reasonable form for
the concept. With the components of Table 5.5 and the established analogies, the
final step of concept generation can begin.
The biological system of the hair cell prompts two concept variants. Both
concept variants use wires and circuit boards to perform the functions involving
electrical energy and would likely be powered by an external, low-voltage power
supply. Regardless of the concept, a flexible circuitry layer and buffer layer would
need to be underneath the visible flooring layer to connect the array to a processing
unit and to protect the underlying circuitry, respectively. These are represented in
the concept sketch close-up views as the orange layer between two black layers.
Wanting to remotely monitor a space means that the solution to indicate status
signal must also be remotely located to the smart flooring. For both concept
variants, the LCD screen is a viable option as the detection mechanism will
produce a measurable result, which can be displayed on the LCD screen. The
concept variants below focus on the biologically inspired detection mechanism;
therefore, the power supply and LCD screen are not shown.
Recall that the critical need is unseen by the human eye. Considering the tactile
response of the hair cell as a cantilever and flooring shaped as individual tiles, each
tile could act as one cantilever to detect a load. The array of cantilever load sensors
would then sense a pressure differential as a person walks across the smart
flooring. A spring is added to the design to keep the tile from collapsing while
allowing the cantilever to bend. A concept sketch is shown in Fig. 5.8. Detail
design would need to be completed to determine the cantilever material that would
best respond and last in a high-traffic environment.
The second concept variant simultaneously exploits the tactile response of the
hair cell and the carotid and aortic stretch receptors and the change in electric
Fig. 5.8 Concept variant one for smart flooring
potential from the sensory cells. Hair cells are vertical structures, while stretch
receptors are found in multiple orientations. Taking inspiration from the hair cell
and stretch receptor morphology leads to a detector design that is comprised of a
vertical structure that can be stretched in multiple orientations.
Offering flexibility and ruggedness for repeated deflection, this detector design
could work for woven flooring such as carpet. Taking inspiration from the sensory
cells, flexion of the carpet fibers would result in a change in resistivity, similar to a
strain gage, or generate a voltage by the principle of piezoelectricity. Polyamide is
a high-performance synthetic polymer and is commonly used in textiles. Fabri-
cating polyamide tubes with a conductive gel or paste that can be woven into
carpet to form an array would achieve the biologically inspired design. Materials
research would need to be completed to determine whether the polyamide and
conductive gel or paste would last in a high-traffic environment.
Alternatively, conductive thread, another detector solution that can be stretched
in multiple orientations, could easily be woven into carpet and offer a change in
resistivity. Conductive thread exists and is used in garments and accessories that
merge technology into clothing. Materials research would need to be completed to
Fig. 5.9 Concept variant two for smart flooring

determine whether the conductive thread would provide a significant change in

resistively or conductivity once woven into a carpet. A concept sketch is shown in
Fig. 5.9. Shaded strands in the close-up view of the carpet fibers in Fig. 5.9 rep-
resent durable feedback elements woven into the carpet, providing hidden sensing
capabilities.
Looking into the literature for a similar surveillance device uncovered a handful
of attempts to create a ‘‘smart’’ floor or flooring. The first concept described above,
individual tiles that detect a load placed in an array, has been done. Richardson
et al. have developed hexagonal, puzzle-like pieces called nodes that are placed in
an array to detect pressure (Richardson et al. 2004). Their tiles contain force-
sensitive resistors and interlock to form a self-organizing network that passes data
to the tile with an external data connection. An earlier approach to the smart floor
involved only one measuring tile made of load cells, a steel plate, and data
acquisition hardware and was not intended to be hidden (Orr and Abowd 2000).
Rather, it was created as an alternative to biometric identification by recognizing a
person’s unique footstep profile.
Two other approaches that utilize load cells and layered flooring to conceal the
sensors are nearly identical to the first concept variant. Liau et al. place a sensor in
the center of every 60960 cm wood-covered tile (Liau et al. 2008), where Ad-
dlesee et al. place a sensor at each intersection of four carpet-covered tiles (Ad-
dlesee et al. 1997). Load cells are similar in principle to cantilever beams in that
deflection is transduced into an electrical signal that can be interpreted. Here, a
literature review revealed that the first biologically inspired concept is feasible and
has been attempted.
Investigating the second concept variant for smart flooring revealed only one
existing design that is similar. Researchers at Infineon Technologies have woven
conductive fibers into carpet and attached them to tiny sensor modules inlaid into
the fabric to build a mesh network (IEE-Institution of Electrical Engineers 2003).
The flooring can report where a person is located, which way they are moving, and
if a sensor module has failed. Each conductor in the design is a copper wire coated
with silver to prevent corrosion and then covered with polyester (IEE-Institution of
Electrical Engineers 2003). A German textile company, Vorwerk, has teamed up
with Infineon to develop the smart carpet (Vorwerk and TGCK 2004; Crane 2005).
The Vorwerk/Infineon product is similar in structure to the second concept variant
created here in that a conductor is concealed and woven into a textile product.
Again, a literature review revealed that the biologically inspired concept is
feasible.
This case study demonstrated that it is possible to systematically design using
the search heuristics and take inspiration from biology in the process. By analyzing
the biological system and making connections, a designer can become inspired. It
is through these correlations that the designer can recognize existing designs that
are similar or develop an innovative design.
Table 5.6 Chemical sensing needs mapped to conceptual flows

Customer need/Constraint Functional basis flow
Selectivity Status signal
Response time Electrical energy
Reusable Sensing layer (material)
Indirect sensing mechanism Sensing layer (material)/Chemical stimulus (energy)
Fig. 5.10 Chemical sensor

black box model
5.4.3 Chemical Sensor Case Study
The systematic biologically inspired design method presented here is able to

inspire more than mechanical and electromechanical devices. This case study
presents the design of a chemical sensor following the approach closest to tradi-
tional systematic design (flowchart A of Fig. 5.3).
For this study, the needs and constraints of the chemical sensing device are
derived from the Handbook of Modern Sensors (Fraden 2004): selectivity (only
senses the desired chemical in the presence of other species), quick response time,
reusable, and utilizes an indirect sensing mechanism. These needs and constraints
are mapped to flows as shown in Table 5.6 to complete the first step of the
methodology. Next, the black box model is created to guide the decomposition of
the flows into a functional model. Figure 5.10 provides the black box model.
Figure 5.11 provides the conceptual functional model to complete Step 2 of the
methodology.
The chemical sensing device black box and conceptual functional models show
the generalized form of a chemical stimulus (i.e., chemical energy). This allows
Fig. 5.11 Chemical sensor conceptual functional model

Table 5.7 Design repository solutions mapped to function/flow pairs

Function/Flow Solutions from design repository
Import/material Housing, reservoir, spring
Import/chemical energy Container, nozzle
Guide/chemical energy Tube
Couple/chemical energy to Basket, container, iron, nozzle, carburetor, burner, housing
material
Import/electrical energy Battery, wire, circuit board, motor, cord, switch
Transmit/electrical energy Wire, battery contacts, circuit board, compound eye
Change/material Blade, impeller, heating element, punch, filter, staple plate,
popcorn popper
Detect/chemical energy Fly chemoreceptor, protein, Animalia chemoreception, Plantae
chemoreception
Regulate/electrical energy Circuit board, actuator, heating element, switch, resistor, diode
Convert/status to control Circuit board
signal
Export/chemical energy Nozzle, bowl, tube, exhaust, bucket
Export/electrical energy Wire, circuit board, cord, switch
Export/status signal LCD screen, circuit board, wire, cord, level, speaker
the designer to query all possible forms of a chemical stimulus. The device sub-
strate is also generalized as material to include all possible forms of material in the
knowledge base. Figure 5.11 demonstrates the indirect sensing mechanism with
couple and change, and the sensing element or transducer with detect. Electrical
energy is utilized to power the sensor and transfer the detection status signal to the
device capable of interpreting such signals, such as a computer. The boundary of
the conceptual functional model includes the sensing layer and powered sensing
element.
Following Step 3, the Design Repository was queried for engineering and
biological entries. The components for each function/flow pair returned are shown
in Table 5.7. For 10 of the 13 sensor functions, the component list was short and
easy to choose from. The functions of change, detect, and export signal returned
many possible components.
Considering the conceptual device as a whole and how one would use the
device is an advantageous thought process for determining suitable components
from a list. With regard to changing the material as the first step in the indirect
sensing mechanism, the impeller, blade, and punch require mechanical movement
to change a material, whereas the staple plate, filter, and heating element could
change the chemical stimulus without mechanical movement. The conceptual
design is not fully determined at this point and could be influenced by the com-
ponent(s) chosen from biological inspiration. Therefore, the functions of change
and detect will be considered together.
Following Step 4 of the methodology, connections need to be established to
assist concept generation. Bacteria employ the two-component regulatory system
for detection of extracellular signals, and the signaling pathways consist of
modular units called transmitters and receivers, both of which are proteins (Stock
et al. 2000). Fly antennae contain chemically sensitive cells (chemoreceptors)
hidden deep within pores, which allow the insect to experience olfaction (i.e.,
sense of smell) (Mitchell 2003). The Animalia and Plantae mechanisms of che-
moreception are not descriptive and are meant to guide a designer in a direction of
research. Analysis of the biological components leads to the choice of fly che-
moreceptor for the detect chemical energy function block.
Further exploration of the fly antennae reveals that the insect cuticle (a chitin-
protein outer cover) has elaborations in the form of trichoids (hairs), pegs, pegs in
pits and flat surfaces, all of which provide multiple pores for chemicals to travel
through (Mitchell 2003). Within the pore is a fluid-protein pathway to the dendritic
(sensory) cell membrane. Once the chemical molecule reaches the fluid sur-
rounding the dendrite membrane, it bonds to an odorant-binding protein and is
carried to one of the receptor sites of the membrane (Mitchell 2003). When the two
make contact in the cation-concentrated fluid, a signal occurs as a voltage potential
change across the membrane, which is the signal to be transduced. The sensing
principles of fly antennae are complex and offer the designer inspirations for the
function of detect and, as expected, for the function of change (Nagel et al. 2010).
A filter is analogous to the porous cuticle, which would narrow down the
selection of chemicals or allow only one stimulus to interact with the sensing layer.
The heating element is analogous to the odorant-binding proteins and cell mem-
brane surface with receptor sites, in that an electrified element is capable of
attracting polarized molecules (disregarding the heating aspect). An impeller could
be used to steer the desired chemical stimulus to the sensing element, which is
analogous to the odorant-binding proteins that shuttle stimuli to receptors.
Biological inspiration considering morphology leads to a sensing element that
has specifically shaped cavities or is uniformly porous, and is a good conductor.
Any material that can be patterned by photolithography can achieve the desired
surface. Morphology of the fly antennae itself offers inspiration for a ‘‘stick-like’’
sensing element. Another connection exists between the engineering component
filter and permeable or ion-selective membranes, which are used in current sensor
technology. Therefore, permeable or ion-selective membranes are analogous to the
porous cuticle. Further analogies exist between the heating element and electrical
energy traveling through a conductor, which could be a copper wire, semicon-
ductor, conducting polymer, etc. Semiconductor electrodes that allow absorption
and desorption of chemical species are analogous to binding and removal of
odorants from the cell membrane receptor sites within the fly antennae, which
could also be used in the final concept.
Performing Step 5, concept generation leads to two concept variants. The first
device supports a housing containing an electrified element (not for the production
of heat) acting as a barrier to the transducer that chemical energy is guided to from
the container, or space. The subsequent chemical energy is attracted to the elec-
trified element, and once bonding occurs, the electrical properties of the electrified
element change and generate a signal to be transduced as shown in Fig. 5.12. The
electrical property change in the electrified element fulfills the requirement of an
Fig. 5.12 Concept variant

one for chemical sensor
Fig. 5.13 Concept variant

two for chemical sensor
indirect sensing mechanism, which also supports reusability as the absorbed par-
ticles could be removed by heating the element. An electronic circuit powers the
transducer, decodes the sensor signal, and produces an electrical signal analogous
to the input. The second device supports a housing containing a filter covering the
sensing layer, which rests on the sensing element. Only the chemical species that
pass through the filter interact with the sensing element. This interaction generates
a signal to be transduced as shown in Fig. 5.13, but does not fulfill the requirement
of an indirect sensing mechanism. The chemical stimuli, although filtered, still are
allowed to interact with the detection layer. An electronic circuit powers the
transducer, decodes the sensor signal, and produces an electrical signal analogous
to the input. Further material research is needed to accurately define the sensing
layers for both concept variants.
When researching whether a sensor with an electrified element on a transducer
exists, the first concept described above, conducting polymer sensors that are
typically used for detection of gases, was found. The conducting polymer is
deposited atop interdigital electrodes, which make it the dielectric material for the
electrodes. When the gas interacts with the polymer, the capacitance between the
electrodes changes and the output signal decreases (György 2008; Bai and Shi
2007). These sensors are sometimes referred to as chemiresistors. Functionally, a
chemiresistor is identical to the first concept.
The second concept is similar to an ion-selective electrode (ISE) or ion-
selective field effect transistor (ISFET) used to measure pH. An ISE has high
specificity to single-charged ions and is made of a doped glass. The glass allows
only single-charged ions, such as hydrogen, to pass through to an internal solution
of neutral pH monitored by an electrode. Concentration of hydrogen ions is cor-
related with a pH value by taking the logarithm of the concentration (Grundler
2007; Eggins 2002). An ISFET is the microelectronic version of an ISE and is
similar in structure to an MOSFET (Grundler 2007; Eggins 2002; Liao et al. 1999).
Instead of glass, the membrane is made as a thin film over an insulation layer of
metal oxide or nitrate on a p-type doped substrate. Two n-type doped regions are
added to the substrate for connection between the source and drain, while the gate
is connected to the sensing layer beneath the membrane. Here, the sensing layer is
a liquid or insulation material and the sensing element is an electrode. Function-
ally, an ISE is identical to the second concept.
This case study demonstrated that it is possible to systematically design a
sensor and take inspiration from biology in the process. By analyzing the bio-
logical system and making connections, a designer can become inspired. It is
through these correlations that the designer can recognize existing designs that are
similar or develop an innovative design.
5.4.4 Heat Exchanger Case Study
This case study presents the design of a heat exchanging component for use in
future hydrogen vehicles following the approach of a known process before
querying is performed. Needs for the heat exchanger design are set by an external
research group that wanted to explore biologically inspired concepts for distrib-
uting hydrogen on the microscale. The external group had exhausted all their
avenues and expressed a need for more ideas, which satisfies Step 1 of the
methodology. Through discussion, a common understanding of distributing
hydrogen on the microscale and the role heat exchange would play in that process
was reached. From here, the black box model, as shown in Fig. 5.14, was created.
A known biological process that distributes material is circulation of blood by the
vascular system of the human body. This is a known biological process that is used
Fig. 5.14 Black box model

of a heat exchanger
to begin the biologically inspired design process. When querying the biological
knowledge base for circulation, many results lead to descriptions of the heart, but
also mention gas exchange organs. This preliminary search did not result in a wide
variety of inspiring biological systems. Therefore, to widen the inspiration space,
the biological terms that correspond to the engineering function of distribute were
retrieved from the engineering-to-biology thesaurus.
In the engineering-to-biology thesaurus, the following seven biological terms
correspond to the engineering term of distribute: circulate, diffusion, exchange,
disperse, scatter, spread, and spray. Circulate resulted in a narrow set of results.
The list does reveal that the term exchange is used in biological literature. With a
new keyword chosen, Step 3 of the methodology can be performed. The biological
knowledge base was queried to identify biological systems that perform the
function of exchange. The resulting query results from Purves et al. (2001) are as
follows:
1. skin
2. gas exchange surfaces of animals
3. fish gills
4. lungs of a bird
5. respiratory system of insects
6. countercurrent heat exchanger in fish
7. stomata
8. abscisic acid.
It is important to note that the listed results are not the only results provided by
the query. These eight were the most interesting and provided the most informa-
tion for biological inspiration. Plants, fish, birds, mammals, and insects all offer
inspiration for this case study. Notice that not all the biological solutions deal with
gas exchange. Exploring other implementations of exchange increases the design
space and creative solution potential.
As this design is for another research group, a meeting was held to discuss the
biological systems that were identified to choose a biological system for further
investigation as well as work through the remaining steps of the methodology. A
connection established during the meeting is the similarity of engineering mass
transfer concepts and the diffusion of respiratory gases into an insect. Although a
direct relationship between the domains was recognized, the external research
group did not see value in pursuing the relationship further. While reviewing a
detailed description of the avian respiratory system, the discussion led to a con-
nection that revealed the solution of a competing research group in the area of heat
exchangers for hydrogen vehicles. During the first inhale/exhale cycle of a bird,
the air travels to the trachea and into posterior air sacs and then is moved to the
parabronchi area where the exchange takes place (Campbell and Reece 2003).
During the second inhale/exhale cycle, the air moves forward to the cranial air sacs
and then exits the bird (Vattam et al. 2008). Realizing that the air within the bird’s
lungs essentially travels in a circle sparked discussion of a spiral-shaped surface
Fig. 5.15 Fish gills (Raven and Johnson 2002)
Fig. 5.16 Concept sketch for heat exchanger
that would effectively distribute hydrogen. That is, in fact, what a competing
research group did (Venere 2010).
While reviewing a detailed description of gills along with Fig. 5.15, however, a
researcher immediately made a connection between the gills of fish and hydrogen
distribution. Considering the fish blood as the coolant and the water surrounding
the fish as hydrogen, a distribution scheme inspired by the morphology and
physiology of fish gills was discussed. The concept sketch is shown in Fig. 5.16. In
this case, the biological functional model was not required, nor the engineering
knowledge base query. The concept generation step began during the meeting, and
the final concept was finalized by the external research group.
This case study demonstrated that it is possible to systematically design a
portion of a larger problem with only a small amount of information as well as take
inspiration from biology in the process. By analyzing the biological system and
making connections, a designer can become inspired. It is through these correla-
tions that the designer can recognize existing designs that are similar or develop an
innovative design.
5.5 Summary
The design methodology presented here represents a function-based design

approach devoted to developing biologically inspired solutions. It challenges and
guides a designer to make connections between the engineering and biology
domains to facilitate innovative design. Although systematic, the methodology

provides a designer multiple avenues toward achieving a biologically inspired
design. It is also envisioned as an inventive and iterative process, in terms of
developing connections between systems at multiple levels of fidelity. As one
level, or scale, of the biological system becomes understood, it leads to a deeper
understanding and a greater curiosity to explore further, thus leading to multiple
innovative designs. The four parts of the function-based framework, which can be
used as stand-alone design tools or methods, coalesce in the systematic biologi-
cally inspired design method as demonstrated by the smart flooring, chemical
sensor, and heat exchanger cases studies. Validation of the design methodology
was proven through matching each of the generated concepts in the case studies
with existing solutions or technologies. It was shown how the connections made
between biological and engineered systems were key to arriving at the biologically
inspired designs.
References
Addlesee MD, Jones A, Livesey F, Samaria F (1997) The ORL active floor. IEEE Pers Commun
4(5):35–41
Bai H, Shi G (2007) Gas sensors based on conducting polymers. Sensors 7:267–307
Balazs M, Brown D (2002) Design simplification by analogical reasoning. In: Cugini U, Wozny
M (eds) From knowledge intensive cad to knowledge intensive engineering, vol 79. Springer,
US, pp 29–44
Bhatta S, Goel A (1997) An analogical theory of creativity in design. In: Leake D, Plaza E (eds)
Case-based reasoning research and development, vol 1266. Springer, Berlin Heidelberg,
pp 565–574
Biomimicry Institute (2010) Biomimicry: a tool for innovation. http://www.biomimicry
institute.org/about-us/biomimicry-a-tool-for-innovation.html
Campbell NA, Reece JB (2003) Biology. Pearson Benjamin Cummings, San Francisco
Casakin H (2006) Visual analogy as a cognitive strategy in the design process: expert versus
novice performance. J Des Res 4(2), doi:10.1504/JDR.2004.009846
Casakin H (2006b) Assessing the use of metaphors in the design process. Environ Planning B:
Planning Des 33(2):253–268
Casakin H (2007) Metaphors in design problem solving: implications for creativity. Int J Des
1(2):21–33
Crane D (2005) New high-tech sensor-Laiden smart carpet may revolutionize building security.
http://www.defensereview.com/new-high-tech-sensor-laiden-smart-carpet-may-revolutionize-
building-security/
Dollens D (2009) BioDigital architecture uses metaphor to design living systems. http://
sensingarchitecture.com/3832/biodigital-architecture-uses-metaphor-to-design-living-systems-
dennis-dollens-video/
Dym CL, Little P (2004) Engineering design: a project-based introduction. Wiley, New York
Eggins BR (2002) Chemical sensors and biosensors. Analytical techniques in the sciences. Wiley,
Chichester
Fan Z, Chen J, Zou J, Bullen D, Liu C, Delcomyn F (2002) Design and fabrication of artificial
lateral line flow sensors. J Micromech Microeng 12:655–661
Forty A (1989) Of Cars, Clothes and Carpets: design metaphors in architectural thought. J Des
Hist 2(1):1–14. doi:10.1093/jdh/2.1.1
Fraden J (2004) Handbook of modern sensors : physics, designs, and applications. Springer,
New York
Gentner D (1983) Structure-mapping: a theoretical framework for analogy. Cogn Sci 7:155–170
Gentner D (1988) Analogical inference and access, vol Analogica. Lecture notes in artificial
intelligence. Morgan Kaufmann Publishers, Los Altos
Gick M, Holyoak K (1980) Analogical problem-solving. Cogn Psychol 12:306–355
Gnatzy W, Grunert U, Bender M (1987) Campaniform sensilla of Calliphora vicina (Insecta,
Diptera) I. Topography. Zoomorphology 160:312–319
Goel A (1997) Design, analogy and creativity. IEEE Expert Intell Syst Appl 12(3):62–70
Grundler P (2007) Chemical sensors an introduction for scientists and engineers. Springer, Berlin
Grunert U, Gnatzy W (1987) Campaniform sensilla of Calliphora vicina (Insecta, Diptera) II.
Typology. Zoomorphology 106:320–328
György I (2008) Conducting polymers: a new era in electrochemistry. Springer, Berlin
Helms M, Vattam SS, Goel AK (2009) Biologically inspired design: products and processes. Des
Stud 30(5):606–622
Hey J, Linsey J, Agogino AM, Wood KL (2008) Analogies and metaphors in creative design. Int
J Eng Educ 24(2):283–294
Hofstadter DR (1995) Fluid concepts & creative analogies: computer models of the fundamental
mechanisms of thought. Basic Books, New York
Hubka V, Eder EW (1984) Theory of technical systems. Springer, Berlin
Hyman B (1998) Engineering design. Prentice-Hall, New Jersey
IEE-Institution of Electrical Engineers (2003) Research news—Walk this way for the smart floor.
Electron Syst Softw 1:5–7
Johnson-Laird P (1989) Analogy and the exercise of creativity. In: Vosniadou S, Ortony A (eds)
Similarity and analogical reasoning. Cambridge University Press, Cambridge, pp 313–331
Liao HK, Yang ES, Chou JC, Chung WY, Sun TP, Hsiung SK (1999) Temperature and optical
characteristics of tin oxide membrane gate ISFET. IEEE Trans Electron Devices
46(12):2278–2281
Liau WH, Wu CL, Fu LC (2008) Inhabitants tracking system in a cluttered home environment via
floor load sensors. IEEE Trans Autom Sci Eng 5(1):10–20
Lindemann U, Gramann J (2004) Engineering design using biological principles. In: International
design conference—DESIGN 2004, Dubrovnik, 2004
Linsey J, Wood K, Markman A (2008) Modality and Representation in Analogy. AIEDAM
22(2):85–100
531(1):117–120
McKean E (2005) The new Oxford American dictionary. Oxford University Press, New York
Mitchell BK (2003) Chemoreception. In: Vincent HR,Ring TC (eds) Encyclopedia of insects.
Academic Press, Amsterdam, pp 169–174
Motamed M, Yan J (2005) A review of biological, biomimetic and miniature force sensing for
microflight. In: Paper presented at the IEEE/RSJ international conference on intelligent robots
and systems (IROS)
Nagai Y, Taura T (2006) Formal description of concept-synthesizing process for creative design.
In: Gero JS (ed) Design Computing and Cognition’06. Springer, Dordrecht, pp 443–460
Nagel JKS, Stone RB (2012) A computational approach to biologically-inspired design. Artif
Intell Eng Des, Anal Manuf, special issue DCC 2010 26(2):161–176
Nagel JK (2010) Systematic design of biologically-inspired engineering solutions. Doctoral
Dissertation, Oregon State University, Corvallis
Nagel JKS, Stone RB, McAdams DA (2010) An engineering-to-biology thesaurus for
engineering design. In: ASME IDETC/CIE 2010 DTM-28233, Montreal, Quebec, Canada
Nagel JKS, Stone RB, McAdams DA (2010) Exploring the use of category and scale to scope a
biological functional model. In: ASME IDETC/CIE 2010, DTM-28873, Montreal, Quebec,
Canada
Nagel JKS, Stone RB (2011) A systematic approach to biologically-inspired engineering design.

Paper presented at the ASME IDETC/CIE 2011, DTM-47398, Washington, D.C., USA 2011
Nagel RL, Tinsley A, Midha PA, McAdams DA, Stone RB, Shu L (2008) Exploring the use of
functional models in biomimetic conceptual design. J Mech Des 130(12):11–23
Nagel JKS, Stone RB, McAdams DA (2010) Function-Based Biology Inspired Concept
Generation. In: Mukherjee A (ed) Biomimetics, Learning From Nature. In-Tech, Croatia
Nagel JKS, Nagel RL, Stone RB, McAdams DA (2010d) Function-based biologically-inspired
concept generation. Artif Intell Eng Des Anal Manuf 24(4):521–535
Orr RJ, Abowd GD (2000) The smart floor: a mechanism for natural user identification and
tracking. Paper presented at the Conference on Human Factors In Computing Systems (CHI),
Hague, Netherlands
Otto KN, Wood KL (2001) Product design: techniques in reverse engineering and new product
development. Prentice-Hall, Upper Saddle River
Pahl G, Beitz W, Feldhusen J, Grote KH (2007) Engineering design: a systematic approach, 3rd
edn. Springer, Berlin
Purves WK, Sadava D, Orians GH, Heller HC (2001) Life, the science of biology, 6th edn.
Sinauer Associates, Sunderland
Raven PH, Johnson GB (2002) Biology. McGraw-Hill, Boston
Richardson B, Leydon K, Fernström M, Paradiso JA (2004) Z-Tiles: building blocks for modular,
pressure-sensing floorspaces. Paper presented at the Conference on Human Factors In
Computing Systems (CHI), Vienna, Austria
Shu LH, Hansen HN, Gegeckaite A, Moon J, Chan C (2006) Case study in biomimetic design:
handling and assembly of microparts. Paper presented at the ASME 2006 IDETC/CIE,
Philadelphia, PA
Smith GF (1998) Idea generation techniques: a formulary of active ingredients. J Creative Behav
32(2):107–133
Stock AM, Robinson VL, Goudreau PN (2000) Two-component signal transduction. Annu Rev
Biochem 69:183–215
Stone RB (1997) Towards a theory of modular design. University of Texas at Austin, Austin
Stone R, Wood K (2000) Development of a functional basis for design. J Mech Des
122(4):359–370
Tsujimoto K, Miura S, Tsumaya A, Nagai Y, Chakrabarti A, Taura T (2008) A method for
creative behavioral design based on analogy and blending from natural things. In: 2008
ASME IDETC/CIE, New York, USA, 2008. DETC2008-49389
Ullman DG (2009) The mechanical design process, 4th edn. McGraw-Hill, Inc., New York
Ulrich KT, Eppinger SD (2004) Product design and development. McGraw-Hill/Irwin, Boston
Vattam S, Helms M, Goel A (2008) Compound analogical design: interaction between problem
decomposition and analogical transfer in biologically inspired design. In: Third international
conference on design computing and cognition, Atlanta, 2008. Springer, Berlin, 377–396
Venere E (2010) Engineers design, build major component for hydrogen cars. http://
www.physorg.com/news186221598.html
Vincent JFV, Mann DL (2002) Systematic technology transfer from biology to engineering.
Philos Trans Royal Soc Lond A 360:159–173
Voland G (2004) Engineering by design, 2nd edn. Pearson Prentice Hall, Upper Saddle River
Vorwerk & Co. TGCK (2004) Infineon thinking carpet. http://www.vorwerk-carpet.com/sc/
vorwerk/bildmeldung_thinkCarpet_en.html
Wicaksono DHB, Pandraud G, Craciun G, Vincent JFV, French PJ (2004) Fabrication and initial
characterisation results of a micromachined biomimetic strain sensor inspired from the
campaniform sensillum of insects. In: IEEE Sensors 2004, 542–545
Chapter 6
Information-Processing Theories
of Biologically Inspired Design
Ashok K. Goel, Swaroop Vattam, Bryan Wiltgen and Michael Helms
Abstract Starting from in situ studies, we develop an information-processing

theory of biologically inspired design. We compare our theory with two popular
theories of biologically inspired design: Biomimicry 3.8 Institute’s Design Spiral
and Vincent et al.’s BioTRIZ. While Design Spiral and BioTRIZ are normative
and prescriptive, our information-processing theory provides a descriptive and
explanatory account of the design paradigm. We examine if and how the process
of biologically inspired design is different from that of other design paradigms
beyond the differences between biological and technological systems. We posit
that biologically inspired design appears to be a distinct design paradigm in part
because it entails solution-based analogies in addition to the problem-driven
analogies typical of other design paradigms.

Keywords Biologically inspired design Biomimicry Biomimetics Bionics

Compound analogy Creativity Cross-domain analogy Design Innovation

Problem decomposition Problem-driven analogy Problem–solution coevolution

Solution-based analogy Task analysis Task model
A. K. Goel (&) S. Vattam B. Wiltgen M. Helms

School of Interactive Computing, Georgia Institute of Technology,
Atlanta, Georgia 30332, USA
S. Vattam
e-mail: svattam@cc.gatech.edu
B. Wiltgen
e-mail: bwiltgen@cc.gatech.edu
M. Helms
e-mail: mhelms3@cc.gatech.edu
A. K. Goel M. Helms
Center for Biologically Inspired Design, Georgia Institute of Technology,
Atlanta, Georgia 30332, USA

128 A. K. Goel et al.
6.1 Introduction
Biologically inspired design (Bar-Cohen 2011; Benyus 1997; Chakrabarti and Shu
2010; Shu et al. 2011; Vincent and Mann 2002; Yen and Weissburg 2007) perhaps
is one of the most important movements in contemporary design of technological
systems. The paradigm espouses use of analogies to biological systems in gen-
erating conceptual designs for new technological systems. This paradigm has
inspired many designers in the history of design, such as Leonardo da Vinci and
the Wright brothers. However, it is only over the last generation that the paradigm
has become a movement, pulled in part by the growing need for environmentally
sustainable design and pushed partly by the desire for creativity and innovation in
design. The design of windmill turbine blades mimicking the design of tubercles
on the pectoral flippers of humpback whales is one example of biologically
inspired design (Biomimicry 3.8 Institute 2011a). The tubercles are large bumps
on the leading edges of the flippers, which create even, fast-moving channels of
water flowing over them (Fish and Battle 1995). The whales thus can move
through the water at sharper angles and turn tighter corners than if their flippers
were smooth. When applied to wind turbine blades, they improve lift and reduce
drag, improving the energy efficiency of the turbine (Ashley 2004). The design of
fog harvesting devices (Shanyengana et al. 2002) inspired by the arrangement of
hydrophilic and hydrophobic surfaces on the back of the Namibian beetle
(Hamilton and Seely 1976) is another example of biologically inspired design
(Biomimicry 3.8 Institute 2011b).
Note that the designs of biologically inspired wind turbine blades and fog
harvesting devices illustrate both sustainable design as well as creative design.
However, although biologically inspired design is rapidly growing as a design
movement, its practice at present is ad hoc, with little systemization of either
biological knowledge from a design perspective or of the processes of biologically
inspired design. Transformation of the promising paradigm of biologically inspired
design into a principled methodology requires rigorous development of precise
theories of biologically inspired design (e.g., Vincent and Mann 2002).
For the purposes of our discussion, it is useful to make three distinctions
between different kinds and parts of design theories. Firstly, a design theory
provides an account of both the contents of knowledge and the processes of design
in some class of domains. Hubka and Eder (1988), for example, provide both a
content account of technological systems and a process account of designing them.
Similarly, in the context of biologically inspired design, Vincent et al. (2006)
provide both a content account of biological systems in terms of energy, infor-
mation, structure, substance, space, and time, and a process account of biologically
inspired design called BioTRIZ. In previous work, we have described contents of
knowledge of biological and technological systems in terms of the structure,
behavior, and function of specific systems (Goel et al. 2012), the function and
behavior of abstract design patterns (Goel et al. 2011a), and types of knowledge
that are transferred from biological systems to technological systems in
6 Information-Processing Theories of Biologically Inspired Design 129
biologically inspired design (Vattam et al. 2010a). We also described partial

process accounts of biologically inspired design such as compound analogies
(Vattam et al. 2010a), problem-driven and solution-based design processes (Helms
et al. 2009), and problem–solution coevolution (Helms and Goel 2012). Our goal
in this chapter is to further develop the process account of biologically inspired
design.
Secondly, process accounts of design can be normative and prescriptive or
descriptive and explanatory. Pahl et al.’s (2007) theory of systematic design, for
example, is normative and prescriptive; Simon’s (1996) theory of design, on the
other hand, is mostly descriptive and explanatory. Most information-processing
theories of biologically inspired design are normative and prescriptive. This
includes BioTRIZ (Vincent et al. 2006), Biomimicry 3.8 Institute’s Design Spiral
(Biomimicry 3.8 Institute 2009), Srinivasan and Chakrabarti’s (2010) GEMS
model, as well as Nagel et al.’s (2013) account. In contrast, our information-
processing theory of biologically inspired design is descriptive and explanatory; it
emerges out of a series of in situ studies of the practice of biologically inspired
design (e.g., Helms and Goel 2012; Helms et al. 2009, 2010; Vattam et al. 2007,
2010a, b; Wiltgen and Goel 2011). Linsey et al. (2008) and Mak and Shu (2008)
too have empirically investigated biologically inspired design from an informa-
tion-processing perspective.
Thirdly, descriptive theories can focus on the microstructure or macrostructure
of cognition. Newell and Simon (1972), for example, describe specific memory,
inference, and attention processes for some kinds of human problem solving and
decision making. Their work focuses on the microstructure of cognition. Similarly,
Gentner (1983), Hofstadter (1996), Holyoak and Thagard (1996), and Kolodner
(1993) provide alternative cognitive accounts of analogy. In contrast, we conduct a
task analysis of the macrostructure of behavior in biologically inspired design.
(See Crandall et al. (2006) and Shraagen et al. (2000) for the methodology of task
analysis.) Our task analysis helps generate a task model of biologically inspired
design. A task model describes the processes and the knowledge that result in the
accomplishment of complex, extended, open-ended tasks such as design (e.g.,
Chandrasekaran 1990) as well as design methods such as design by plan instan-
tiation (Brown and Chandrasekaran 1989) and case-based design (Goel and
Chandrasekaran 1992).
To summarize, our main goal in this chapter is to develop a task model of cross-
domain analogies from biology to engineering in the conceptual phase of bio-
logically inspired engineering design. In general, an information-processing theory
of biologically inspired design may describe the behaviors of an individual
designer, the interactions among a team of designers, or the behaviors of a design
team viewed as a unit. Although we are interested in all three levels of aggrega-
tion, in this work, we focus on interdisciplinary design teams of biologists and
engineers viewed as the unit of analysis. We view this work as a step in a bigger,
longer-term agenda of developing (1) a design methodology that promotes sys-
tematization of biologically inspired design, (2) pedagogical techniques that foster
education and training in biologically inspired design, and (3) interactive
technologies that facilitate the work of teams engaged in biologically inspired

design. In this paper, first we briefly summarize the context and methodology of an
empirical study of biologically inspired design. Then, we conduct a task analysis
and develop a task model of creative analogies in biologically inspired design.
Next, we compare our theory with the two best-known information-processing
theories of biologically inspired design: Design Spiral and BioTRIZ. Finally, we
examine if and how biologically inspired design is fundamentally different from
other design paradigms.
6.2 Research Context and Methodology
In this section, we briefly describe the context of our observations and the
methodology of our analysis of biologically inspired design practice. Since 2006,
we have observed ME/ISyE/MSE/BME/BIOL 4740, a yearly, interdisciplinary,
project-based course on biologically inspired design taught jointly by biology and
engineering faculty at the Georgia Institute of Technology. The class is composed
of mostly senior-level undergraduate students from biology, biomedical engi-
neering, industrial design, industrial engineering, mechanical engineering, and a
variety of other disciplines. Although it evolves a little every year, the course is
consistently structured around lectures, found object exercises, journal entries, and
one or more design projects. Some lectures focus on case studies of biologically
inspired design. Some lectures pose problems for students to solve in small groups.
Other lectures discuss design processes involved in biologically inspired design,
such as reframing engineering problems in biological terms, functional decom-
position of a design problem, use of analogy in design, and quantitative analysis of
designs. Yen et al. (2010, 2011) provide a detailed account of the teaching and
learning in the course.
We base this work on the 2006–2010 classes of the ME/ISyE/MSE/BME/BIOL
4740 course. The extended design projects in the classes were the focal points of
our data collection. The projects involved identification of a design problem of
interest to the team and conceptualization of a biologically inspired solution to the
identified problem. Each design project grouped together an interdisciplinary team
of typically 4–5 students. The design projects included both problem-driven and
solution-based designs. Each team had at least one student with a biology back-
ground and a few from different engineering disciplines. Each team identified a
problem that could be addressed by a biologically inspired solution, explored a
number of solution alternatives, and developed a final solution design based on one
or more biologically inspired designs.
In this chapter, we analyze two design projects by one team from one class of
ME/ISyE/MSE/BME/BIOL 4740: the Shark Attack project and the Levee project.
Note that we have changed the project names for the sake of anonymity. The same
design team within the class conducted both projects. We report on this design
team in particular because one of the authors (Wiltgen) participated in the team;
Wiltgen and Goel (2011) provide details of the participant observation. Wiltgen
collected artifacts generated by the design team, including team presentations and
reports. He also collected all his individual class assignments and made audio
recordings of all team meetings. We then performed a task analysis informed by
Wiltgen’s observations of the design process. (Figures 6.1 and 6.2 are our
reconstructions of parts of the design process).
We also collected design reports from other design teams in various classes of
ME/ISyE/MSE/BME/BIOL 4740. Although we report in detail only on the two
projects of Wiltgen’s team, our task model is intended to capture a variety of
design processes that we observed in our studies of biologically inspired design.
For additional examples of design projects in the course, please see Helms and
Goel (2012), Helms et al. (2009, 2010), Nelson et al. (2009), Wilson et al. (2010),
and Vattam et al. (2007, 2010a, b).
6.2.1 The Shark Attack Project
The Shark Attack project lasted a little more than a month. In this project, each
team in the class first selected a small set of biological organisms. Afterward, each
team member was assigned to research a particular organism. Each team member
was next assigned to find a human problem that he or she thought his or her
researched organism could help solve and to investigate any existing solutions to
Fig. 6.1 The level of functional decomposition of the target design problem and the source
biological system in the Shark Attack project at which analogical transfer from the biological
source to the target problem occurred. The design team was inspired by the way the snapping
shrimp (or pistol shrimp) creates sound. They used this inspiration to design a device to prevent
shark attacks on humans by generating sounds using cavitation. (This figure was derived from the
design artifacts of the Shark Attack project.)
Fig. 6.2 The process of compound analogy in the Levee project during one iteration. Part 1
illustrates the decomposition of the target design problem into four sub-problems that the design
team addressed. Parts 2 and 3 depict a high-level mapping between a sub-problem and a
biological system. Part 4 shows the individual solutions to the sub-problems in relation to the
original design problem. Note that only two analogies made by the design team are shown. (This
figure was derived from the design artifacts of the Levee project.)
the problem. Each team member was then assigned to create a mapping between
the researched organism and the design problem. Finally, the design team was
assigned to pick one team member’s work to pursue and to decide what organisms
the team should keep or discard. The designing continued from there. For details of
the assignments, please see Yen et al. (2011, 2013).
The Shark Attack project was directed at preventing shark attacks off of the
coast of the United States without harming the sharks. The students designed an
underwater sound-based shark repellant device inspired by the snapping shrimp
(Ritzmann 1974), a small shrimp with the ability to create loud, underwater sound
waves using one of its claws. The device worked by emitting sounds, generated by
the same cavitation mechanism that the snapping shrimp uses to emit sound but at
a frequency that sharks dislike. The design team envisioned that the device would
repel sharks without harming them.
6.2.2 The Levee Project
Following the Shark Attack project, the Levee project lasted about two and a half
months. Instead of starting with a biological system as in the Shark Attack project,
the Levee project began by identifying a design problem. The Levee project
occurred across two major iterations, with the first iteration being shorter in time
than the second. For this project, each team member was first assigned to come up
with a set of design problems. The team was then assigned to decide upon and
investigate one of those problems as well as existing solutions to the problem. At
the same time, each team member was assigned to find one natural system that
solves all or part of the problem. The designing continued from there. (Again, see
Yen et al. 2011, 2013 for details of the assignments).
The goal of the Levee project was to prevent another flood disaster like that
caused by 2005’s Hurricane Katrina by strengthening the levee system in New
Orleans. The design team identified several modalities of failure for the levees in
New Orleans, such as scouring, overtopping, and joint failure, and devised bio-
logically inspired solutions for each except for overtopping. For example, the team
drew inspiration from scaly-foot snails (Yao et al. 2010) based on which the levee
design would be resistant to incoming water forces in part by being multilayered
like the snails’ shells.
6.3 Task Analysis
From our observations and analysis, we identified several tasks and processes of
biologically inspired design. In this section, we analyze some of those tasks and
processes.
6.3.1 Cross-Domain Analogies
By definition, biologically inspired design engages cross-domain analogies, for

example, analogies from biology to engineering. Although we have observed that
extended episodes of biologically inspired design involve both within domain and
cross-domain analogies (Vattam et al. 2010a), it is the essentialness of cross-
domain analogies that defines the paradigm of biologically inspired design. Both
the Shark Attack and the Levee projects illustrate cross-domain analogies from
biology to engineering.
6.3.2 Problem–Solution Coevolution
Conceptual design in biologically inspired design entails problem–solution

coevolution (Helms and Goel 2012). That is, the design process iterates between
defining and refining the problem and the solution, with both the problem and the
solution influencing each other (Maher and Tang 2003; Dorst and Cross 2001). As
a solution (S) is developed and evaluated for a given problem (P), it reveals
additional issues, spawning a new conceptualization of the problem (P ? 1). The
process continues with the development of a new solution (S ? 1) and will iterate
until a final solution is decided upon.
For example, consider the Shark Attack project. The team originally pursued
Wiltgen’s individual idea, which was inspired by the snapping shrimp, to design a
decoy-like device that would attract the sharks to a location away from human
population using sound. Analysis of design artifacts suggests that the team dis-
covered new sub-problems upon evaluation of the idea, such as durability of the
decoy if sharks were going to attack it instead of humans. The overall problem
then evolved to account for these newly identified issues, for example, that one
must design a device that can handle getting attacked by a shark. The final design,
however, did not incorporate a solution to this sub-problem. We hypothesize that
this is because the team switched from a shark-attracting system to a shark-
repelling system (a radical evolution!), as a result of which the system was no
longer in danger of getting attacked by a shark, thus eliminating the sub-problem
altogether; this illustrates problem–solution coevolution in biologically inspired
design.
6.3.3 Problem Decomposition
Biologically inspired design engages decomposition of the target design problem

as well as functional decomposition of the biological system that acts as a source
analogue to the design problem (Vattam et al. 2007, 2010a). Decomposition of a
large, complex problem into smaller, simpler problems is a familiar idea in design
(e.g., Brown and Chandrasekaran 1989; Chandrasekaran 1990; Dym and Brown
2012; Hubka and Eder 1988; Pahl et al. 2007). Functional decomposition of a
system into sub-systems too is a familiar idea in design (e.g., Dym and Brown
2012; Hubka and Eder 1988; Pahl et al. 2007; Simon 1996). Thus, it is not
surprising that decomposition of target design problems and functional decom-
position of source biological systems permeate biologically inspired design.
However, these decompositions appear to play a special role in biologically
inspired design. The decomposition of the target design problem and the functional
decomposition of the source biological system help identify the appropriate level
for the analogical transfer from the biological system to the design problem.
Figure 6.1 illustrates both functional decomposition of the design problem and
functional decomposition of the biological system in the Shark Attack project.
Here, both the design problem and the biological system are functionally
decomposed until a level is reached where a function in the problem decompo-
sition is similar enough to a function in the system decomposition for analogical
transfer to occur. In this example, the ability of the snapping shrimp to defend
itself and hunt using sound was an inspiration for the design of a device to prevent
shark attacks using sound.
6.3.4 Problem Decomposition and Compound Analogies
Although problem decomposition is common in almost all kinds of design, it

appears to play a second special role in biologically inspired design. In particular,
we found that biologically inspired design often entails compound analogies in
which a new design concept is generated by composing the results of multiple
cross-domain analogies (Vattam et al. 2007, 2010a). This process of compound
analogical design relies on an opportunistic interaction between the processes of
memory and problem solving. In this interaction, the target design problem is
decomposed functionally, solutions to different sub-functions in the functional
decomposition are found through analogies to different biological systems
retrieved from memory, and the overall solution is obtained by composing the
solutions for achieving the different sub-functions. Thus, the sub-functions in the
functional decomposition of the design problem act as probes into a functionally
indexed memory of biological systems.
As an example, let us consider the Levee project. Figure 6.2 illustrates a part of
the design process in the project. During the two design iterations in the Levee
project, the design team identified six sub-problems that were related to levee
failure: scour (the eroding of soil by water), regional failure (local levee failures
causing whole regions to flood), cracking (water forces breaking and breaking
through the levee), foundation destruction (water undermining the foundation of
non-earthen levees), overtopping (water going over the levee), and joint failure
(water forces breaking and breaking through the joints of concrete levees). Let us
just consider the first design iteration here. Our analysis of the design artifacts
produced by the design team in the first iteration suggests that the team identified
the first five sub-problems and addressed the first four. The team did not identify a
single biological source that would overcome all four of these sub-problems.
Instead, the final design solution for the first iteration, which was more along the
lines of a set of solutions rather than a tightly integrated device, was a compound
solution; it took inspiration from (1) riparian buffers for the scour sub-problem, (2)
riparian buffers and Bacillus pasteurii bacteria for the foundation destruction sub-
problem, (3) scaly-foot snails for the cracking sub-problem, and (4) Polyrachis
sokolova ant nests for the regional failure sub-problem. Figure 6.2 depicts the
processes of analogy for the scour and cracking sub-problems; it also depicts a
high-level breakdown of the Polyrachis sokolova biological analogue.
6.3.5 Problem-Driven and Solution-Based Analogies
We observed the existence of two high-level analogical processes for biologically

inspired design based on two different starting points—problem-driven analogy
and solution-based analogy (Helms et al. 2009; Vattam et al. 2007, 2010a; see also
Wilson et al. 2010). As Fig. 6.3a illustrates, in a problem-driven analogical pro-
cess, designers identify a problem that forms the starting point for subsequent
problem solving. They usually formulate their problem in functional terms (e.g.,
stopping a bullet). In order to find biological sources for inspiration, designers
‘‘biologize’’ the given problem; that is, they abstract and reframe the function in
more broadly applicable biological terms (e.g., what characteristics do organisms
have that enable them to prevent, withstand, and heal damage due to impact?).
Designers use a number of strategies for finding biological sources relevant to the
design problem at hand based on the ‘‘biologized’’ question, and then they research
the biological sources in greater detail. Important principles and mechanisms that
are applicable to the target problem are then extracted to a solution-neutral
abstraction and applied to arrive at a trial design solution.
The Levee project is an example of the problem-driven analogical process of
biologically inspired design. The design team began with the problem of pre-
venting levee failure and conducted a search for biological sources. The search
returned such organisms as the scaly-foot snails and riparian buffers. Principles
were extracted from these solutions (such as multilayered structures and dense
underground networks, respectively), which were then applied to the design
problem.
On the other hand, in the solution-based analogical process, as illustrated in
Fig. 6.3b, designers begin with a biological source of interest. The designers
understand (or research) their biological source to a sufficient depth to support the
extraction of deep principles from it. Then, they find human problems to which the
principle can be applied. Finally, they apply the principle to develop a design
solution to the identified problem.
6 Information-Processing Theories of Biologically Inspired Design
Fig. 6.3 The twin analogical processes of biologically inspired design: a problem-driven analogy and b solution-based analogy
137
The Shark Attack project is an example of the solution-based analogical process

of biologically inspired design. Early in the design project, each designer in the
ME/ISyE/MSE/BME/BIOL 4740 class was given an assignment to investigate a
biological source, to determine a design problem for which the biological source
might be useful, and to derive an analogical mapping between the two. Wiltgen
selected the snapping shrimp. After doing some amount of research on the snap-
ping shrimp, he picked the shark attack problem as a potentially good match and
came up with the initial solution to the problem. The design team chose to pursue
Wiltgen’s design idea. During the course of the design project, the team developed
an understanding of the snapping shrimp and the principle of cavitation as a means
to create sound. Although other factors also likely influenced the design episode,
evidence for the influence of the snapping shrimp analogue is apparent in the final
design of the shark-repelling device that used cavitation as its sound generation
mechanism.
We note that the problem-driven and solution-based analogical design pro-
cesses have different characteristics. Compared to problem-driven analogical
processes, solution-based analogical processes tend to exhibit not only design
fixation but also a fixation on the structure of the biological design (Helms et al.
2009; Vattam et al. 2007). Again compared to problem-driven processes, solution-
based design processes also tend to more often result in the generation of multi-
functional designs, that is, where a single design principle meets multiple func-
tional goals (Helms et al. 2009; Vattam et al. 2007).
It is important to note also that our findings about problem-driven and solution-
based analogical processes of biologically inspired design are not limited to the
design projects in the ME/ISyE/MSE/BME/BIOL 4740 class. Vattam et al. (2007)
describe a narrative analysis of about seventy real-world case studies reported in
the literature on biologically inspired design. We found that a majority of the
successful case studies entailed solution-based analogy. Further, Vattam et al.
(2007) found that some case studies of biologically inspired design described in
the literature as based on problem-driven analogical design in fact appear to have
been based on solution-based analogical design. Solution-based analogy also
characterizes the two examples we briefly described in the introduction to this
chapter: the design of wind turbine blades inspired by the design of pectoral
flippers of humpback whales and the design of fog harvesting systems inspired by
the design of hydrophilic surfaces on the back of the Namibian beetle. These
findings have had a substantial impact on the teaching and learning in the ME/
ISyE/MSE/BME/BIOL 4740 course. For example, until 2008, the course used to
have design projects that emphasized only problem-based analogical design as is
commonly prescribed by normative theories of biologically inspired design. As a
result of our empirical findings and the descriptive account we developed to
explain the findings, the instructors of the class introduced design projects oriented
toward solution-based analogical design. Yen et al. (2013) describe the evolution
of the class in detail.
It may be important to note also that our notions of problem-driven and solu-
tion-based analogical processes in biologically inspired design have little in
common with similarly named processes described by Kruger and Cross (2006).
Kruger and Cross are referring to the observation that once a designer has pro-
posed a conceptual solution to a design problem, subsequent information pro-
cessing often is driven by both the design problem and the proposed conceptual
solution. In contrast, our problem-driven and solution-based analogical processes
pertain to two different kinds of analogies entailed by biologically inspired design:
one starts with the design problem and makes an analogy to a biological solution
(problem-driven analogy), and the other starts with a biological solution and
makes an analogy to a design problem (solution-based analogy).
6.4 A Task Model
In this section, we first summarize a generic task model of analogical design. We

will then elaborate this generic task model to incorporate the findings presented in
the previous section to obtain our task model of biologically inspired design.
Figure 6.4 illustrates a generic task model of analogical design based on our
earlier work (Goel 1997; Goel and Bhatta 2004) that is consistent with informa-
tion-processing theories of analogical reasoning in general (Gentner 1983; Gentner
et al. 2001). The overall task is design (see Fig. 6.4). This is accomplished by
using the method of analogical reasoning. The analogical reasoning method sets up
sub-tasks like retrieval of a source analogue, mapping and transfer of relevant
knowledge across source and target to obtain the new solution, and evaluation and
storage of the new solution. Each sub-task (e.g., retrieval) might, in turn, be
accomplished by one of several methods (e.g., feature-based similarity matching
for retrieval). Knowledge, here, refers to the knowledge inputs and outputs asso-
ciated with the processing of each task, sub-task, or a method. For example, the
knowledge associated with the sub-task of transfer includes what may get trans-
ferred between the source and the target design situations; this can include, among
Fig. 6.4 A generic task model of analogical design

others, elements of a previous design like components and relationships between

components.
Our task model of biologically inspired design is based on the above generic
task model of analogical design, but will extend the generic task model to
incorporate two key findings from our study described above: (1) problem-driven
analogy and solution-based analogy and (2) compound analogy.
6.4.1 Incorporating Problem-Driven and Solution-Based

Analogies
Earlier, we identified two processes followed by designers engaged in biologically

inspired design, suggesting two methods for the task of biologically inspired
design: the problem-driven analogy and the solution-based analogy. The problem-
driven analogical process incorporates the design sub-tasks of problem formula-
tion, problem reframing, biological solution search, biological solution definition,
principle extraction, and principle application. Similarly, the solution-based ana-
logical process incorporates the design sub-tasks of biological solution identifi-
cation, biological solution definition, principle extraction, solution reframing,
problem search, problem definition, and principle application.
As one might expect given that biologically inspired design is based on cross-
domain analogies between biological and technological systems, there are corre-
spondences between many of the sub-tasks in the generic task model of analogical
design (Fig. 6.4) and the sub-tasks in the problem-driven and solution-based
analogical processes of biologically inspired design (Fig. 6.3). Figure 6.5 illus-
trates some of these correspondences. For example, the ‘‘biological solution
search’’ sub-task in the problem-driven analogical process and ‘‘problem search’’
sub-task in the solution-based analogical process correspond to the ‘‘retrieval’’
sub-task in the generic task model of analogical design. The aggregate of ‘‘bio-
logical solution definition,’’ ‘‘principle extraction,’’ and ‘‘principle application’’
sub-tasks in the problem-driven analogical process corresponds to the ‘‘mapping’’
and ‘‘transfer’’ sub-tasks in the generic task model of analogical design; similarly,
the aggregate of ‘‘problem definition’’ and ‘‘principle application’’ in the solution-
based analogical process corresponds to the ‘‘mapping’’ and ‘‘transfer’’ sub-tasks
in the generic task model of analogical design.
On the other hand, there are also sub-tasks in the problem-driven and solution-
based analogical processes of biologically inspired design (Fig. 6.3) that are not
directly matched by sub-tasks in the generic task model of analogical design
(Fig. 6.4). In particular, there are ‘‘problem abstraction’’ and ‘‘solution abstrac-
tion’’ sub-tasks in our task model of biologically inspired design that are prepa-
ratory to the sub-tasks of retrieval, mapping, and transfer that follow. Figure 6.6
illustrates the task model for biologically inspired design after incorporating
problem-driven analogy and solution-based analogy.
6 Information-Processing Theories of Biologically Inspired Design
Fig. 6.5 Correspondences between the generic task model of analogical design and the processes of (a) problem-driven analogy and
(b) solution-based analogy
141
142
Fig. 6.6 A task model of biologically inspired design after incorporating the two processes of problem-driven analogy and
solution-based analogy
A. K. Goel et al.
6.4.2 Incorporating Compound Analogies
Our task model of biologically inspired design should also account for compound
analogies. Figure 6.7 illustrates the task model of biologically inspired design after
incorporating the processes of compound analogy, problem-driven analogy, and
solution-based analogy. Here, S1 represents the initial solution obtained. We add a
new sub-task ‘‘evaluate’’ to both problem-driven and solution-based methods. This
sub-task evaluates the initial solution S1 generated by a method. If the evaluation
of S1 indicates that S1 addresses only a part of the design problem, then a new
design sub-problem is spawned to address the remaining part(s) of the problem.
Addressing the new sub-problem may lead to another partial solution S2. The sub-
task ‘‘compose’’ composes S1 and S2 to obtain a more complete solution to the
original problem. For expediency, it is assumed here that sub-task execution for
compound analogy is sequential, represented by one-way arrows between the
circles denoting the evaluation, designing, and composition. The actual process
may in fact involve much more complex interactions.
6.5 Comparative Analysis
In this section, we will compare our task model of biologically inspired design
with existing information-processing theories along some of the dimensions that
appear to characterize the design paradigm: (1) cross-domain analogies, (2)
compound analogies, and (3) the twin processes of problem-driven analogy and
solution-based analogy. We will not consider here the issues of problem–solution
coevolution or problem decomposition because they appear to characterize most
all design paradigms. In particular, we will compare our task model with the two
best-known information-processing theories of biologically inspired design: Bio-
mimicry 3.8 Institute’s (2009) Design Spiral and Vincent et al.’s (2006) BioTRIZ.
We should note, however, that Design Spiral, BioTRIZ, and our own task
model are all still under development and evolving. Deldin et al. (2013) describe
the AskNature Webportal that provides a functionally indexed database of bio-
logical systems in support of Design Spiral. Similarly, Vincent (2013) provides an
ontology for representing biological systems in support of BioTRIZ.
6.5.1 Design Spiral
Perhaps the most popular design process used for biologically inspired design is
Biomimicry 3.8 Institute’s (2009) Design Spiral illustrated in Fig. 6.8. In partic-
ular, Fig. 6.8a illustrates Biomimicry 3.8 Institute’s Challenge to Biology Design
Spiral and Fig. 6.8b illustrates Biology to Design Spiral.
Fig. 6.7 Our task model of biologically inspired design after incorporating the processes of
compound analogy, problem-driven analogy, and solution-based analogy
The basic idea of design spiral as a design process has been in the design
literature for some time. For example, Evans (1959) describes design spiral for
ship design and Boehm (1988) describes design spiral for software design and
development. However, the Biomimicry 3.8 Institute pioneered the Design Spiral
process in biologically inspired design and related it to its AskNature database
(Biomimicry 3.8 Institute 2008).
A comparison between our task model and the Biomimicry 3.8 Institute’s
Design Spiral reveals the following similarities and differences:
• Both Design Spiral and our model address cross-domain analogies between
biological and technological systems.
Fig. 6.8 a Biomimicry 3.8 Institute’s Challenge to Biology Design Spiral (adapted from
Biomimicry 3.8 Institute 2009). b Biomimicry 3.8 Institute’s Biology to Design Spiral (adapted
from Biomimicry 3.8 Institute 2009)
• Design Spiral appears to be a normative and prescriptive theory for biologically

inspired design. In contrast, our task model is a descriptive and explanatory
theory based on in situ observations of biologically inspired design in practice.
• Both Design Spiral and our model characterize biologically inspired design as
an iterative process. Design Spiral also captures the notion that the first iteration
is at a high level of abstraction, the next is more detailed, and so on.
• Although there are important differences, Challenge to Biology Design Spiral
appears similar to our problem-driven analogical process, and Biology to Design
Spiral seems similar to our solution-based analogical process.
• Design Spiral does not directly address compound analogies. However, the
iterative nature of Design Spiral suggests that the process can support compound
analogies. Our model explicitly accounts for compound analogies.
• The steps in Design Spiral are a little informal. We are trying to incrementally
formalize the tasks in our problem-driven and solution-based analogical pro-
cesses. For example, Wiltgen et al. (2011) describe methods of retrieval of
biological systems relevant to a design problem from a digital library based on
discrimination networks.
6.5.2 BioTRIZ
BioTRIZ (Vincent et al. 2006) is an information-processing theory of biologically

inspired design derived from the earlier theory of engineering invention called
TRIZ (Altshuller 1984). The TRIZ theory begins with a repository of design cases
with known solutions, where each case is indexed by contradictions that arose in
the original design situation. For example, consider a case in the repository that
represents the design of an airplane wing. In this case, the designer faces the
contradiction of obtaining a material that is both strong and lightweight and solves
it using a solution, say S1. This case is then indexed by the contradiction ‘‘strong
yet lightweight material.’’ Additionally, if the particular solution S1 belongs to a
more general way of resolving contradictions of a particular kind, it may be
categorized as a generic abstraction, such as ‘‘use porous materials’’ (to resolve the
contradiction of strong yet lightweight material). TRIZ posits the existence of forty
such generic ways of resolving contradictions, called inventive principles. The
inventive principles were extracted by dropping the specifics of a particular case
and domain and retaining the essence of how a particular class of contradictions is
solved, so we can imagine each principle pointing to numerous cases (potentially
belonging to different domains) in which that principle was used to resolve a
contradiction. The contradictions and the principles typically are organized in a
contradiction matrix (TRIZ Matrix 2004).
When the designer is presented with a design problem and intends to use TRIZ,
she reformulates the problem to identify certain key contradictions in the
requirements of the design. For each contradiction, she is reminded of a inventive
principle that is applicable for resolving that contradiction. In addition to sug-
gesting the essence of a solution for resolving that contradiction, the inventive
principle also points to a number of cases in which that inventive principle was
instantiated (See Fig. 6.9). These cases can originate from domains different from
the one in which the designer is currently working. TRIZ, however, does not
address the issue of how transfer occurs.
Vincent et al. (2006) recently developed a modified version of TRIZ, called
BioTRIZ, specifically for biologically inspired design. The primary difference
between the two theories is a change in the features that compose the contradiction
matrix. Whereas TRIZ defines 39 features with which to determine contradictions
Fig. 6.9 The TRIZ information-processing model of invention

and index into inventive principles, the current version of BioTRIZ has six
‘‘operational fields’’: substance, structure, space, time, energy, and information.
A comparison of our task model and BioTRIZ reveals the following similarities
and differences:
• Both BioTRIZ and our model address cross-domain analogies between bio-
logical and technological systems.
• BioTRIZ is a prescriptive theory of biologically inspired design, derived from
best practices in mechanical engineering design. In contrast, our task model is a
descriptive theory based on in situ observations of biologically inspired design.
• The processing in BioTRIZ is problem-driven. The processing in BioTRIZ
always begins with a specification of a design problem. It does not directly
address the solution-based analogical process. Our task model accounts for both
problem-driven and solution-based analogies.
• BioTRIZ does not directly address compound analogy. However, since a design
problem may contain multiple contradictions, and resolving those contradictions
may require the invocation of multiple inventive principles, compound analogy
appears to be feasible in BioTRIZ.
• BioTRIZ provides a detailed account of cross-domain analogical retrieval. Our
information-processing theory too provides a detailed account of analogical
retrieval (Wiltgen et al. 2011).
6.6 What Makes Biologically Inspired Design Different?
The above comparative analysis of theories of biologically inspired design brings

us to questions we are often asked by other design theorists: is biologically
inspired design different from other design paradigms? Or, put a little differently,
what precisely makes biologically inspired design a novel design paradigm from
an information-processing perspective? Note that the question here is not whether
biological and technological systems are different. As Vincent et al. (2006) note,
‘‘biology and technology solve problems in design in rather different ways’’;
biological systems often use information for functions for which technological
systems tend to use energy. French (1994) and Vogel (2000) make detailed
analyses of the similarities and differences between biological and technological
systems; biological systems in general tend to be more multifunctional than
technological systems. Instead, the question here is the following: are the pro-
cesses of biologically inspired design fundamentally different from that of other
design paradigms?
As we noted above, Design Spiral and BioTRIZ are normative and prescriptive
theories of biologically inspired design; they can be, and have been, applied to
many kinds of design. In fact, TRIZ, from which BioTRIZ evolved, arose out
of a study of mechanical engineering inventions, and it has been applied to many
kinds of design, ranging from mechanical engineering to organizational design
(TRIZ Matrix 2004). Similarly, the design spiral process is extensively used in the
design of artifacts ranging from ships to software.
However, our task analysis offers some insights into what may make biologi-
cally inspired design a different design paradigm from an information-processing
perspective. Firstly, biologically inspired design by definition is based on cross-
domain analogies. While many design processes in and out of biologically inspired
design sometimes engage cross-domain analogies (e.g., Goel 1997; Goel and
Bhatta 2004), and while biologically inspired design also frequently engages
within domain analogies (Vattam et al. 2010a), insofar as we know there are not
many other kinds of design that by definition are based on cross-domain analogies.
Secondly, biologically inspired design often entails compound analogies. For
example, the target design problem is decomposed functionally, solutions to dif-
ferent sub-functions in the functional decomposition are found through analogy to
different biological systems retrieved from a functionally indexed memory, and the
overall design solution is obtained by composing the solutions for achieving the
different sub-functions. While problem decomposition is common to almost all
kinds of design (e.g., Brown and Chandrasekaran 1989; Chandrasekaran 1990) and
while other kinds of design too sometimes engage in compound analogies (e.g.,
Goel and Chandrasekaran 1992), compound analogy appears to be a stronger
characteristic of biologically inspired design compared to other kinds of design.
Thirdly, biologically inspired design engages two different analogical pro-
cesses, namely problem-driven analogy and solution-based analogy. We first
observed these two analogical processes in our in situ studies of biologically
inspired design in practice. Subsequently, we conducted a narrative analysis of
some seventy case studies of biologically inspired design and confirmed this
observation. Insofar as we know, information-processing theories of analogy (e.g.,
Clement 2008; Dunbar 2001; Gentner 1983; Gentner et al. 2001; Gick and Ho-
lyoak 1983; Goel 1997; Holyoak and Thagard 1996; Keane 1988; Kolodner 1993)
focus on and emphasize problem-driven analogy. Further, insofar as we know,
information-processing theories of all other kinds of design focus on and
emphasize problem-driven design (e.g., Boehm 1988; Brown and Chandrasekaran
1989; Chandrasekaran 1990; Dorst and Cross 2001; Dym and Brown 2012; French
1996; Goel and Bhatta 2004; Goel and Chandrasekaran 1992; Hubka and Eder
1988; Maher and Tang 2003; Pahl et al. 2007; Simon 1996). Therefore, that
biologically inspired design entails both problem-driven and solution-based
analogies appears to be another definitional characteristic of biologically inspired
design.
Veros and Coelho (2011) take a different approach to comparing various
methods of biologically inspired design. In particular, they apply Biomimicry 3.8
Institute’s Challenge to Biology Design Spiral and our problem-driven and solu-
tion-based analogical design processes to a class of design problems. One of their
key findings is that the solution-based analogical process tends to more often
produce multifunctional designs, which confirms our own empirical observations
(Helms et al. 2009; Vattam et al. 2007).
6.7 Conclusions
In this chapter, we first presented a task analysis of biologically inspired design based
on our empirical observations of its practice in ME/ISyE/MSE/BME/BIOL 4740,
Georgia Tech’s introductory course on biologically inspired design. Our task analysis
indicates some of the fundamental processes of biologically inspired design: (1)
Biologically inspired design by definition engages cross-domain analogies. (2)
Problems and solutions in biologically inspired design coevolve. (3) Problem
decomposition is a fundamental process of biologically inspired design. (4) Bio-
logically inspired design often involves compound analogy, entailing a complex
interplay between the processes of problem decomposition and the processes of
analogical retrieval from memory. (5) Biologically inspired design entails two dis-
tinct but related processes: problem-driven analogy and solution-based analogy.
We then presented a task model of biologically inspired design. Since problem–
solution coevolution and problem decomposition are common to most design
paradigms, our task model focused on two other results of our task analysis: that
biologically inspired design often involves compound analogy and that biologi-
cally inspired design entails both problem-driven analogy and solution-based
analogy. Our task model of biologically inspired design is a descriptive theory of
biologically inspired design that has had significant influence on the teaching and
learning in ME/ISyE/MSE/BME/BIOL 4740 (Yen et al. 2011, 2013).
Next, we compared our task model with the two best-known information-pro-
cessing theories of biologically inspired design: Biomimicry 3.8 Institute’s Design
Spiral and Vincent et al.’s BioTRIZ. Both Design Spiral and BioTRIZ are nor-
mative and prescriptive theories of biologically inspired design. Both address
cross-domain analogies and, at least implicitly, allow compound analogies.
However, BioTRIZ appears to support only problem-driven analogies.
Finally, we analyzed what makes biologically inspired design a novel design
paradigm and, in particular, how the processes of biologically inspired design
differ from that of other design paradigms. Our task model suggests that biolog-
ically inspired design differs from other kinds of design in the use of cross-domain
analogies, the use of compound analogies, and the use of both problem-driven and
solution-based analogies. We are presently investigating the methodological,
technological, and pedagogical implications of our information-processing theory
of biologically inspired design.
Acknowledgments We are grateful to the instructors and students of the ME/ISyE/MSE/BME/

BIOL 4740 class from 2006 through 2010, especially Professors Jeannette Yen, Craig Tovey, and
Marc Weissburg, as well as the design team that developed the Shark Attack and Levee projects.
We thank the US National Science Foundation for its support of this research through a Cre-
ativeIT Grant (#0855916) titled ‘‘Computational Tools for Enhancing Creativity in Biologically
Inspired Engineering Design,’’ and a TUES Grant (#1022778) titled ‘‘Biologically Inspired
Design: A Novel Biology-Engineering Curriculum.’’ This chapter augments and amplifies the
information-processing analysis of biologically inspired design developed in Vattam et al. 2010b
and Goel et al. 2011b. Figures 6.1 through 6.7 and Figure 6.9 have been adapted from Goel et al.
2011b. We thank the reviewers of the earlier papers and this chapter for their critiques.
References
Altshuller G (1984) Creativity as an exact science. Gordon and Branch Publishers, Luxembourg
Ashley S (2004) Bumpy flying Scalloped flippers of whales could reshape wings. Sci Am
291(2):18, 20
Bar-Cohen Y (ed) (2011) Biomimetics: nature-based innovation. CRC Press, USA
Benyus J (1997) Biomimicry: innovation inspired by nature. William Morrow, New York
Biomimicry 3.8 Institute (2008) AskNature. http://www.asknature.org/. Last retrieved on April
25, 2011
Biomimicry 3.8 Institute (2009) Biomimicry: a tool for innovation. \http://www.
biomimicryinstitute.org/about-us/biomimicry-a-tool-for-innovation.html[. Last retrieved on
April 25, 2011
Biomimicry 3.8 Institute (2011a) Energy. \http://www.biomimicryinstitute.org/case-studies/
case-studies/energy.html[. Last retrieved on April 28, 2011
Biomimicry 3.8 Institute (2011b) Water vapor harvesting: namib desert beetle. \http://
www.asknature.org/strategy/dc2127c6d0008a6c7748e4e4474e7aa1[. Last retrieved on
April 28, 2011
Boehm BW (1988) A spiral model of software development and enhancement. IEEE Comput
21(5):61–72
Brown D, Chandrasekaran B (1989) Design problem solving: knowledge structures and control
strategies. Morgan Kaufmann, San Mateo
Chandrasekaran B (1990) Design problem solving: a task analysis. AI Mag 11(4):59–71
Clement J (2008) Creative model construction in scientists and students: the role of imagery,
analogy, and mental simulation. Springer, Dordrecht
Crandall B, Klein G, Hoffman R (2006) Working minds: a practitioner’s guide to cognitive task
analysis. MIT Press, Cambridge
Deldin JM, Laux J, Shuknecht (2013) The AskNature database: enabling solutions in biomimetic
design. In: Goel, McAdams, Stone (eds) Biologically inspired design: computational methods
and tools (this volume)
Dorst K, Cross N (2001) Creativity in the design process: co-evolution of problem-solution. Des
Stud 22:425–437
Dunbar K (2001) The analogical paradox. In: Gentner D, Holyoak KJ, Kokinov BN (eds) The
analogical mind: perspectives from cognitive science. MIT Press, Boston
Dym C, Brown D (2012) Engineering design: representations and reasoning, 2nd edn. Cambridge
University Press, Cambridge
Evans JH (1959) Basic design concepts. J Am Soc Naval Eng 71(4):671–678
Fish F, Battle J (1995) Hydrodynamic design of the humpback whale flipper. J Morphol
225:51–60
French M (1994) Invention and evolution: design in nature and engineering, 2nd edition.
Cambridge University Press, Cambridge, 1988
French M (1996) Conceptual design for engineers, 3rd edn. Springer, London
Gentner D (1983) Structure mapping: a theoretical framework for analogy. Cogn Sci
7(2):155–170
Gentner D, Holyoak K, Kokinov B (eds) (2001) The analogical mind: perspectives from
cognitive science. MIT Press, Cambridge
Gick M, Holyoak KJ (1983) Schema induction and analogical transfer. Cog Psych 15:1–38
Goel A (1997) Design, analogy, and creativity. IEEE Expert 12(3):62–70
Goel A, Bhatta S (2004) Use of design patterns in analogy-based design. Adv Eng Inform
18(2):85–94
Goel A, Bras B, Helms M, Rugaber S, Tovey C, Vattam S, Weissburg M, Wiltgen B, Yen J

(2011a) Design patterns and cross-domain analogies in biologically inspired sustainable
design. In Proceedings AAAI spring symposium on AI and sustainable design. Stanford
University, Palo Alto, 45–51 Mar 2011
Goel A, Chandrasekaran B (1992) Case-based design: a task analysis. In: Tong C, Sriram D (eds)
Artificial intelligence approaches to engineering design, vol II: innovative design. Academic
Press, San Diego, pp 165–184
Goel A, Vattam S, Wiltgen B, Helms M (2011b) An information-processing account of creative
analogies in biologically inspired design. In: Proceedings eighth ACM conference on
creativity and cognition, Atlanta, 71–80 Nov 2011
Goel A, Vattam S, Wiltgen B, Helms M (2012) Cognitive, collaborative, conceptual and
creative—Four characteristics of the next generation of knowledge-based CAD systems: a
Hamilton W, Seely M (1976) Fog basking by the Namib Desert beetle, Onymacris unguicularis.
Nature 262(5566):284–285
Helms M, Goel A (2012) Analogical problem evolution in biologically inspired design. In:
Proceedings 5th international conference on design computing and cognition, College station.
Springer, Berlin
30(5):606–622
Helms M, Vattam S, Goel K (2010) The effect of functional modeling on understanding
biological systems. In: Proceedings ASME 2010 international design engineering technical
conferences & computers and information in engineering conference (IDETC/CIE 2010).
Montreal, Quebec, Canada, 15–18 Aug 2010
Hofstadter D (ed) (1996) Fluid concepts & creative analogies: computer models of the
fundamental mechanisms of thought. Harvester Wheatsheaf, London
Holyoak K, Thagard P (1996) Mental leaps: analogy in creative thought. MIT Press, Cambridge
Hubka V, Eder W (1988) Theory of technical systems: a total concept theory of engineering
design. Springer, Dordrecht
Keane M (1988) Analogical problem solving. Ellis Horwood, Chichester
Kolodner J (1993) Case-based reasoning. Morgan Kauffman, San Francisco
Kruger C, Cross N (2006) Solution-driven vesus Problem-driven design: strategies and outcomes.
Des Stud 27(5):527–548
Linsey J, Wood K, Markman A (2008) Modality and representation in analogy. AI Eng, Des,
Manuf 22:85–100
Maher ML, Tang H (2003) Co-evolution as a computational and cognitive model of design. Res
Design 19(1):21–28
Nagel J, Stone R, McAdams D (2013) Function-based biologically inspired design. In: Goel,
McAdams, Stone (eds) Biologically inspired design: computational methods and tools (this
volume)
Nelson B, Wilson J, Yen J (2009) A study of biologically-inspired design as a context for
enhancing student innovation. In: Proceedings 39th ASEE/IEEE frontiers in education
conference, 2009
Newell A, Simon H (1972) Human problem solving. Prentice-Hall, Upper Saddle River
Pahl G, Beitz W, Feldhusen J, Grote K (2007) Engineering design: a systematic approach, 3rd
edition. Translated and edited by Wallace K, Blessing L, Springer, London
Ritzmann R (1974) Mechanisms for the snapping behavior of two alpheid shrimp, Alpheus
californiensis and Alpheus heterochelis. J Comp Physiol 95:217–236
Shanyengana E, Henschel J, Seely M, Sanderson R (2002) Exploring for as a supplementary
water resource in Namibia. Atmos Res 64(1–4):251–259
Shraagen J, Chipman S, Shalin V (eds) (2000) Cognitive task analysis. Lawrence Erlbaum, USA
Shu L, Ueda K, Chiu I, Cheong H (2011) Biologically inspired design. CIRP Ann Manuf Technol
60:673–693
Simon H (1996) Sciences of the artificial, 3rd edn. MIT Press, Cambridge
Srinivasan V, Chakrabarti A (2010) An integrated model of designing. ASME JCISE 10(3).
doi:10.1115/1.3467011 (from http://computingengineering.asmedigitalcollection.asme.org/
article.aspx?articleid=1402109)
TRIZ Matrix (2004) Interactive TRIZ contradiction matrix. (http://www.triz40.com/)
Vattam S, Helms M, Goel A (2007) Biologically inspired innovation in engineering design: a
cognitive study. Technical report GIT-GVU-07-07, graphics, visualization & usability center,
inspired design. AI Eng Des, Anal Manuf 24:467–481 Special issue on biologically inspired
design
Vattam S, Helms M, Goel A (2010b) Biologically inspired design: a macrocognitive account. In:
Proceedings ASME 2010 international design engineering technical conferences & computers
and information in engineering conference (IDETC/CIE 2010). Montreal, Quebec, Canada,
15–18 Aug 2010
Veros C, Coelho D (2011) Biologically inspired design: methods and validation. In: Coelho D
(ed) Industrial Design—New Frontiers. InTech open access books; Available in creative
commons: http://www.intechopen.com/books/industrial-design-new-frontiers/biologically-
inspired-design-methods-and-validation
Vincent J (2013) Finding the gaps in BID. In: Goel, McAdams, Stone (eds) Biologically inspired
design: computational methods and tools (this volume)
Vincent J, Mann D (2002) Systematic transfer from biology to engineering. Philos Trans R Soc
Lond 360:159–173
Vogel S (2000) Cat’s paws and catapults: mechanical worlds of nature and people. W.W Norton
and Company, New York
Des Stud 31(2):169–186
Wiltgen B, Goel A (2011) Case-based reasoning all over the place: the multiple roles of CBR in
biologically inspired design. In: Proceedings workshop on human-centered and cognitive
approaches to case-based reasoning at the 19th international conference on case based
reasoning. Greenwich, London, 2011. (An updated version is available at http://
dilab.gatech.edu/publications/wiltgengoel-icccbr-mbr-2011.pdf.)
Wiltgen B, Goel A, Vattam S (2011) Representation, indexing, and retrieval of biological cases
for biologically inspired design. In: Proceeding 19th international conference on case based
reasoning. Greenwich, London, Sep 2011
Yao H, Dao M, Imholt T, Huang J, Wheeler K, Bonilla A, Suresh S, Ortiz C (2010) Protection
mechanisms of the iron-plated armor of a deep-sea hydrothermal vent gastropod. Proc Natl
Acad Sci (PNAS) 107(3):987–992
engineering design. Adv Nat Sci 3(2):27–40
Yen J, Helms M, Tovey C, Weissburg M, Goel A (2013) Adaptive evolution of a biologically
inspired design course. In: Goel, McAdams, Stone (eds) Biologically inspired design:
computational methods and tools (this volume)
collected contributions. J Bioinspiration Biomimetics 2. doi:10.1088/1748-3190/2/4/E01
(from http://iopscience.iop.org/1748-3190/2/4/E01)
interdisciplinary education. In: Bar-Cohen Y (ed) Biomimetics: nature-based innovation.
Taylor & Francis, USA
Chapter 7
Adaptive Evolution of Teaching Practices
in Biologically Inspired Design
Jeannette Yen, Michael Helms, Ashok Goel, Craig Tovey

and Marc Weissburg
Abstract At Georgia Tech in 2005, we developed an interdisciplinary under-

graduate semester-long course, biologically inspired design (BID), co-taught each
year by faculty from biology and engineering. The objective of this chapter is to
share our teaching experience with those interested in teaching such a course
themselves. The specific curriculum of a BID course must depend on the student
mix, the institutional context, and instructor goals. Therefore, rather than pre-
senting a particular curriculum, we present key problems that we encountered in
our 8 years of teaching and how we addressed them. We expect that any who try to
teach such a course will face one or more of the same challenges, and we offer
numerous pedagogical approaches that can be tailored to their specific circum-
stances. By describing our solutions, their consequences, and the extent to which
they met our expectations, we also point out where tough student challenges still
exist that are in need of attention from the community.

Keywords Teaching biologically inspired design Learning biologically inspired

design Problem-driven design Solution-based design Analogical design

Cross-domain analogy Design by analogy Understanding biological systems
J. Yen (&) M. Weissburg
School of Biology, Georgia Institute of Technology, Atlanta, GA 30332, USA
e-mail: jeannette.yen@biology.gatech.edu
M. Weissburg
e-mail: marc.weissburg@biology.gatech.edu
M. Helms A. Goel
School of Interactive Computing, Georgia Institute of Technology, Atlanta, GA 30332, USA
e-mail: mhelms3@cc.gatech.edu
A. Goel
C. Tovey
School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta, GA
30332, USA
e-mail: ctovey@isye.gatech.edu

154 J. Yen et al.

Functional decomposition Structure-Behavior-Function Design evaluation

Team design Interdisciplinary design Interdisciplinary education Design

creativity Engineering design Engineering creativity Multi-disciplinarity

Team-based learning Analogical reasoning
7.1 Introduction
Biologically Inspired design (BID) is highly interdisciplinary. The following four

examples of BID illustrate the importance of contributions from different disci-
plines. For Velcro (see, for example, Simonton 2004), a close morphological
examination of a plant burr provided the inspiration for a successful design of a
fastener. For RHex (Altendorfer et al. 2001), a deep understanding of cockroach
locomotion and stability of multi-legged organisms and robotic engineering was
needed for success in an all terrain vehicle. The development of the Geckel wet
adhesive (Lee et al. 2007) required a thorough understanding of the chemical and
material properties of the biological solutions for gripping by geckos and sticking
underwater by mussels. For the honeybee web-hosting algorithm, which dynam-
ically allocates Web server resources to hosted services (Nakrani and Tovey
2007), the inventors needed mathematics to understand the bee behavior, the
nectar output of flower patches, and the patterns of internet traffic. Thus, taken as a
whole, BID spans science and engineering. Since few designers are likely to have
deep enough knowledge in a wide range of fields required for any given design,
BID often is collaborative. To teach a course well in this design paradigm likewise
requires expertise in biology, engineering, and design.
At Georgia Tech in 2005, we developed an interdisciplinary undergraduate
semester-long course [ME/ISyE/MSE/BME/BIOL 4740: BID] taught in the fall
semester each year (Weissburg et al. 2010; Yen et al. 2010, 2011). The course
recruits students from these majors: mechanical engineering, industrial and sys-
tems engineering, materials science engineering, biomedical engineering, and
biology; in addition, we sometimes get majors from industrial design, architecture,
chemistry, mathematics, or nuclear engineering. The course is co-taught by faculty
from biology and engineering. Although it is not possible to teach deep knowledge
in all these disciplines in a one-semester course, we can show the students that
such knowledge can take them on the exciting path of BID. At Georgia Tech, we
restrict this course to juniors and seniors who have established their majors, thus
are able to bring specialized knowledge to the table. One of the goals of the course
is to show the students how a deep understanding of their field and experience in
working on an interdisciplinary team can enable them to be more inventive and
creative. The goal is to motivate them to learn as much as they can in their field,
then come to class to practice how to collaborate. In today’s information-rich
setting with easy access to knowledge resources, and in an increasingly interdis-
ciplinary and collaborative design world, we turn our emphasis in this chapter on
how to retrieve that knowledge, communicate it effectively across disciplines, and
7 Adaptive Evolution of Teaching Practices in Biologically Inspired Design 155
utilize it to solve problems in interdisciplinary teams. We think BID is a useful

environment for learning and applying these skills.
In Yen et al. (2011), we defined the following five learning goals of the BID
course: (1) Novel design techniques; (2) Interdisciplinary communication; (3)
Science and Engineering knowledge outside core domain; (4) Interdisciplinary
collaboration; and (5) Application of existing knowledge to a new field. To reach
these goals, we presented the skills we taught, the exercises that we developed to
enable the students to practice these skills, and the format of the class to indicate
how we deployed these exercises. Our pedagogical techniques were grounded in
the theory and practice of interdisciplinary research and education, recommended
in the cognitive and learning sciences (e.g., Ausubel 2000; Bransford et al. 2000;
Bybee 1997; Lave and Wenger 1991; Vygotsky 1978) as well as recommendations
for teaching science (e.g., National Research Council 2011) and biology (e.g.,
National Research Council 2009). In Yen et al. (2011), we provided details on
some of the more complex exercises such as biological design in natural evolution,
problem decomposition, and analogical reasoning. We presented lessons learned
from the first 3 years of teaching this course (2005–2007). To summarize, we
found that creativity improved through the use of analogical reasoning to link
biological functions to engineering challenges. We experienced clear differences
in how biologists and engineers solve problems, which identified interdisciplinary
communication gaps that were overcome, to some extent, by giving students
practice in both domains in their interdisciplinary projects as well as by motivating
them to apply their own knowledge to new problems and domains.
During the last 5 years (2008–2012), the BID course retained all the learning
goals mentioned above. As this course and the field at large mature, additional
challenges have arisen and become part of our learning objectives. In this chapter,
we relay the triumphs and tribulations encountered in our ambitious plan to pro-
vide students with the opportunity to collaborate across disciplines through BID as
well as learn about BID itself. The objective of this chapter is to share our teaching
experience with those interested in teaching a BID course themselves. The cur-
riculum of a BID course is flexible and depends on the student mix as well as the
institutional context and instructor goals. Therefore, rather than presenting a
specific curriculum, we present key problems that we encountered in the 8 years of
teaching and how we addressed them. We expect that anyone who tries to teach
such a course will face one or more of the same challenges, and we offer numerous
pedagogical approaches that can be tailored to different circumstances. By
describing our solutions, their consequences, and the extent to which they met our
expectations, we also point out where tough student challenges still exist that are in
need of attention from the community.
156 J. Yen et al.
7.2 Key Challenges
While the inventory of skills required to generate successful BIDs is vast, we focus
on the following key challenges that we see students struggle with year after year.
1. Searching for biological systems
2. Understanding biological systems
3. Identifying and understanding good design problems
4. Analogy mapping and transfer
5. Communicating across discipline boundaries
6. Communicating complex system knowledge
7. Teaming in an interdisciplinary environment
8. Maintaining equal engagement throughout the process
9. Evaluating designs
7.2.1 Searching for Biological Systems
In past studies (e.g., Vattam and Goel 2011), we have documented that up to 25%
of out-of-class time can be spent simply searching for the right biological organism
to solve a particular problem. While tremendous resources already are being
devoted to solving this problem technologically (as this volume attests), currently
students comb through volumes of textbooks, scientific databases, and the Web to
find what they need. When they do find something promising: (a) it is often written
in academic or technical language that is difficult for a non-expert to understand;
(b) it takes time to determine whether it will be applicable to their design problem;
and (c) it is usually not design oriented and requires translation before it is useful.
These problems are magnified when the design teams do not have individuals with
broad-based biological knowledge. Even when biologists are represented, they
may not have the background that would be most desirable. For example, students
with strong knowledge of basic organismal biology (e.g., comparative physiology,
functional morphology, behavior, invertebrate and vertebrate biology) are the most
well-equipped to search and identify appropriate systems for human-scale prob-
lems. Student designers typically (although not always) attempt to solve human-
scale problems as this familiar scale is where humans have the most experience.
For design problems at smaller or larger scales (e.g., at the molecular scale such as
filtering pharmaceuticals from the water supply or at large scales such as city
planning), different specializations may be helpful (from organic chemistry to
ecosystem structure). Programs such as ours, where biologists are less well rep-
resented than engineers, require specific curricular elements to increase the ability
of students to mine the biological literature.
7.2.2 Understanding Biological Systems
Our course has a mix of engineering, biology, architecture, and design students
who must work as a team to understand the key mechanisms of the biological
system so that they are capable of abstracting the mechanism and applying it to an
engineering design problem. Whereas the biologists may have a deep understanding
of a given biological phenomenon in its biological context, it is a challenge for them
to communicate that understanding in such a way that the non-biologists understand
it well enough to use in a design. This is exacerbated by the number and breadth of
biological systems the students are asked to learn about, the limited time available
for deep understanding, and the natural tendency of students to either focus on
structural details, and/or use improper analogies to facilitate or communicate their
understanding (e.g., the analogy: xylem in a tree acts like a straw in a drink—is not
accurate at the mechanistic level since a pressure gradient is used to transport water
by the straw while the molecular force of cohesion is used to transport water up to
its leaves from its roots). Biologists who have experience examining biological
systems in terms of function (e.g., biomechanics, physiology, and behavior) initially
may be more able to communicate their understanding of biological systems in an
appropriate manner. We find that architects and designers tend to focus on the
structural elements of the system, at least initially, and require practice in thinking
about function in biology. Engineers think about function, but generally lack the
requisite biological knowledge.
7.2.3 Identifying and Understanding Good Design Problems
Throughout their scholastic careers, students are taught how to solve problems that
are given to them. Less frequently faced in an academic context, this course
presents a unique set of challenges when students are asked to identify and define a
problem of their own choosing. Students in our BID class have (in early course
iterations) devoted up to half the semester defining their design problem when
challenged with a wide-open problem landscape. We have learned that BID may
originate with the standard process of problem-driven design or may begin from a
solution-based approach, where the unique mechanisms of a biological solution of
interest determine which problems one may wish to explore (Helms et al. 2009).
Thus, in solution-based design, problem identification is a critical aspect of BID.
As instructors, we must balance the requirements of good problem identification
and formulation against the needs of teaching a complete BID process.
7.2.4 Analogy Mapping and Transfer
Students often manifest cognitive limitations, biases, and errors (Helms et al.
2009). Whereas students naturally and effortlessly make analogies during the
158 J. Yen et al.
process of design, their analogies can be superficial. Students fall prey to a kind of
confirmation bias, focusing on initial superficial alignment between analogue and
problem, while ignoring deeper dissimilarities until they are forced to confront
them late in the design process. For example, a student team in 2011 attempted to
design a collapsible bicycle helmet inspired by the girdled lizard, which bends
itself into a circle of spiked bands for protections against predators. The design
failed because ‘‘protection’’ in the case of a bicycle helmet means dissipation and
absorption of energy from a collision, whereas in the case of the lizard, it means
resistance to penetration by claws or teeth.
7.2.5 Communicating Across Discipline Boundaries
Communication often is hampered by differences in the specialized terminology of

different disciplines. For example, for biologists, ‘‘stress’’ represents extreme
conditions such as heat, lack of water, or predators, to which organisms must
respond using physiological, behavioral, genetic, developmental, or other mech-
anisms. For mechanical engineers, ‘‘stress’’ is the measure of force per unit area in
a deformable body. Such differences occur even within the broad field of engi-
neering, but become increasingly large as more disciplines participate.
7.2.6 Communicating Complex Systems Knowledge
In many disciplines, there are systems so complex that it seems impossible to draw
possible analogies to another field without extensive research and teaching
experience. We have found that decomposing a particular function of a system into
subfunctions allows others to understand at least the interactions occurring at one
level accurately even without gaining a full understanding of how all the functions
in a complex system are integrated. If a subfunction still remains out of grasp of
understanding, then it too must be decomposed further into its underlying mech-
anisms until a principle, common to both disciplines, is reached. This journey may
take the designers several levels deep down in the hierarchical breakdown of the
problem or the natural system, but success is more likely when the team reaches
this common understanding.
7.2.7 Teaming in an Interdisciplinary Environment
Students taking classes within their field of study often work alone, or in teams
with others in their field. Few if any entering students in our BID class have shared
a course with someone outside their major. Hence, it may be difficult initially for
students to recognize the value of knowledge and approaches outside their
discipline. This can be abetted by the institutional persona that encourages divi-
sions in the perceived utility of different fields of knowledge. For instance,
engineers at a technology institute may think their expertise is more valuable than
others. An appreciation for everyone’s talent needs to be nurtured throughout the
time the team is working toward a common goal.
7.2.8 Maintaining Equal Engagement Throughout

the Process
The roles of each discipline may change throughout the course, depending on the
stage in the design process and a design’s specific requirements. Initially, emphasis
is placed on biological knowledge since the teams have to select an organism and
understand how the biological system works. Once the teams enter the design
process, everyone is actively engaged because BID is unfamiliar to most students,
with more weight placed on the biologists to find, understand, and explain solu-
tions. The engineers are the most engaged when there is a required feasibility or
performance assessment, work which biologists are not as experienced to perform.
Under deadline pressure to complete a design, team members who cannot con-
tribute directly can feel marginalized or devalued.
7.2.9 Evaluating Designs
A good design for our purposes must simultaneously satisfy the following criteria:
functionality, potential market, manufacturability, novelty or competitive advan-
tage, and reasonable cost. Although different ways of teaching BID may not
emphasize all of these criteria, student designs become amorphous and speculative
without a focus on functionality, novelty, and manufacturability, whereas failure to
consider market and advantage results in designs that do not solve real problems or
do so in a way little different from current designs. Challenges occur because: (a)
students may not be familiar with using some of these criteria in their design
analysis; (b) applying some of these criteria may require students to apply quan-
titative methods outside of their domain; and (c) students have trouble balancing
conflicting criteria. These challenges are exacerbated by the profusion of possible
quantitative analyses that could be performed. It is difficult for students to select
the few analyses that are crucial.
7.3 Summary of Core Development Areas
In this work, we present our efforts to identify and solve problems in the teaching
of BID, as embodied in the following five areas.
160 J. Yen et al.
7.3.1 Content
In moving away from a lecture-based course to a problem-based collaborative

learning environment, we need to balance between providing knowledge about
biology, engineering, and design (content) and hands-on practice engaging in the
BID process.
7.3.2 Representation and Tools
For students to find and learn about biological systems, to communicate that
knowledge to people from different backgrounds, and to apply that learning to ill-
defined engineering problems of their own making, we must equip them with tools
that facilitate understanding and communication and focus attention on aspects of
systems that are important for design.
7.3.3 Design Process
We have implemented several design process formalizations to contend with the

special needs of a process focused on analogical design. As more experience is
gained in teaching BID, we see similarities to and differences from more stan-
dardized design process approaches. One key difference is solution-based design:
this process starts with solutions presented in natural biological systems and
translates appropriate functions to solve design challenges in an inventive fashion.
7.3.4 Design Evaluation
Students produce conceptual designs in this course. Given the need to teach the
process in 15 weeks, and the emphasis on student-identified design problems,
building and testing a prototype to demonstrate feasibility is not possible. Nev-
ertheless, even for conceptual designs, students must convince themselves, the
class, and instructors that the design could work and would have some advantage
over existing products. Throughout the semester, examples of quantitative anal-
yses give students practice in addressing issues that often crop up in BID, such as
scaling, materials selection, and environmental impact.
7.3.5 Interdisciplinary Teaming
BID can serve as a catalyst for innovation because of the mix of disciplines. But
just throwing the students together would not lead to success. With different
cultures, values, processes, and vocabularies, as well as different technical back-
grounds, we have learned that a number of different teaming techniques are nec-
essary to ensure proper communication, balance, and respect in a properly
functioning team.
In the next sections, we document the challenges faced in each core develop-
ment area.
7.4 Content
Given the multiple course objectives, balancing content is a difficult task. We must
communicate a breadth of biology and engineering knowledge, accounts, and tools
for design processes and facilitate interdisciplinary communication. Additionally,
there must be sufficient time for the students to practice with the tools they have
learned. We describe specific elements of content that we have identified to help
meet those challenges and ways to maximize the effectiveness of this content to
avoid overloading the students (Table 7.1).
7.4.1 BID Stories
To maintain student enthusiasm, we began every class with what we call a BID-
wow story: an account about an exciting, innovative bioinspired design. These
consist of examples such as: the whale fin inspired windmill blade (Miklosovic
et al. 2004) which is more efficient, quieter, and able to capture wind energy at
lower wind speeds; the slime mold that connected nutrient sources placed in a petri
dish in the same pattern as major cities around Tokyo and grew a transport system
as efficient as the Tokyo railway (Tero et al. 2010); the spacious, transparent
cabins of the 2050 AirBus concept plane (http://www.airbus.com/innovation/
future-by-airbus/) with a bionic structure mimicking bird bones to make planes
lighter and stronger; the butterfly-inspired sensor that responds to different
chemical vapors using the ordered arrays of iridescent scales to outperform
existing nano-engineered photonic sensors (Potyrailo et al. 2007), or; the cat’s eye
retro reflector (Percy Shaw’s patent No. 436,290 and 457,536) that reflects light
back to its source with minimum scattering, similar to eye shine created by the
tapetum of a cat’s eye. These fascinating stories are told as soon as the class bell
rings, encouraging the students to be in class on time, and keeping them focused
162
Table 7.1 Five course elements (down) pertaining to content that address the 9 key challenges (across)
Content Searching Understanding Identifying Analogy Communicating Communicating Teaming in an Maintaining Evaluating
for biological and mapping across discipline complex system interdisciplinary equal designs
biological systems understanding and boundaries knowledge environment engagement
systems good design transfer throughout
problems the process
BID stories X X
Case studies X X X X X X X X X
Found objects X X X X X
Evolution X X
Focused reading X X
J. Yen et al.
on the thrill of invention. While these news stories pop up frequently, they do not
provide the details of the source of inspiration nor the process of transfer. For this,
we turn to case studies.
7.4.2 Case Studies
Case studies presented by local experts provide information that can meet a variety
of challenges from increasing subject knowledge to developing design skills and
can be an integral part of any BID class. Locally, we have many to choose from, and
the research described by familiar and respected teachers at one’s home institute
adds to the impact. Many of these bioinspired designs have taken years of research
and development. As a result, the lectures given by the BID practitioners have a
wealth of very detailed knowledge that students sometimes find difficult to absorb.
Although the stories are all astounding and fascinating, what should/could a student
get from this? One strategy is not to be concerned about content but to use these
meetings to give the students the chance to meet and talk to the people behind the
design, and we did that initially. We invited a parade of professors who used two of
the 30 class periods (nearly 3 h in all), sharing their excitement about the process
and product. This was great motivation, and students still rate these kinds of expert
lectures as a favorite part of the class, but it did not teach the student ‘‘how to.’’
Over the years, we reduced the number of lectures and the length (45 min plus time
for discussion) and provided the following guidelines to the lecturer:
1. Describe the key feature of the natural system that provided your inspiration. In
particular, we asked experts to focus on 3 things regarding their inspiration.
What were the structures that come from the biological system? In this case,
structure refers to the system components that perform the function of interest
in the system. Why did this function help the organism survive? How did the
organism achieve that function? This is the deep biological knowledge.
2. Decompose the challenge you faced into its functions and describe the function
that your design addressed. What structures are needed for this function, what
use is this function to humans, how do existing solutions achieve this function,
and what are the limitations of existing solutions? This is the deep engineering
knowledge.
3. How did you translate the biology into the engineered design? This is the
design process. From this, we saw how analogical reasoning was a key element
in this translation process.
4. Provide the 3 best articles on your BID, one on the biological inspiration, one
on the details of the specific biological mechanism of interest, and one on how
the biological system were translated into an engineering design that worked.
This teaches the students how to read scientific literature.
164 J. Yen et al.
This format is consistent with our emphasis on system components, interac-

tions, and functions as tools to help students define effective analogies [see Sect.
7.6.6, Structure–Behavior–Function (SBF), and structured representation for BID
(SR.BID)]. This narrative produces a balanced mix of both biological content and
design process and leaves the students wanting to hear more about what the
scientist–engineer did. There were always a handful of energetic students who
asked a lot of good questions and we would have a lively discussion that spilled
out into the hallway after class. One of the professors remarked: ‘‘I got more
questions in this class than when I gave the same seminar to faculty and grad
students in my discipline!’’ However, in our experience, these case studies (even
when presented by an individual involved in the research) are not sufficient for the
students to actually grasp the BID process, even when presented in the uniform
way described above. Moreover, the journal readings and technical depth of the
papers were clearly overloading the students with scientific publications that were
difficult to read and understand. Despite this being a favorite activity of the stu-
dents, we limit these lectures to 4–6 per semester, and instead, we focus on more
hands-on learning strategies.
7.4.3 The Found Object Exercise
Observing, experimenting with, researching, and describing the functions of bio-

logical objects is a central curricular element that meets a variety of challenges,
but is particularly well suited to increasing the ability to search biological systems
and increasing interdisciplinary communication. Students are asked to go outside,
find something in nature, play with it until something intriguing is noticed, then
find an article that explains how the natural system works (Yen et al. 2011).
Through this exercise, we want the students to reconnect with their natural sur-
roundings, spend enough time interacting with nature to find something that is
marvelous, and then deepen their knowledge by reading about it first in general
biology/ecology/behavior texts that point to good articles. This develops a sense of
connection to nature or biophilia (Wilson 1984). The objective of this part of the
exercise is to figure out what search strategy to use to find the information needed
for BID and how to get this information out of articles from the primary literature.
We use their interaction with nature as the stimulus for deepening their biological
knowledge base. In class, the team members share what they found and decide
who has the best found object. That person tells the class about it using a succinct
knowledge representation template (explained in the representations section). For
a class of 40 with 8 interdisciplinary teams of 5, it takes the entire class period to
do this. At the end, we review the 8 best objects and discuss whether enough was
presented to understand how the system works. We found that when students
became facile with the knowledge representation template, they could zero in on
the key function of interest without getting caught up in all the other fascinating
details inherent in complex biological entities. This helps the finders hone their
analytical skills, focusing them on only the most salient features for design. This
also helps the speakers to hone their communication skills, conveying the key
principles that a biologist and an engineer need in order to see the value of the
biological strategy of interest. Additionally, through active participation in the
process, these exercises add a breadth of amazing local biological systems to each
student’s repertoire of biological knowledge, strengthening their appreciation for
the local natural environment around them.
7.4.4 Perspective on Evolution
At the other end of the spectrum from the problem of providing sufficient depth,
we have the problem of providing an overall perspective on BID. Engineering
students can find the variety of biological organisms and functions to be bewil-
dering. Biology students can have difficulty establishing and maintaining focus on
biology in the context of design. We provide a lecture on evolution early in the
course to help students gain a perspective on biology in the context of design. This
lecture includes the concepts of common ancestry and convergent evolution,
multi-function optimization, and local versus global optimization. This lecture has
been given every year and receives consistently positive comments from students.
It appears to help them understand differences between evolution as a design
process and intentional design, which enables a more sophisticated view of how to
search and evaluate potential biological solutions.
7.4.5 Focused Readings in BID
One of the nagging struggles of this design class is to provide the correct depth of
information at the right time to students, and to do it without overburdening the
students, or suppressing their motivation to continue reading throughout the
course. Early iterations of the class asked students to find technical, academic
papers for biological systems of interest. Considering that a student may do five
found object assignments, five case study lectures, and must research up to ten
biological systems, a requirement consisting of two documents per assignment
results in a massive overload of technical documentation (up to 40 technical
papers!). Adding requirements on top of the technical reading, such as formulating
summaries or key questions for presenters only exacerbated the situation. The
problem still remains: how do we ensure that students engaged in BID conduct
deep explorations of a select few organisms, while getting a broad range of
exposure to many and in such a way that given some biological system, they are
capable of acquiring the knowledge on their own?
One means to address the breadth and depth issue was the use of a general
purpose textbook, such as Vogel’s (2000) Cat’s Paws and Catapults, instead of
166 J. Yen et al.
technical papers prior to the case studies. The textbook is written to be broadly
applicable, yet provides sufficient depth to highlight the key principles as well as
the challenges of applying those principles. By aligning the themes in the textbook
chapters with the themes of the case studies, we provide salient real-world
examples to reinforce the reading.
The problem of finding and understanding deep technical references for a few
systems, balanced against the need to understand, for example, found objects,
remains an unsolved challenge in the class. Simply put, students give higher
priority to generating exciting designs than to time-consuming deep reading.
7.5 Representations and Tools
As we have emphasized, BID requires students to find and learn about interesting
biological systems, to communicate that knowledge to people with backgrounds
different than their own, and to apply that learning to ill-defined engineering
problems of their own choosing. In this fast-paced, novel context, students become
easily overwhelmed and unable to identify clear learning objectives. Student
presentations of biological systems found locally (found objects) in years 2006 and
2007 best exemplify these early struggles. When asked to summarize the most
interesting aspects of found objects: (a) discussion is dominated by the structural
details of biological objects; (b) students superficially associate a wide variety
functions to the design; (c) though we emphasize mechanistic explanations, they
are rarely provided; (d) if mechanistic explanations are offered, they are often
provided by reference to a common sense analogy (often incorrectly); and (e)
technical explanations employing terms from one domain are not understood by a
majority of students from a different domain. In this context, what specific tools or
representation strategies can help focus students on aspects of systems that are
important for design? Table 7.2 below lists five that we have used in our course. In
Sects. 7.4.1–7.4.4, we describe the first four. We postpone the description of the
fifth, SR.BID, until Sect. 7.6.
7.5.1 Search Strategies
Since our initial classroom deployments, we realized that students are challenged
with converting engineering-centric design problem language to the corresponding
biological terms necessary to find biological systems in the external information
environment. This is a particular problem when classes are dominated by engi-
neers who have had little basic biology and are unfamiliar with potentially relevant
biological systems. Starting in 2006, we offered three useful techniques for helping
students to identify keywords that might lead to fruitful searches. These techniques
are documented in (Yen et al. 2011), but in brief, we asked students to: (1) identify
7
Table 7.2 Four course elements (down) pertaining to representation and tools that address the 9 key challenges (across)
Representations Searching Understanding Identifying Analogy Communicating Communicating Teaming in Maintaining Evaluating
and Tools for biological and mapping across complex an equal designs
biological systems understanding and discipline system interdisciplinary engagement
systems good design transfer boundaries knowledge environment throughout
Search strategies X
SBF X X X
DANE and biologue X X X X
Functional X X X
decomposition
Adaptive Evolution of Teaching Practices in Biologically Inspired Design
167
168 J. Yen et al.
key functions of interest; (2), invert the function to reveal generic principles (e.g.,
heating and cooling both concern principles of heat transfer); and (3) identify
extreme environments in which high-performing biological systems might be
found. Each of these techniques provides one or more keywords that may be useful
when browsing large collections of literature for biological systems. Regardless of
the technique used, search remains challenging; students report that as much as a
quarter of their design time is spent searching for information on biological
organisms and functions (Vattam and Goel 2011).
7.5.2 Structure–Behavior–Function Analysis
As noted above, students have trouble articulating the properties and functions of
biological systems in a way that facilitates abstraction and transfer of design
principles. Our first iteration of this course in 2005 lacked any advice about
representation, which diminished our ability to teach students how to transfer
knowledge from one domain to the other. The basic problem is that both engi-
neering and biological systems can be described in a variety of ways, which
obscures the fundamental cognitive step of transferring biological mechanisms as
solution principles.
In 2006, we introduced a single lecture on SBF analysis (Bhatta and Goel 1997;
Goel et al. 2009), which is grounded in cognitive theories of systems thinking.
Structure–Behavior–Function
• Structure, behavior, and function form an abstraction hierarchy for systems
thinking; behavior is an intermediate level of abstraction between structure and
function.
• Structure specifies the components of the system as well as the connections
among them. For example, the structure of the electrical circuit in an ordinary
household flashlight comprises of an electrical battery, a light bulb, a switch,
electrical connections among the battery, bulb and switch.
• Behaviors specify the causal processes occurring in the system. For example, the
behavior of the flashlight is that when the switch on the flashlight is pressed,
current flows from the battery to the bulb, and the bulb converts electrical into
light energy.
• Functions specify the outcomes of the system. For example, the function of the
flashlight is to produce light when the switch is pressed.
• Behaviors provide causal mechanistic explanations of how the structure of the
system accomplishes its functions. For example, the behavior of the flashlight
explains how its structure accomplishes its functions.
• A behavior of a system specifies the composition of the functions of its sub-
systems into the system functions. For example, the behavior of the flashlight
composes the functions of its components—the battery, bulb, and switch—into
the function of the flashlight.
• A subsystem or component of a complex system can itself comprise a system

and thus have its own SBF model. Hence, SBF models of a system can have a
hierarchical structure. For example, consider the system of the basilisk lizard,
which is well known for its ability to run across water. If the function (F) of
interest of the basilisk lizard is ‘‘run on top of water,’’ one can consider the
opposing limbs, tail, and wide flat feet as part of the structural (S). The way in
which the feet move in opposition are counter-balanced by the tail, and how the
feet slap the water generating lift, then extend down and back creating more lift,
thrust and a pocket of air in the water, and are then withdraw up and out through
the air pocket could be considered the behavior (B) that generates the ‘‘run on
top of water’’ function. One could consider the muscular-skeletal system of the
legs as a subsystem of this system used to create a subfunction ‘‘generate
movement of legs’’ which causes the higher-level ‘‘run on top of water’’
function. Likewise, one can consider the form the foot takes throughout the
process as another subfunction, ‘‘change foot surface area.’’ In this way one can
decompose the ‘‘run on top of water’’ function into a number of subfunctions,
including ‘‘generate movement of legs’’ and ‘‘change foot surface area,’’ each of
which could entail another SBF model. Similarly, one can consider the function
‘‘run on top of water’’ to be part of the function ‘‘escape predator’’ showing that
one can navigate both up and down the levels of functional abstraction in
the SBF model hierarchy. This kind of hierarchy will be discussed further in
Sect. 7.5.4.
The origin of SBF analysis lies in Chandrasekaran’s functional representation

scheme (Chandrasekaran 1994; Chandrasekaran et al. 1993). Other researchers
have developed similar cognitively oriented approaches to thinking about complex
systems, for example, Rasmussen (1985). Gero and Kannengeisser (2004) describe
the design process itself in terms of function, behavior, and structure. Erden et al.
(2008) provide a recent review of functional modeling. Note that in SBF analysis,
functions are mental abstractions chosen by the modeler, and not intrinsic to the
complex system. In the case of engineering systems, a functional abstraction
corresponds to an intended output behavior of a system, subsystem, or component.
However, since functions are mental abstractions, we can also use SBF modeling
to model natural systems, including biological systems, such as the human heart,
and ecological systems, such as forests. Even more so than engineered systems,
natural systems exhibit layers of varied functionality at different scales, feedback
loops, and other types of causal processes that characterize complex systems.
Recently, there has been considerable interest in the use of SBF modeling in
science education. Goel et al. (1996) proposed the use of SBF models for
explaining complex systems in science education. Ebert-May et al. (2010) and
Speth et al. (2011) found that construction of SBF models in college-level courses
helped expose students’ misconceptions of ecological and biological systems,
respectively. Chan et al. (2010) found that high-achieving students in a college-
level course on biomedical engineering paid more attention to behavior and
function than did low-achieving students, and that attention to behavior and
170 J. Yen et al.
function improved student performance. Helms, Vattam and Goel (2011) found
that SBF models of biological systems enable complex inferences that were not
readily enabled by textual or diagrammatic representations of the systems. Vattam
et al. (2011) discovered that use of SBF modeling for learning about ecosystems in
middle school science classes resulted in statistically significant improvement in
students’ understanding of the structures, behaviors, and functions of aquatic
ecosystems. Silk and Schunn (2008) summarize some of the benefits of SBF
analysis in science education.
In 2007, we introduced SBF analysis as a framework for organizing found
object exercises. Students were asked as part of the found object homework
assignments (Sect. 7.4) and in their discussions to (a) focus on a single function of
the organism in question, (b) identify the structures relevant to accomplishing that
function, and (c) provide a behavioral explanation for how those structures give
rise to the function. Instructors facilitated these discussions as necessary to guide
students. (In the SBF vocabulary, behavior is synonymous with causal mechanistic
explanation.)
As expected, students discussed structure at length, although they were unable
to limit themselves to the discussion of a single function. As noted earlier, SBF is a
hierarchical representation and systems are naturally functionally hierarchical. As
a result, it was difficult for students to maintain a single level of functional
abstraction during their discussions. Often students travelled ‘‘up’’ the functional
hierarchy attempting to explain why the organism performed the function in
question such as reproduction, survival, and escape from predators. The result was
discussions about many high-level functions that lacked in detail. Less frequently,
students travelled ‘‘down’’ the functional hierarchy, explaining a small portion of
how the organism performed a function. These discussions usually resulted in very
detailed, technical low-level discussions that only a few students could follow.
One must continually emphasize to the students that, while the number of levels in
a decomposition is very large, functions expressed at much lower or higher levels
than the original problem may not always be useful for the purpose at hand,
because they introduce constraints (lower levels) or goals (higher levels) not
present in the initial problem definition. In addition to traversing levels of
abstraction, students frequently confused the different senses of the word
‘‘behavior.’’ Students often associate behavior with higher-level actions at the
organism level, for example, mating behavior, territory marking behavior, seeking
shelter from the heat rather than addressing the causal mechanisms, as this word is
used in the cognitive sciences (Gero and Kannengeisser 2004; Goel et al. 2009).
To simplify the vocabulary, in 2008, we changed the SBF vocabulary to a
What-Why-How vocabulary, mapping ‘‘What’’ to ‘‘Structure,’’ ‘‘Why’’ to
‘‘Function,’’ and ‘‘How’’ to ‘‘Behavior.’’ This was an attempt to both remove the
ambiguous interpretation of ‘‘behavior’’ and to formalize the levels of functional
abstraction. Functional abstraction was considered in terms of ‘‘why’’ moving up
the hierarchy (more abstract, superfunctions), and ‘‘how’’ moving down (more
detailed, subfunctions). Again, students were asked to describe all biological
systems in these terms, both conversationally and in formal homework
assignments and design reports. Despite removing potentially confusing SBF

language, students continued to describe these systems in a way that enabled them
to avoid providing a mechanistic explanatory account. Table 7.3 gives an example
from a midterm presentation in 2008.
It is illuminating to characterize the failure in Table 7.3 as a traversal in
hierarchical levels. A hierarchical level corresponds to the vertical location of a
function in a problem decomposition. Going down one level means to think about
the subfunctions of the function under focus. Going up one level is to consider to
what the focal function directly contributes. For example, if irrigation is the focal
function, then acquiring, transporting, and dripping or spraying water would be
subfunctions one level below. Sustaining plant growth would be a function one
level higher, perhaps at the same level as harvesting; feeding the hungry would be
a function considerably higher in the hierarchy.
In the case shown in Table 7.3, we see ‘‘What’’ addressing a function, ‘‘Why’’
addressing a higher-level function, and ‘‘How’’ addressing a structure, in this case
color patterns. However, ideally, the ‘‘What’’ would address the components of the
solution (e.g., structural color patterns), the ‘‘Why’’ would address the functions of
solar absorbance and energy capture, and the ‘‘How’’ would explain the mecha-
nism by which structural color patterns cause solar absorbance and energy capture.
The example in Table 7.3 uses ‘‘Why’’ to capture a function several hierar-
chical levels higher than the one that is really being considered. That is, while
maintaining body temperature may be the top-level function or goal, it is several
levels displaced from energy absorbance. This suggests an overloading of the term
‘‘Why’’ as both ‘‘the function of interest’’ and ‘‘the reason for the function of
interest.’’
In our experience, students need a large amount of practice with these repre-
sentations to employ them correctly. As a rule of thumb, we have found that
restricting the analysis to one to three hierarchical levels above or below the
‘‘What’’ function is useful to focus the student’s attention on the right structures,
functions, and mechanisms. Levels above this cutoff often take the students to the
ultimate evolutionary objective of a given biological ‘‘solution,’’ which may not
match the engineering problem for which a given function may be useful. For
example, the ultimate evolutionary objective, to survive, is so universal that it
gives no additional guidance in the search for connections between biology and
engineering. Going too many levels down may introduce constraints specific to the
particular way the biological function is achieved, and which may not be relevant
if the goal is to abstract the function rather than copy precisely the mechanism.
Table 7.3 Using What– Iridescent butterfly wings

Why–How vocabulary fails
to generate a mechanistic ‘‘What’’ Solar absorbance and energy capture
explanatory account ‘‘Why’’ To maintain body temperature
‘‘How’’ Structural color patterns
172 J. Yen et al.
Despite student problems with correctly identifying answers to the ‘‘why,’’

‘‘how,’’ and ‘‘what’’ questions, we have found the SBF schema helpful because it
helps students ask useful questions in trying to understand complex biological
systems. Such hierarchical analysis is used to decompose complex concepts into
manageable pieces of information. Thus, SBF has become part of the language of
discourse in the BID class.
7.5.3 DANE and Biologue
We also have experimented with interactive tools that use SBF models to help
enhance student understanding of biological and engineering systems. Given the
importance of knowledge representations and interactive tools, it is not surprising
that recently there has been enormous amount of work on devising representations
and tools to support BID (Biomimicry 3.8 Institute 2008, 2009; Bruck et al. 2007;
Chakrabarti et al. 2005; Chakrabarti and Shu 2010; Cheong et al. 2011; Chiu and
Shu 2007a, b; Nagel et al. 2008, 2010; Sarkar and Chakrabarti 2008; Sarkar et al.
2008; Sartori et al. 2010; Shu 2010). These tools differ in their representations of
biological and engineering designs, strategies for searching for a biological solu-
tion potentially relevant to a design problem, the (implied) process of BID,
evaluation of design solutions, and so on. However, these representations and tools
for BID are normative and prescriptive. We believe that it is important to situate
the development of representations and tools in real-life contexts. The BID course
has provided a motivation and a context for using, often in new ways, existing
knowledge representations such as SBF, and also for developing and evaluating
new representations such as SR.BID (see Sect. 7.6.6) and new interactive tools
such as DANE (Goel et al. 2012; Vattam et al. 2010b) and Biologue (Vattam and
Goel 2011) (see below).
DANE provides a digital library of SBF models of biological and engineering
systems, as well as tools for constructing SBF models of new systems. We
introduced DANE into the BID class in 2009. Some students in the BID class
found DANE useful for making sense of complex biological systems and con-
structing a conceptual understanding of the systems. (DANE can be downloaded
from \http://dilab.cc.gatech.edu/dane/[.)
Biologue enables students to annotate and share documents on biological sys-
tems as a team, to tag the documents with SBF models, and to search for additional
documents based on SBF tags. We introduced Biologue into the BID class in 2012.
Some students in the BID class found Biologue useful for online annotation and
sharing of biology articles. In controlled experiments, we discovered that Biologue
enables subjects to more easily and accurately locate relevant biology documents
online (Vattam and Goel 2011). While initial results from DANE and Biologue are
promising, identifying ways to search for interdisciplinary analogies remains an
open research area (as this volume attests).
7.5.4 Functional Decomposition
In addition to the SBF analysis introduced in 2007, we began introducing

decomposition diagrams in the class. Our early cognitive studies (Vattam et al.
2007) provided some evidence that analogical matching between problems and
biological solutions was taking place functionally, but implicitly so. That is to say,
we had no formal method of determining the quality of a match between the
problem and the biological solution identified to help solve it. If we could for-
malize the decomposition of both problem and solution functionality, we would
provide a more formal method for making and evaluating the connection between
problem and solution.
In 2007, we introduced functional decomposition (Dym and Brown 2012;
French 1996; Pahl et al. 2007; Simon 1996). An interactive lecture on problem
decomposition was provided where the class participated in a group decomposition
exercise for designing a search and rescue vehicle that could walk on uneven and
shifting surfaces, such as sand. Figure 7.1 shows the decomposition that was
created interactively with students during that lecture. Assuming functional
matching was the primary index used to retrieve biological solutions from mem-
ory, a diagram such as this should provide a number of functions, each of which
may lead to an array of biological solutions that could be applied to solve one or
more of the subfunctions identified. Thus, multiple biological solutions could be
used in solving a single problem. This phenomenon, termed compound analogical
design, is well documented in class (Helms et al. 2009; Vattam et al. 2010a, b).
Students were asked to provide similar ‘‘solution-neutral’’ decompositions of
their problem for all presentations and reports, as well as to justify their analogies
by matching the functions provided by a biological solution and the function in the
decomposition. Figure 7.2 is a typical example of a student’s problem decompo-
sition. We found students consistently tailor these decompositions to the biolog-
ical, and sometimes technological, solutions that they are already considering.
That is, the functions that appear in the functional decomposition are nearly always
aligned directly with a system that the students are considering. It is unclear
whether this is selective pruning of the decomposition so that there are no func-
tions for which there is no solution, or whether the decomposition is a result of a
bottom-up approach where students fit functions from solutions to the problem
decomposition. This bottom-up composition suggests that functional decomposi-
tion of a problem is not only formed in a top-down manner, but also may be
partially formulated based on the availability of solutions.
7.6 Design Process
Students in BID class are asked to invent their own design challenge and to generate
a creative, biologically inspired conceptual design that solves that challenge.
174 J. Yen et al.
Fig. 7.1 Problem decomposition of a search and rescue vehicle to transport over uneven and
unstable ground
Undergraduate students enrolling in a BID class may enter with little or no formal
design process training. Even for engineering students, design is often in the context
of a problem with very specific functional requirements, that is, the problem and
evaluative criteria often are very clear. It is important to monitor the typical design
experience of the student pool to determine how much to coach students through the
process, particularly during problem definition.
Fig. 7.2 Decomposition of the problem of maintaining a comfortable home temperature

Generating a new design problem, as well as learning a process for solving

design problems adds an additional level of complexity to the design process. In
early iterations of the class, we used a typical problem-driven design process, but
we recognized that student designers spontaneously adopted a second design
process, solution-based design. After close study of this process, we recognized
that not only was the second process useful for students, but it also appeared to
represent the more successful modern design approach outside of the classroom
(Vattam et al. 2007). Beginning with the recognition that we needed to teach not
one, but two design processes, we implemented a number of strategies for assisting
students in structuring their design process (Table 7.4).
7.6.1 Problem-Driven and Solution-Based Design
Perhaps the most significant reworking of the class involved the organization of
the class into sections representing these two dominant process modes for BID.
Initially, we instructed students to find a problem, find a biological source, and
apply the source to the problem to generate a solution. With regularity, half of the
design teams would follow the problem-driven approach, and half of the design
teams would instead fixate on an interesting biological solution and then find an
appropriate problem to solve. Since each process seemed useful in different cir-
cumstances, we decided to formalize the different approaches and allow students
to experience both.
On the very first day of class, students are now introduced to dozens of inter-
esting biological systems in our ‘‘biology auction’’ exercise. The auction engages
student’s curiosity and imagination with a wide range of possible biological sys-
tems that can serve as design inspiration, either directly or indirectly. In addition,
the immediate emphasis on biological systems reinforces the validity of biological
knowledge and engages the biologists.
Over the next 6 weeks of the class, students identify one interesting biological
system and figure out a means for using the interesting principles of that solution
to solve a human-scale problem. We teach this process in class more formally as
solution-based design and scaffold the process with exercises meant to help stu-
dents (a) understand the mechanism of interest in their biological system, (b)
abstract the mechanism used in their system, (c) identify a number of problems for
which their system may provide a solution, and (d) formally analyze the analogy
between their system and the problems they propose to solve in order to identify
the best solution-problem match. Only in weeks five and six are students asked to
begin producing conceptual designs.
We institute a more compressed problem-driven design cycle during weeks
seven through ten. This begins with students: (a) defining a problem; (b)
abstracting the problem; (c) finding biological solutions to the abstract problem;
and (d) formally analyzing the analogy between their problem and the solutions
they propose will solve their problem. Students craft a design solution in the last
176
Table 7.4 Six course elements (down) pertaining to the design process that address the 9 key challenges (across)
Design process Searching Understanding Identifying Analogy Communicating Communicating Teaming in an Maintaining Design
for biological and mapping across discipline complex system interdisciplinary equal evaluation
Problem-driven and
solution-based
processes
Source breadth X
Problem definition X
Problem focus X X X
Project format
Analogical evaluation and X X X X
SR.BID
J. Yen et al.
week of this second cycle. This process of problem-driven design is an instanti-

ation of the more basic cognitive process of analogical reasoning (Clement 2008;
Dunbar 2001; Gentner 1983; Gick and Holyoak 1983; Goel 1997; Hofstadter 1996;
Holyoak and Thagard 1995; Keane 1988; Kolodner 1993; Nersessian 2008).
Solution-based design appears new and different from the perspective of design
theory, all of which is problem-driven (e.g., Dym and Brown 2012; French 1996;
Pahl and Beitz 1996). Thus, the BID course acts as a research laboratory for
developing, identifying, and studying new BID constructs and processes.
With the remaining time, we allow students to continue to develop their ideas
for either the first or second project. This allows students to experience both
solution-based and problem-driven processes, while balancing the need to get deep
into the design, permitting students to deal with the complex problems that come
with more detailed design.
7.6.2 Source Breadth
The challenges associated with defining a problem, and finding and understanding
biological systems, often result in students exploring few potential biological
solutions (Wilson et al. 2010). This compounds the tendency of all novices to
engage in design fixation. Students in early iterations of the course were required
to explore only one problem and solve the problem with one or more biological
solutions. Student designers in this environment investigated between two and ten
biological solutions, while applying one or two biologically inspired mechanisms
to solve the problem. However, in about two-thirds of the design projects, students
fixated on the first biological system they encountered and only superficially
explored other systems. Thus, in early classes student design teams shallowly
investigated a handful of biological systems and came to deeply understand one or
maybe two biological systems.
To counteract the effects of solution-fixation, student design teams are now
required to report on at least thirty biological systems throughout the course. Each
student examines a minimum of five biological sources before selecting one for
their solution-based design, which means for a team of five, the entire team learns
about 25 biological systems. These systems can be related to each other via con-
vergent evolution, phylogeny, or exaptation. Furthermore, during problem-driven
design, each student is required to report on five biological sources, or 25 natural
systems for the team. In their final team design reports, a deep analysis is required
of at least five of these systems. These systems may or may not overlap with
systems discussed in their five found object exercises, again including up to twenty-
five additional systems per team. Thus, student teams may explore and share
knowledge about as many as seventy-five biological systems over the course of a
semester. Furthermore, because students are trained in formally representing these
systems using the function-based representation tools we provided (see Sect. 7.5),
178 J. Yen et al.
exploration of these systems is a structured process that (in theory) allows students
to functionally index each system in their own memory for later retrieval during
design episodes.
7.6.3 Problem Definition
In a semester-long design project, students (who again, are largely naive with
respect to open-ended design problems) spend approximately half of the semester
learning about and grappling with the complexity of their own design problems.
Students seem to be motivated by tackling complex, often topical, issues such as
oil-spill cleanup or eliminating traffic congestion, for which they may have little
familiarity. In a class where students are expected to learn an incredible breadth of
content and process knowledge, our task is motivating students to find interesting,
challenging problems, without letting the definition of the problem itself become
the core challenge.
Problem discovery and definition is usually the first step in the design cycle
(Dym and Brown 2012; French 1996; Pahl et al. 2007). Even when the design
cycle is solution-based, problem definition is quickly derived by working back-
ward from a potential solution. Moreover, problem definition is inherently itera-
tive. We have found, for example, that 70 % of the function requirements
considered during the semester are discarded by the final design and as many as
one-third of the final function requirements were identified during the final few
weeks of a semester-long design.
We provide three scaffolds for students to help with problem definition. First,
we give a lecture early in the semester that is inspired by Ron Bills, the CEO of
Envirofit, entitled ‘‘What makes a problem a good problem?’’ this lecture pro-
vides an answer in terms of three W’s: what stinks, who cares, and what are you
going to do about it? This lecture sticks in the students’ minds and keeps their
focus on the practical. To reinforce this perspective, student design teams are
required to answer the three W’s during their preliminary design evaluation.
Second, we provide the problem-structuring tool (SR.BID) described in Sect.
7.6.6. These tools provide a handle for students to gain traction on defining their
problems. Third, we embed problem definition formally in many of the design
assignments, forcing students to reflect, to explicitly represent their design
problems, and come to a shared team understanding of them. Our observations in
2011 suggest that these interventions: (a) reduce the time students spend on
problem development; (b) reduce the problem scope; and (c) enhance the level of
detail, especially with respect to the number of performance criteria and speci-
fications that are considered. We believe projects in 2011 were among the most
practical designs produced since we began teaching this class, with no sacrifice in
perceived creativity.
7.6.4 Problem Focus
The extent to which student problems should be defined by the instructor versus by
the students involves some trade-offs that may strongly affect student performance.
We have tried limiting problems to specific areas (for example, sustainable housing)
and allowing students to choose their own problems. Constraining student problems
to specific areas potentially allows student groups to share information and come to a
deeper understanding of the problem as a result of their joint efforts. It also may
ameliorate some of the difficulties associated with problem formulation, and iden-
tifying and understanding biological systems. However, many students express
disappointment that they were not allowed to choose problems with which they were
comfortable. Moreover, constraining the problem sometimes limited the opportu-
nity of certain engineering disciplines to participate. We also learned that unless
problems are highly constrained, student teams were able to find a wide variety of
problems such that benefits of shared knowledge were weakened considerably.
Ultimately, in our highly interdisciplinary environment, allowing students to self-
identify problems leads to stronger application of engineering knowledge and
student motivation and is preferred even though problem definition remains chal-
lenging. Constraining problems to certain domains may be more productive when
students share a greater amount of disciplinary knowledge or attitudes.
7.6.5 Project Format
The project format can serve a variety of goals, some of which may be important,
but are ancillary to BID practice. For instance, it is common practice at Georgia
Tech for instructors to require capstone design projects to be sponsored by industry
partners, where industry partners then can participate in the project as real-world
customers. We initially framed the project format in an entrepreneurial context,
that is, student design teams were expected to pitch their designs to appeal to a
group of venture capitalists. While students found this quite motivating, they also
spent a lot of time on branding, marketing, and selling their concept, rather than
understanding and articulating the underlying principles. Still, the emphasis on
design feasibility is important. Thus, we currently frame the design process in a
more pragmatic sense. We ask students to prepare a presentation that validates the
design concept in a way that would convince us (the instructors, and other guest
evaluators) to invest in creating a prototype of their design.
7.6.6 Analogical Evaluation and SR.BID
Since BID is a cross-domain activity between biology and engineering, analogy

making is a cornerstone process of BID but formalizing instruction for analogical
180 J. Yen et al.
mapping continues to challenge us. Students asked to find a biological analogy for
a problem like ‘‘create a device for collecting water samples at a fixed depth
underwater in a lake,’’ can almost immediately produce answers like ‘‘puffer fish,’’
‘‘pelican,’’ and ‘‘whale.’’ Remarkably, students generate these analogies naturally,
without instruction, and within minutes given a particular design challenge. When
asked to describe why the analogy is a good fit, however, students require much
more time and often are at a loss for a description that they themselves find
satisfactory. When asked to evaluate these analogies for applicability or ‘‘goodness
of fit’’ for generating a solution to the design problem, students can often cate-
gorize one analogy as better than another, but they are generally incapable of
producing a consistent rationale for why one is better than the other. These
observations have led us to seek better ways to focus student attention on
appropriate analogical mapping and evaluation of goodness of fit.
Students seem to use more than functional similarity as memory probes for
arriving at appropriate analogies. Similarity of structure, external environment,
and performance characteristics too are often involved; for instance, the search
technique of ‘‘finding extreme adapters’’ is used to find high-performing biological
solutions by environmental similarity of problem–solution pairs. Capitalizing on
these findings, we introduced in 2011, a new representation framework called
Structured Representation for Biologically Inspired Design (SR.BID) that extends
and expands the SBF representation. The core aim of SR.BID was to create a
comprehensive representation linking biological solution descriptions and problem
descriptions over the broader range of concepts used for analogical indexing,
mapping, and evaluation. The SR.BID framework borrowed the structure (called
specifications in SR.BID) and function concepts from SBF, and appended envi-
ronment and performance concepts noted earlier. The ‘‘Four Box Method’’ uses
the four concepts pictured in Figure 7.3 to organize student thinking about prob-
lems and solutions.
Fig. 7.3 Four Box Method

of SR.BID
Table 7.5 Using the Four Box Method to describe the problem of building a better bike lock
Operating environment Functions Performance Specifications
Outdoors and indoors Protect bicycle from Weigh less than 5 lbs. Adjustable
theft
Wind and Prevent potential Withstand 4500 lbs. Weatherproof
precipitation damage to of force
Daytime and nighttime bicycle caused Withstand temperatures Flexible
by contact with below 32 F and above
lock 100 F
Temperature variations Waterproof, freeze proof, Strong materials
and shockproof
Easy to use
Deter from cutting
Competitively
priced
All assignments in the 2011 class were structured using the SR.BID framework.
The following Table 7.5 provides an example of a four box diagram provided to
describe the problem of building a better bicycle lock.
Furthermore, once given a problem specification and a specification of the
biological solution in the same format, students were asked to do a side-by-side
comparison where they identified whether elements were the same, similar, or
different. In Table 7.6, we see a side-by-side comparison of the North American
Elk antler with the bicycle lock problem. This side-by-side comparison forces
students to consider not only where the problem and solution align, but also where
solutions do not line up. Markman and Gentner (1993) have suggested that
comparison of source and target problems in analogical transfer often is based on
such alignment. We believe that in BID, highlighting these differences early in the
analogical mapping stage helps students consider their sources more deeply, as
well as identify potential transfer issues such as performance, size scaling, material
composition, and manufacturing earlier in the design process.
From 2006 through 2011, we have experimented with a number of represen-
tations and tools to help students overcome common interdisciplinary design
challenges. We have less experience with SR.BID than with other representations.
However, student surveys indicate that the SR.BID organizational framework
resulted in more pragmatic final projects and provided students with a more robust
method for evaluating analogies, especially for identifying potential transfer
failure points. On the other hand, students continued to provide shallow mecha-
nistic explanations of biological systems. In 2011, we did not teach diagrammatic
functional decomposition, and instructors felt as a result students lacked a deeper
understanding of the connectedness of functions in both problem and solution
descriptions.
182 J. Yen et al.
Table 7.6 Side-by-side comparison of the biological solution (Elk antler) with the problem
(bicycle lock) using SR.BID
Problem target Biological source (Elk antler)
Operational environment Operational environment
College students/adults N/A N/A
Global use in all habitats Different North America/Eastern Asia/forest habitat
Bike racks, poles, sign posts, fixed Different Elk head
structures
Usable in all seasons/usable at all times Different Usable only during mating season
of day
Temperature range: adaptable to outside Same Temperature range: adaptable to outside
temperature temperature
Weather resistant Same Weather resistant
Functions Functions
Prevent bike theft N/A Fight/protect against other male elks during
mating season
Withstand applied stress Same Withstand applied stress
Maintain temperature Same Maintain temperature
Deter possible thieves Similar Deter predators
Specifications Specifications
Stress withstanding materials Same Stress withstanding
Lightweight materials Different Strong materials
Inexpensive materials Similar Relatively low energy cost
Inert materials Same Inert materials
Detachable from bike Similar Ability to shed antlers after the end of
mating season
Lifespan of over 5 years Different Lifespan equals the duration of mating
season
Criteria Criteria
Weight \5 pounds Different Weight of up to 40 pounds
Fits around average sized tree trunk Different Height of up to 3.9 feet
Fits on/around average sized bike frame N/A N/A
7.7 Evaluation
One of us once overheard an alumnus from our 3rd iteration of the course describe
it to a prospective student. He said, ‘‘It’s different from any other course. There are
assignments like ‘go outside and find something.’ It’s hard to know exactly what
you’re supposed to do.’’ In most courses, it is clear to the students what specific
information and skills they must master. Because BID education is process ori-
ented, as opposed to content oriented, the students often have trouble gauging their
own performance, particularly before their first projects are vetted. Evaluation by
the faculty is necessary throughout, and most important quite early, to help stu-
dents realize what it is they should be working on, how good their work is, and
what mental activities are leading to productive outcomes. The first two course
7
Table 7.7 Six course elements (down) pertaining to evaluation that address the 9 key challenges (across)
Evaluation Searching Understanding Identifying Analogy Communicating Communicating Teaming in an Maintaining Evaluating
for biological and mapping across discipline complex system interdisciplinary equal designs
Three W’s X X X X X
In-class X X
feedback
Environmental X X
impact
assessment
Make-or-break X
QA
Materials X
assessment
Reports X X X X X X
183
184 J. Yen et al.
elements in Table 7.7 below are the main methods we have found to get students
on a productive track early in the course.
Ideas are commonplace; good ideas less so. Many of the ideas in the student
journals were novel, but impossible to implement so as to achieve the desired
functionality. Quantitative analysis is usually the key to assessing feasibility of a
design. For example, how much must the surface area of a shoe expand to prevent
sand from liquefying when an adult walks at normal speed? How much will
serrations at the leading edge of a lawn mower blade reduce noise? Course ele-
ments 3–5 in Table 7.7 represent the quantitative assessments we require. In
addition, we ask the students to perform some quantification of the three W’s.
The last course element in Table 7.7 is the final report. This pulls together the
biological sources, problem description, design description, analogical evaluation,
and all of the quantitative analyses that are described in this and the previous
sections. If the students can write a persuasive project summary and have correctly
performed the underlying analysis, they can feel confident that they have delivered
a good BID.
7.7.1 Three W’s
The three W’s, ‘‘What stinks,’’ ‘‘Who cares,’’ and ‘‘What are you going to do about
it,’’ were introduced in Sect. 7.6.3 as scaffolding for student problem definition.
For the oral presentations of the first two projects, we require the students to state
the three W’s of their problem definition. This helps them select a worthwhile and
well-defined problem. For the final reports, we also require quantitative justifi-
cations of each W. For example, the first W would ask how wasteful are lawn
sprinklers compared with drip irrigation? The second W would ask how much
clean water is wasted annually by lawn sprinklers and of what fraction of total
clean water use does that consist? A more thorough answer to the second W would
calculate the annual cost of the wasted water to a typical owner of a water
sprinkler. If the annual cost is a few dollars, who is going to care, even if the
overall cost is a hundred million? The third W would call for the quantitative
analysis of the design to be sure it saves the amount of water claimed, and an
estimate of the production cost.
7.7.2 In-class Feedback
During in-class work sessions, we circulate among groups, answering questions,

critiquing designs, helping with analyses, and suggesting ideas. We have not kept
records of these interactions, but we are sure that this feedback is indispensible to
the students during the first and second design projects. We often have the ready
knowledge to tell a group that an idea has already been tried, or that an organism’s
mechanism is not what they think it is, or that the basic nature of their problem is
different from what they suppose. This kind of feedback helps eliminate dead ends
early, before the team sinks much time into them.
The other kind of feedback that is very helpful during the early stages of work
has to do with problem focus. Students frequently begin with too broad a problem
and need to be advised to narrow their focus, often drastically. It has become much
easier for us to provide this feedback now that we have taught the course for
several years, because we have acquired some problem domain knowledge. For
example, every year since 2006 at least one team has wanted to solve the problem
of water. We have learned that worldwide water problems range from aquifer
depletion, desertification, inefficient irrigation, and leaky toilets to collection, non-
point-source pollution, filtration, and millions of children’s deaths annually. Each
of these differs by geographical region, culture, and other factors. We might
suggest a focus on collecting potable water from the air for a hundred thousand
refugees living in makeshift tents in Haiti, or on finding gray water alternatives to
pure aquifer sources for farmers in the midwestern USA.
Occasionally, a group will have too narrow a problem focus. If their solution is
good, it is usually enough to point out that there is not a sufficiently important
‘‘who cares,’’ and urge the students to find a broader scope of application.
Therefore, it is not usually vital to detect this flaw very early. Feedback during in-
class presentations, discussed next, is sure to reveal such flaws that have not yet
been detected.
We invite experts from various departments such as mechanical engineering,
materials engineering, architecture, chemistry, psychology, and civil engineering,
as well as local firms such as Perkins+Will, Interface, and David Oakey Designs
and that are interested in sustainability, to attend the oral and poster presentations
of the student projects. These presentations are typically given a week or two
before the final project reports are due. Each team gets feedback from the visiting
experts, the course instructors, and their fellow students. Surprisingly, we have
found that some of the toughest questions come from other students. The visitors
are the most likely to challenge fundamental assumptions or parameters of the
entire project. We instructors, perhaps because we have been providing feedback
all along, tend to ask the least unsettling questions. Instead, we usually probe to
test whether or not the students have a deep understanding of the biologically
inspired mechanism that is being transferred into the design.
Several times visitors have expressed regret that they had not been brought in
earlier, because there is not enough time for the student team to act on their
criticisms or ideas. On the other hand, these visitors are a scarce resource and we
are leery of imposing too much on them. The best use of this resource seems to be
during the presentations of the first and second designs, because the teams will
choose one of those two to refine for their third design and therefore have several
weeks to take an expert’s comments into account.
186 J. Yen et al.
7.7.3 Environmental Impact Assessment
Student responses to our course suggest that BID captures the imagination and
attunes students to values of sustainability. In fact, many engineering students in
our early courses reported they were more likely to consider sustainability and
environmental impact of their designs as a consequence of learning BID, even
though sustainability was not a design requirement. Subsequently, we added an
environmental impact assessment (EIA) assignment to align student output with
their greater sensitivity to environmental concerns.
The EIA assignment creates a number of challenges, given many engineering
and biology curricula do not cover this kind of evaluation. We discovered it was
necessary to familiarize students with the major environmental impact categories
and their associated metrics (e.g., greenhouse gases in CO2 kilogram equivalents
and solid non-toxic waste in cubic feet). We identified some of the most common
difficulties and created a quantitative homework assignment that forced the stu-
dents to navigate them. The assignment was to compare the environmental impact
of travel by air and travel by car. This forced students to understand the need to
express the cost per function achieved (e.g., amount of CO2 equivalents released
per passenger miles travelled), and how to prioritize potential costs (e.g., the
amount of clean water used per passenger mile is negligible compared to the
impact of greenhouse gas emission). Afterward, when teams were working on their
projects, we met with each group to discuss which impacts were important and
how they were to be measured.
Several of the changes that we have described, namely identifying pitfalls and
environmental impact categories in lectures, tailoring quantitative analysis
assignments to these lectures, and discussing these issues with each team while
they were developing their designs, had the net effect of changing quantitative
assessment from a burdensome requirement of a final report to a key tool used
during much of the design process.
7.7.4 Make-or-Break Quantitative Analysis
In the first three years of the course, we gave three quantitative homework
assignments, each analysis tied to a specific reading or lecture. Our aim was to
stimulate students to perform quantitative evaluations of their projects. These
assignments were unpopular; many students, especially biologists, found them
difficult. Starting in the fourth year, we changed these from individual to group
assignments. To keep the biologists engaged, we offered extra credit to teams if a
biologist presented the group’s solution to the class. We observed that the quality of
the student solutions improved, and that the student satisfaction with the assign-
ments increased when the design team was jointly responsible for the exercise.
However, the degree to which all team members, in particular the biologists,
learned how to do quantitative analysis is unknown.
Though we do not know whether everyone learned how to perform quantitative
analysis, we do know that the students did not learn how to choose what quanti-
tative analyses were worth doing. All final design reports were supposed to include
a quantitative assessment that related to how well the design functioned. In the first
few years, we were usually dissatisfied with their quality. Many assessments
lacked depth or importance. Teams frequently analyzed aspects of the design that
were not critical to its performance, choosing analyses with straightforward
techniques as opposed to relevance. We elected to address this problem with a
‘‘make-or-break’’ lecture, in which we stress that usually there is a single quan-
titative issue of function that is critical to the success of the design. A bicycle
helmet must be able to protect against a certain speed of collision; a condensation
device for desert use must produce a certain amount of water per day; a levee must
withstand a certain flood height. We told each team to figure out what would make
or break their design. We then met with each team to discuss their choice. This
discussion was important because otherwise a difficult time-consuming technical
analysis could turn out to be irrelevant or a major design infeasibility could go
undetected.
Our subsequent experience has led us to identify common issues that students
confront in this analysis. One pitfall has to do with scaling. For example, a human-
sized gecko could not climb walls easily because the mass increases as the cube of
the length, but the surface area of the foot increases only as the square of the
length. The adhesive force is proportional to the surface area, as a simple thought
experiment will show. We created a new quantitative homework assignment for
which scaling was the key. Since biological solutions often are scale-dependent,
students often will have to deal with this issue, and some discussion of scaling
seems key for successful analogical transfer of principles. The other common
pitfall had to do with materials. This was so important that we made a materials
assessment a separate requirement, as described in the next subsection.
7.7.5 Materials Assessment
Students in the first 3 years of the course would often base their design on a
hypothesized material with certain physical properties, when no such material
existed. When we reviewed the course after 3 years, we were a bit shocked to see
that this single weakness rendered about one-third of all the designs infeasible! We
began to warn students not to rely on imagined materials, encouraging them to use
existing material or to design a hybrid material from known materials. Now, we
require a materials analysis in the second or third week of the third project. The
final reports typically incorporate this materials analysis.
Often the properties of a material have turned out to be the ‘‘make-or-break’’
quantitative question. In several cases, teams performed a computation-intensive
188 J. Yen et al.
finite-element analysis to answer the question. Usually, only one member of the
team, a mechanical or materials engineer, knew how to perform such an analysis.
In our experience, therefore, a materials analysis seems necessary to prevent
situations in which students produce unfeasible designs. In the most recent itera-
tion of the course, fall 2012, only one of the eight final designs (a radically
different toothbrush) depended on material of dubious manufacturability.
7.7.6 Reports
We have always required students to deliver both oral and written reports. In the
first year, we tried different formats. For written reports, we asked for traditional
write-ups of about 10 pages, posters, and pamphlets of 4–8 sides. For oral reports,
we asked for either short poster presentations or PowerPoint presentations. We
observed that students were highly motivated by poster presentations, and we have
retained them. We found that written reports got much better if we specified a
template in advance and tied all of the elements in the template to previous
assignments. In this way, final reports served as a reflective synthesis of previous
work and provided an opportunity for improvement. Report templates also pro-
vided students with focus. There are so many aspects of the process of BID that
could be included in a report that students are at a loss for what to include or not
include. In particular, in 2007, about 40 % of the final report documented the
design process, while 60 % documented the actual final design. The template
seemed to reinforce both the process learning goals for the students and the
product/design goals. Creating good rubrics for a class is an extremely difficult
problem, particularly in design. To grade the final designs, these 10 sections are
awarded a specified portion (%) of the final grade as follows:
1. Summary (5 %). Specify the problem and the biological source; state the key
analogy; describe the design solution and its value proposition as compared
with existing solutions.
2. Biological System Understanding (10 %). For solution-based designs, convey
a deep understanding of the primary natural system, with particular focus on
explanation of the mechanism(s) of interest. For problem-based designs,
provide a deep description of all mechanisms transferred to the design. In
addition, describe at least briefly all natural systems that were considered,
indicating why a system was or was not chosen for inspiration.
3. Design Problem Understanding (10 %). Motivate the problem, including what
stinks, who cares, and what are we going to do about it. Also, give a detailed
problem decomposition showing a logical analysis of the functions involved in
the problem including function, operating environment, performance criteria,
and constraints.
4. Biological System to Design Problem Analogy and Comparison (10 %).
Describe similarities and differences between the biological systems and the
design problem. In addition, present arguments for and against the suitability
of the biological systems to serve as a solution to the design problem.
5. Visualization (10 %). Supplement the written text with a variety of visual
representations such as graphs, figures, drawings (CAD or freehand), and
tables. Legends must be informative.
6. Quantitative Analysis of Biological Mechanism (20 %). Provide a succinct
and quantitative analysis of the mechanics, material properties, or interacting
processes of the biological system(s) that are transferred to the design.
7. Quantitative Analysis of Design (20 %). Provide a succinct and quantitative
analysis of the key functions of the problem. Show how the new design
integrates the principles derived from nature.
8. Design Understanding (10 %). Discuss the principal obstacles to achieving
the design objectives that were encountered. Assess the value of the design
(greater functionality, cost savings, increased sustainability, other potential
applications).
9. Cross-Domain Translation Creativity (±10 %). This portion of the grade
depends upon the creativity of the design based on its novelty with respect to
current technology and previous BID designs, together with the potential
usefulness of the proposed product.
10. Literature (5 %). This must contain key references from the primary literature
(no Weblinks allowed) for the biological systems, existing solutions, similar
problems, materials, and mechanics.
Item 9 in the list above requires clarification. The weights of the other items sum
to 100 %. Item 9 permitted the project grade to go up or down by as much as a full
level, for example, from B to A or C. We instructors did not fully agree as to how
much the designs should be graded on the process rather than on the outcome. Item 9,
being a highly subjective criterion, gave individual instructors leeway with respect
to the rest of the grading rubric. To ensure fairness, we balanced the set of instructors
assigned to each report. Note also that the weights assigned to these categories will
vary, reflecting the instructors’ course goals and institutional context.
In the most recent iteration of the course, we required a complete draft of the
report a few weeks before the final version was due. We graded the drafts as
carefully as we would have graded final versions. About half of the final reports
were much improved over the drafts. (several were already excellent). This process
required a lot of time from both faculty and students, but it significantly improved
the outcome.
7.8 Interdisciplinary Training
Having the opportunity to work in interdisciplinary teams gives students that

chance to examine a problem from a different viewpoint, share uncommon
knowledge between disciplines, enable them to re-examine their own major, and in
190 J. Yen et al.
essence, seed their minds with new ideas. Teams in this class include at least one
of each of these two disciplines: biologist, mechanical engineer, plus a mixture of
these: systems engineer, materials scientist, designer (industrial designer, architect,
and artist). Asking the students to show they are able to use each other’s skills,
starting from the biological inspiration, throughout the design process, to the final
quantitative analysis of feasibility informs them of the importance of the inter-
disciplinary effort. These interactions expand their design space, promoting cre-
ative thinking and innovation in design (Table 7.8).
7.8.1 Faculty Engagement
How often does a biologist work on a team with a biomedical or mechanical

engineer, a materials scientist, an industrial engineer or an architect or city
planner? One key to effective bioinspired design is that it requires expertise in
multiple fields. In our experience, there is no greater influence on the success of a
final design than having a mentor to help facilitate the team design. Even one or
two sessions with an expert can make a dramatic difference. For example, when a
team of mechanical engineers, computer scientists, and biologists tried tackling the
issue of desalination, they classified the problem as one of finding a way to
generate water pressure for reverse osmosis. Having created a biology-based
solution that required ‘‘no input energy,’’ the team thought they had ‘‘solved’’ the
problem. Five minutes with a faculty expert, and suddenly, now recast in terms of
a thermodynamics problem, the team saw they had a major problem with their
design (specifically that the system would quickly reach equilibrium after which
no further desalination would occur). Over the years, we have identified those
faculty who are open to interdisciplinary collaboration, can spare the time to
facilitate a team over several one or 2 h sessions, and evaluate the output in such a
way that makes sure the team correctly understands the principles of interest.
Taking advantage of local expertise helps customize the course to the strengths of
the institute.
7.8.2 Knowledge from Other Domains
As already mentioned, BID draws from many areas of scientific knowledge and
cannot be accomplished without at least two or more disciplines working together.
In our course, we emphasize the essential value of knowledge from other domains.
In particular, it requires a sufficiently broad understanding of biology to facilitate
search and selectively deep understanding once a particular biological source is
targeted as a potential source for innovation. Whereas substitutes exist, there is
still no resource quite as effective as a good biologist.
7
Table 7.8 Five course elements (down) pertaining to interdisciplinary training that address the 9 key challenges (across)
Interdisciplinary Searching Understanding Identifying Analogy Communicating Communicating Teaming in an Maintaining Evaluating
training for biological and mapping Across complex system interdisciplinary equal designs
biological systems understanding and discipline knowledge environment engagement
systems good design transfer boundaries throughout
Faculty X
engagement
Knowledge from X X X X X X X X X
other
domains
Nature auction X X X
ID teams X X
Peer evaluations X X X
191
192 J. Yen et al.
It is not easy for an engineer to identify keywords to search for a system with
properties they seek: a BID thesaurus is useful, and biologists can act as a
‘‘translator.’’ Although several groups are working on techniques, such as context-
based searching, to help engineers bridge the knowledge gap without direct access
to biological expertise, such an approach is neither optimal nor justified when there
is easy access to biologists. Hence, we always place at least one biologist in a team
of 5, although adding additional biologist team members, when possible, is sound
practice. Just as for engineers, there are many kinds of biologists, so the particular
ability of the biology teammate can affect strongly the choice of systems that can
be examined by the team.
7.8.3 Nature Auction
It is important to emphasize the vital role of non-engineering disciplines in what is

(ultimately) an engineering design exercise. One of our techniques is to throw the
students into a fun but unfamiliar situation that establishes the importance of
different types of knowledge. In our first class, we form temporary student teams,
each containing at least one biologist. Then, we engage them in an extraordinary
auction. We are auctioning off nature. The room is lined with often spectacular
images of organisms (e.g., basilisk lizard) performing some uncanny feat (walk on
water) with a caption that describes the behavior. The teams are given an equal
number of ‘‘BID dollars’’ to select at least 3 (usually 4–5) biological systems to
study for their first (solution-based) BID. They examine potential selections as a
team and discuss the value of each natural system as the basis of their choice.
Thus, begins the process of revaluing the role of nature, and the role of their fellow
teammates! The BID auction not only provides them with a jumpstart on their
investigation of interesting biological organisms, but also helps them learn about
the knowledge, values, and perspectives of their teammates.
7.8.4 Interdisciplinary Teams
We continue to use the interdisciplinary team as a way to encourage the impor-

tance of acquiring and communicating knowledge outside of one’s domain. In the
final presentations, extra credit is given to the engineer who can explain the
biological function and the biologist who can explain the engineering function.
This embeds in each team the need for all participants to share their knowledge
and is one of the pedagogical advantages of collaborative inquiry-based learning
(Bransford et al. 2000; Bybee 1997). Eventually, everyone in the group under-
stands the value of the biologist but usually, the biologist remains the source and
search engine for interesting biological strategies. Similarly, not all the biologists
succeed in becoming a materials engineer or mechanical engineer, but often
understand the basic constraints and capabilities of these skills, and learn how to
express themselves using concepts familiar to the engineers. All students come to
the conclusion that they can address this complex problem more effectively by
putting their skills together and learning how to apply their knowledge as a team to
address the challenge.
7.8.5 Peer Evaluations
Team interactions can range from everyone working equally under strong lead-
ership and team spirit to dysfunctional teams ruined by team members who do not
or are unable to engage in the process. We ask each member of a team to evaluate
themselves and their team, using a system based on a fictional reward. Students are
given 1000 fictional dollars per team member, which they distribute between
individuals (including themselves) based on the value to team. We ask each stu-
dent to justify this distribution by commenting on the contribution of each team
member (again including themselves). We alter the grades of students who average
significantly higher or lower than 1,000. Our intentions are both to be fair and to
motivate. Students know in advance that their grades may be lower than their
team’s grade if they do not contribute adequately. In some teams, everyone clearly
valued the expertise offered by each discipline. However, in other teams, the
engineers would not engage in the biological search and the biologists did not
know how to engage in the quantitative assessments. More attention is needed to
find ways to engage all the disciplines throughout the process.
7.9 Synthesis
This is an unusual course. We are not teaching biology or engineering, but we are
asking the students to obtain a deep understanding of the specific biological system
they intend to apply and translate into engineering design. A student’s under-
standing of the biological system has to be deep enough that he or she can identify
the biological knowledge that should be transferred to an engineering problem.
Well-defined grading rubrics are useful so students know what constitute the traits
of a BID expert. These should be directed at project evaluation, but also need to
help students understand what sorts of mental process and activities help produce
novel and useful (e.g., creative) designs.
Despite the unfamiliarity and challenges, students are eager to include this
design process in their skill set because of the lure of invention, the novelty of
BID, and its potential to lead to more sustainable practices. When given the
freedom to work on a problem of their own choosing, motivation is not a sig-
nificant problem. While engagement can wax and wane, depending especially on
the ability of an individual to apply their domain skill set and feel useful, case
194 J. Yen et al.
studies, BIDwow, and auctioning off nature work well to maintain enthusiasm.
Taking advantage of this enthusiasm by teaching bioinspired design allows us to
reach quite a few learning goals, making it well worth the effort to identify some
best practices. These recommendations are specifically for a course where we take
inspiration from biology for design, going beyond copying or using nature.
Our experience has been that for effective BID, having biological expertise is
necessary, either by having biologists on the team or available for consultation.
When there is no option for student teams to include biologists, trying to find the
right function from the right natural system is difficult. We recommend that
experts be consulted for the best understood biological systems. Seasoned biology
faculty that do research on biological systems and attend biology conferences have
an edge in finding the natural systems that are rich in mechanistic details. Whereas
using this expertise to give the students a good starting point may take away the
chance to teach them how to search the biological literature, the students still will
have many chances to hone their search strategies to find other exemplars of
similar or inverted functions in extreme environments. Advanced students may
even learn to find relevant examples based on phylogenetic relatedness, conver-
gent evolution, or exaptations.
Focusing the found object exercise on some key biological functions (sensing,
locomotion, and hierarchy) and comparing conjectured designs based on these
solutions illustrates to the class the myriad of possibilities, teaches them about
these key biological concepts and reduces design fixation. Skillful use of SBF,
functional decompositions, and analogical reasoning to compare biological and
engineering systems enable connections to be made between the biological
functions and engineering needs. The tools for these cross-disciplinary interactions
that we have developed work well. Students read the primary literature carefully
and they can use SBF to focus and to keep them from getting lost in the inherent
biological complexity. The newly developed SR.BID framework applied to the
problem and biological solutions works well to identify the many functions to take
across the divide and to evaluate analogies more deeply across functional, per-
formance, and specification viewpoints.
When the student population is diverse, encouraging team-based problem
solving is not only desirable, but necessary. Using bonus points given to the
biologist who is able to explain the engineering principles or the engineer who is
able to explain the biological principles tells the students that we think this ability
to communicate across disciplines is useful in their training to be a practicing
bioinspired designer. This cross-disciplinary practice serves as one way to keep
disciplines engaged throughout the design process. Oral presentations where each
team member speaks can help encourage all members to be active.
Design evaluation remains one of the more challenging aspects. Students need
clear direction as to what does and what does not constitute a good problem, and to
avoid common pitfalls in arriving at good designs (e.g., poor material selection, lack
of appreciation of scaling, and appropriate EIA). Simple assignments pertinent to
developing specific skills help students to incorporate these considerations in their
final design, particularly when project grading rubrics indicate they are required.
Facilitator interaction is needed to make sure the biological mechanisms are

understood correctly and the match to engineering functions makes sense, or teams
may work diligently but unproductively. This is a particular problem since students
consider the amount of time and energy devoted to a project as a sunk cost, which
discourages them from unbiased evaluations of the potential for a given project.
Timely and useful feedback is not a problem if the facilitator has a vested interest in
the project. But having all these capabilities in a single instructor faced with any
number of biological and engineering functions is rare, particularly when students
are allowed to self-identify projects.
The final output of our course is a conceptual design that identifies the make-or-
break criteria and theoretically testing the design’s feasibility. Convincing tests of
product success requires building and testing a prototype, which requires another
semester of effort.
7.10 Conclusions and Next Steps
Problem areas that require additional attention are the search strategy for bio-
logical systems, a more complete method for teaching analogical mapping and
evaluating good analogies, and evaluating good designs and good design prob-
lems. Searching and identifying useful biological systems with high potential for
transfer to design still can be much improved when an expert biologist makes a
suggestion or guides an exploration into the natural world. Capitalizing on evo-
lutionary knowledge is key here, and computational methods, while promising,
still cannot take the place of a skilled biologist. Although many connections can be
made between biology and engineering, it is still difficult to figure out which
analogical match is the best to pursue to solve the engineering challenge.
The BID class is also a research laboratory for studying BID artifacts and
practices. On one hand, it has allowed us to apply, evaluate, and explore theories
of creative design, analogical reasoning, and knowledge representation. On the
other, in situ studies of BID already have led to the development of new
descriptive theories of BID such as solution-based analogy, new knowledge rep-
resentations such as SR.BID, new interactive tools such as DANE and Biologue,
and new techniques such as the four box method for specifying design problems.
Our current course, focused on idea generation and conceptual design, does not
include instruction in essential skills that can bring a creative idea to fruition.
Translating a biological principle to a functioning device requires fundamental
concept testing and experiments to build and test a BID prototype. Over and over,
these new interdisciplinary designers realize that quantitative analyses can eval-
uate value and feasibility of the design but that implementation and testing provide
the essential proof of success. Long (2012) found that making a physical prototype
solves the problems of functionality, manufacturability, and improper quantitative
analyses. A sequel to this course would be a fundamental concepts testing class to
evaluate whether the make-or-break criterion is feasible. If it is feasible, then the
196 J. Yen et al.
3rd class would be prototype building and testing. Some students in our classes
have gone on to do this independently in other design classes. However, for certain
projects, there may be some value in extremely limited prototyping within the
context of our current 15 week course model. Rapid prototyping combined with
hierarchical scaling have been implemented voluntarily by student groups in our
BID class and this activity has gone far to evaluate designs. For certain classes of
projects, incorporating this requirement would be feasible and useful.
On balance, BID provides a continuous and exciting growth process. Practicing
this approach improves the facility with these new design skills, encouraging us to
make friends outside our expertise. The novelty and allure of the BIDs provide
motivation to go beyond the superficial and deeply understand the complexity of
the problem and the complexity of the natural system.
Acknowledgments We are grateful to several colleagues who have contributed to this work
over the last several years, including Inbal Flash-Gvili, Wendy Newstetter, Swaroop Vattam, and
Bryan Wiltgen. We thank the US National Science Foundation for its support of this research
through a TUES grant (#1022778) entitled ‘‘Biologically Inspired Design: A Novel Interdisci-
plinary Biology-Engineering Curriculum,’’ and a CreativeIT Grant (#0855916) entitled ‘‘Com-
putational Tools for Enhancing Creativity in Biologically Inspired Engineering Design.’’
References
Altendorfer A, Moore N, Komsuoglu H, Buehler M, Brown HB Jr, McMordie D, Saranli U, Full

R, Koditschek DE (2001) Rhex: a biologically inspired hexapod runner. Auton Robots
11:207–213
Ausubel D (2000) The acquisition and retention of knowledge: a cognitive view. Kluwer, New
York
Bhatta S, Goel A (1997) Learning generic mechanisms for innovative design adaptation. J Learn
Sci 6(4):367–396
Biomimicry 3.8 Institute (2008) AskNature—the biomimicry webportal. http://www.
asknature.org/. Retrieved 15 Apr 2012
Biomimicry 3.8 Institute (2009). Biomimicry: a tool for innovation. http://www.
biomimicryinstitute.org/about-us/biomimicry-a-tool-for-innovation.html. Retrieved on 15
Apr 2012
Bransford J, Brown A, Cocking R (eds) (2000) How people learn: brain, mind, experience, and
school. National Academy, Washington, DC
Bruck H, Gershon A, Golden I, Gupta S, Gyger L, Magrab E, Spranklin B (2007) Training
mechanical engineering students to utilize biological inspiration during product development.
Bioinsp and Biomim 2:S198–S209
Bybee R (1997) Achieving scientific literacy. Heinemann, Portsmouth
aiding biomimetic and artificial inspiration of new ideas. AIEDAM 19:113–132
Chan Y, Yu A, Chan C (2010) Assessing students’ integrative learning in biomedical engineering
from the perspectives of structure, behavior, and function. In: Proceedings of 40th ASEE/
IEEE frontiers in education conference, 27–30 Oct 2010, Washington, DC
Chandrasekaran B (1994) Functional representation: a brief historical perspective. Appl Artif
Intell 8:173–197
Chandrasekaran B, Goel A, Iwasaki Y (1993) Functional representation as a basis for design

rationale. IEEE Comput 26:48–56
Cheong H, Chiu I, Shu, Stone R, McAdams D (2011) Biologically meaningful keywords for
functional terms of the functional basis. J Mech Des 133: doi:10.1115/1.4003249
Chiu I, Shu L (2007a) Biomimetic design through natural language analysis to facilitate cross-
domain information retrieval. AIEDAM 21:45–59
Chiu I, Shu L (2007b) Using language as related stimuli for concept generation. AIEDAM
21:103–121
Clement J (2008) Creative model construction in scientists and students: the role of imagery,
analogy, and mental simulation. Springer, Dordrecht
Dunbar K (2001) The analogical paradox. In: Gentner D, Holyoak KJ, Kokinov BN (eds) The
analogical mind: perspectives from cognitive science. MIT Press, Boston
Dym C, Brown D (2012) Engineering design: representation and reasoning. Cambridge
University Press, New York
Ebert-May D, Momsen J, Long T, Wyse S (2010) Dearly departed Grandma Johnson: revealing
student understanding of carbon cycling with structure-function-behavior models. In:
Proceedings of 95th annual meeting of the Ecological Society of America
Erden M, Komoto H, van Beek T, D’Amelio V, Echavarria E, Tomiyama T (2008) A review of
function modeling: approaches and applications. AIEDAM 22:147–169
French M (1996) Conceptual design for engineers, 3rd edn. Springer, London
Gentner D (1983) Structure-mapping: a theoretical framework for analogy. Cog Sci 7:155–170
Gero JS, Kannengiesser U (2004) The situated function–behaviour–structure framework. Des
Stud 25:373–391
Gick M, Holyoak KJ (1983) Schema induction and analogical transfer. Cog Psych 15:1–38
Goel A (1997) Design, analogy and creativity. IEEE Intell Syst 12(3):62–70
Goel A, Gomez A, Grue N, Murdock W, Recker M, Govindaraj T (1996) Towards design
learning environments—exploring how devices work. In: Proceedings of international
conference on intelligent Tutoring systems, Montreal, Canada
Goel A, Rugaber S, Vattam S (2009) Structure, behavior and function of complex systems: the
SBF modeling language. AIEDAM 23:23–35
Goel A, Vattam S, Wiltgen S, Helms M (2012) Cognitive, collaborative, conceptual and
creative—four characteristics of the next generation of knowledge-based CAD systems: a
30:606–622
Helms M, Vattam S, Goel A, Yen J (2011) Enhanced human learning using structure-behavior-
function models. Interactive learning environments. In: Proceedings of 11th IEEE interna-
tional conference on advanced learning technologies, Athens, Georgia, USA, July 2011,
pp 239–243
Hofstadter D (1996) Fluid concepts and creative analogies: computer models of the fundamental
mechanisms of thought. Basic Books, NY
Holyoak KJ, Thagard P (1995) Mental leaps: analogy in creative thought. MIT Press, Cambridge
Keane M (1988) Analogical problem solving. Ellis Horwood, Chichester
Kolodner J (1993) Case-based reasoning. Morgan Kaufmann, San Francisco
Lave J, Wenger E (1991) Situated learning: legitimate peripheral participation. Cambridge
Lee H, Lee BP, Messersmith PB (2007) A reversible wet/dry adhesive inspired by mussels and
geckos. Nature 448:338–341
Long J (2012) Darwin’s devices. What evolving robots can teach us about the history of life and
the future of technology. Basic Books, New York
Markman AB, Gentner D (1993) Structural alignment during similarity comparisons. Cog Psych
25:431–467
Miklosovic DS, Murray MM, Howle LE, Fish FE (2004) Leading-edge tubercles delay stall on
humpback whale (Megaptera novaeangliae) flippers. Phys Fluids 16:L39–L42
198 J. Yen et al.
Nagel R, Midha P, Tinsley A, Stone R, McAdams D, Shu L (2008) Exploring the use of
functional models in biomimetic concept design. J Mech Des 130:1–13
Nagel J, Nagel R, Stone R, McAdams D (2010) Function-based, biologically inspired concept
generation. AIEDAM 24:521–535
Nakrani S, Tovey C (2007) From honeybees to internet servers: management of internet hosting
centers. Bioinsp and Biomim 2:S182–S197
National Research Council (2009) A new biology for the 21st century. http://www.nap.edu/
catalog.php?record_id=12764. National Academies Press, Washington
National Research Council (2011) A Framework for twenty-first century science education:
practices, cross-cutting concepts and core ideas. http://www.nap.edu/
catalog.php?record_id=13165. National Academies Press, Washington
Nersessian NJ (2008) Creating scientific concepts. MIT Press, Cambridge
Pahl G, Beitz W, Feldhusen J, Grote K (2007) Engineering design: a systematic approach.
English edition: Wallace K, Blessing L (trans: 3rd edn.) Springer
Potyrailo RA, Ghiradella H, Vertiatchikh A, Dovidenko K, Cournoyer JR, Olson E (2007)
Morpho butterfly wing scales demonstrate highly selective vapour response. Nat Photonics
1:123–128
Rasmussen J (1985) The role of hierarchical knowledge representation in decision making and
system management. IEEE Trans Syst Man Cybern 15:234–243
Sarkar P, Chakrabarti A (2008) The effect of representation of triggers on design outcomes.
AIEDAM 22:101–116
from nature. J Comp Info Sci Eng 8(3):1–9
design. AIEDAM 24:483–505
Shu L (2010) A Natural-language approach to biomimetic design. AIEDAM 24:507–519
Silk E, Schunn C (2008) Core concepts in engineering as a basis for understanding and improving
K-12 engineering education in the United States. Paper presented to the National academy of
engineering/National research council workshop on K-12 engineering education, Washington,
DC
Simon H (1996) Sciences of the Artificial. 3rd Ed, MIT Press
Simonton D (2004) Creativity in science: chance, logic, genius, and zeitgeist. Cambridge
Speth E, Dirnbeck M, Murrilo L, Momsen J, Long T (2011) Introductory (but not simple)
biology: student’s representation of complex causal relationships in biological systems. Paper
presented to the first meeting of the society for the advancement of biology education
research, Minneapolis
Tero A, Takagi S, Saigusa T, Ito K, Bebber DP, Fricker MD, Yumiki K, Kobayashi R, Nakagaki
T (2010) Rules for biologically inspired adaptive network design. Science 327:439
Vattam S, Goel A (2011) Foraging for inspiration: understanding and supporting the information
seeking practices of biologically inspired designers. In: Proceedings ASME DETC conference
on design theory and methods, Washington, DC
Vattam S, Helms M, Goel A (2007) Biologically inspired innovation in engineering design: a
cognitive study. Technical Report GIT-GVU-07-07, graphics, visualization and usability
center, Georgia Institute of Technology
inspired design. AIEDAM 24:467–481
Vattam S, Wiltgen B, Helms M, Goel A, Yen J (2010b) DANE: fostering creativity in and
through biologically inspired design. In: Proceedings of first international conference on
design creativity, Kobe, Japan
Vattam S, Goel A, Rugaber S, Hmelo-Silver C, Jordan R, Gray S, Sinha S (2011) Understanding
complex natural systems by articulating structure-behavior-function models. J Educ Tech Soc
14:66–81
Vogel S (2000) Cat’s Paws and Catapults: mechanical worlds of nature and people. W.W. Norton
and Co, New York
Vygotsky L (1978) Mind in society: the development of the higher psychological processes. The
Harvard University Press, Cambridge
Weissburg M, Tovey C, Yen J (2010) Enhancing innovation through biologically inspired design.
Adv Nat Sci 3:145–167
Wilson EO (1984) Biophilia. The Harvard University Press, Cambridge
Des Stud 31:169–186
engineering design. Adv Nat Sci 3:168–187
interdisciplinary education. In: Bar-Cohen Y (ed) Biomimetics: nature-based innovation.
Taylor & Francis, Boca Raton
Chapter 8
Supporting Analogical Transfer
in Biologically Inspired Design
Amaresh Chakrabarti
Abstract Biologically inspired design (BID) is an emergent area of research for

understanding design with biological mechanisms as inspiration, and for sup-
porting systematic BID for developing creative designs. Understanding and sup-
porting the processes of analogical transfer, whereby potential biological material
is identified and adapted to solve engineering problems, is the focus of this chapter.
Two questions are asked: At what level does analogical transfer take place? How
to support analogical transfer? Our empirical studies show that transfer generally
takes place at four levels of abstraction: state change, organ, attribute, and part.
When unaided, BID is dominated by transfer at part, attribute, and organ levels,
which reduces potential for creativity. This led to development of new guidelines
for supporting systematic analogical transfer, an Integrated Framework for
designing to encourage transfer at each level of abstraction, and a computational
tool called ‘Idea-Inspire’ to provide analogically relevant biological stimuli for
inspiration at any of these levels. Comparative studies using these interventions
show significant increase in the number of transferred designs when aided by these
interventions and a shift in the majority of the transfer to state change and organ
levels, thereby increasing the potential for greater creativity.

Keywords Analogical transfer Biomimetics Biological stimuli Engineering

design Guidelines Idea-inspire GEMS of SAPPhIRE Novelty Usefulness

Creativity Technical product development
A. Chakrabarti (&)
Innovation, Design Study and Sustainability Laboratory (IdeasLab), Centre for Product
Design and Manufacturing, Indian Institute of Science, Bangalore 560012, India
e-mail: ac123@cpdm.iisc.ernet.in

202 A. Chakrabarti
8.1 Introduction
Biologically inspired design (BID) is an emergent area of research that focuses on

understanding, and supporting designing systematically, with biological mecha-
nisms as inspiration. BID is interesting because nature has ‘invented’ a plethora of
mechanisms that use a variety of phenomena to achieve, in a resource-effective
manner, numerous functions that have their equivalent in engineering problems.
Gordon (1961) identified biology as a major source of analogies for engineering
design.
BID has the potential to enhance creativity due to various reasons; using bio-
logical systems may inspire generation of:
• a large number of designs; previous research shows that generating a large
number of designs enhances creativity (Chakrabarti et al. 1992; Chakrabarti and
Tang 1996; Fricke 1996; Sarkar and Chakrabarti 2007a);
• designs that use a wide variety of actions and phenomena; previous research
shows that designs with greater variety at higher levels of abstraction such as
actions and phenomena lead to designs of greater novelty (Srinivasan and
Chakrabarti 2010a). Also, cross-domain inspiration is identified as a valuable
source of analogical transfer (Benami and Jin 2002; Hon and Zeiner 2004;
Tseng et al. 2008).
• designs that utilise resources efficiently; it is noted that resource effectiveness is
an important criterion for designs of value (Ulrich and Seering 1990; Chak-
rabarti 2000, 2001, 2004a; Chakrabarti and Singh 2007).
Engineers have often been inspired by biological mechanisms. However, these
inspirations traditionally resulted from serendipity and individual interest, and
rarely from systematic study (Chakrabarti et al. 2005). BID research promises to
expand the understanding of how BID is currently carried out, how this can be
enhanced with systematic processes as framework and biological mechanisms as
stimuli, for developing novel and valuable—that is, creative designs (Sarkar and
Chakrabarti 2007b, 2011).
8.2 BID Research: A Brief Overview
The DRM framework proposed by Blessing and Chakrabarti (2009) describes

research into design in terms of the following four stages:
• Research Clarification. This is the stage at which the overall goals of the
research, criteria for assessing research success (e.g. which goals and aspects of
designing could be understood or improved with the results of this research),
and the main research questions and hypotheses are developed.
• Descriptive Study I. During this stage, an understanding of (the current state of)
designing, and the factors that influence designing and its success, is developed.
8 Supporting Analogical Transfer in Biologically Inspired Design 203
This understanding of designing ‘as is’ provides the basis for identifying
possible aspects of designing that could be improved.
• Prescriptive Study. During this stage, specific aspects of (current) designing are
focused on for improvement, and some form of support—which can be guide-
lines, methods, tools, or standards—is developed in order to improve these
aspects of designing. The support is envisaged to influence and change current
designing into a more desired ‘future’ designing, so as to influence success
criteria in a more positive manner.
• Descriptive Study II. During this stage, the support developed is evaluated for its
effectiveness in changing designing as desired and assessing its influence on
success. Depending on the results, iterative steps are undertaken to improve the
support (going back to Prescriptive Study), or even improve the understanding
of current designing (going back to Descriptive Study I).
According to the DRM framework, a specific programme or piece of research
into design can be described as a combination of a subset of these four stages, with
varying emphasis on the stages. DRM categorises design research broadly into two
categories: Descriptive Studies—where the main focus is on developing research
criteria and understanding of designing ‘as is’, and Prescriptive Studies—where
the main focus is on developing and evaluating support so as to change designing
‘as should be’. Using the DRM framework, BID can be categorised into
descriptive (i.e. as is) studies and prescriptive (i.e. as should be) studies
(Chakrabarti and Shu 2010).
Descriptive studies into BID are relatively few and recent (Chakrabarti and Shu
2010). Helms et al. (2009) analyse the processes followed in BID projects, where
student designers carry out given design tasks in order to develop BID. A major
finding is that in BID, both problem and solution decompositions are transferred;
this is consistent with the earlier findings that problems and solutions co-evolve in
designing—and therefore both are design outcomes (Nidamarthi et al. 1997;
Chakrabarti et al. 2004b). Vattam et al. (2008) report that generation of compound
solutions are achieved in these processes through two related approaches: analogy
and problem decomposition. Vattam et al. (2010a) argue that while BID is
inherently analogical in nature, understanding of its analogical basis is currently
limited. They present an observational study of a series of BID sessions towards
developing a content theory of creative analogies in the context of BID. Sartori
et al. (2010) analyse a collection of published biomimetic design cases, and
identify the generic levels of abstraction at which biomimetic transfer occurs in
design (further discussed in Sect. 8.3).
There are many prescriptive studies in this area. Some focus on biomimetic
processes, some on databases and guidelines, and others on support tools.
Hill (1997, 2005) proposes an orientation model for supporting biomimetics
projects; it has two steps: goal setting and solution identification. Based on con-
tradicting demands identified in goal setting, solution identification consists of: (1)
determining the basic function(s) underlying the contradicting demands; (2)
identifying relevant biological structures with similar functional characteristics;
204 A. Chakrabarti
(3) compiling the identified biological structures and analysing each to extract
underlying principles and make preliminary solution associations; (4) transferring
these preliminary solutions into technical solutions according to the requirements
and conditions of the goal; (5) varying and combining relevant characteristics of
these solutions; (6) enlisting alternatives of each characteristic into a morpho-
logical table and identifying possible combinations of these characteristics; (7)
using common evaluation methods, evaluating the solution elements or completing
variants to select the best; and (8) elaborating the chosen solution. Schild et al.
(2004) propose a systematic method for identifying analogue solutions to a given
problem, with the following steps: (1) problem formulation at an appropriate level
of abstraction; (2) evaluation; (3) search for analogies; and (4) verification and
evaluation. Gramann (2004) recommends the following biomimetic process to
support technical problem solving: (1) formulate a search objective; (2) search for
and assign a set of relevant biological systems; (3) analyse these biological sys-
tems; and (4) evaluate.
Based on an analysis of the above processes, Sartori et al. (2010) construct a
generic model of the biomimetic process with these steps: (1) formulate search
objectives; (2) search for biological analogues; (3) analyse biological analogues;
and (4) transfer these analogues to the technical domain.
Several databases of biological systems have been developed to aid the biomi-
metic process. Catalogue sheets by Hill (1997, 2005) aim to capture knowledge
about biological structures and functions. With the goal of developing bio-TRIZ,
the database of biological effects by Vincent et al. (2002, 2006) uses TRIZ methods
(Terninko et al. 1998; Mann 2001). Further, Vincent et al. developed a framework
for capturing biological data in a technology-compatible manner, to support
designers in BID. One issue with these approaches, as Chakrabarti et al. (2011)
point out, is the distribution of biological functionality over several levels of scale
and complexity; as Sartori et al. (2010) argue, developing an appropriate functional
representation of biological systems suitable for engineering design seems to be a
major, unresolved issue. Neither in Hill’s catalogue sheets nor in Vincent et al.’s
database is there an explicit relationship between function and structure of bio-
logical systems with an objective basis. A possible resolution to this is the basis
provided by the SAPPhIRE model of causality (Chakrabarti et al. 2005).
SAPPhIRE model uses seven constructs to explain how system-functions are
achieved, see further details in Sect. 8.4.2. A database of over 1,000 entries of
biological and technical systems structured using SAPPhIRE model has been
developed for a software tool Idea-Inspire (see further in this section) to retrieve
relevant entries as inspiration for a given design problem (further details in Sect.
8.5). There are other approaches of cataloguing biological systems, such as func-
tion-based approaches (e.g. Vakili et al. 2007; Nagel et al. 2010), approaches that
use reverse engineering and ontologies (Wilson and Rosen 2007; Wilson et al.
2009), or SBF model-based approaches (Vattam et al. 2010b).
The database approach to biomimetics is not without criticism; Gramann
(2004), for instance, criticises these due to the vast amount of and variety in
biological knowledge to be captured—a massive task. The natural language-based
approach (e.g. by Hacco and Shu 2002; Chiu and Shu 2007; Cheong et al. 2010) is
a possible resolution to this issue. Their approach uses natural language processing
methods to analyse biological information already available in existing resources,
in order to extract relevant biological phenomena that can be transferred by
designers to the target domain by applying analogical reasoning. A comprehensive
review of this approach can be found in (Shu 2010).
According to Chakrabarti and Shu (2010), each approach has pros and cons.
While the natural language approach avoids the effort involved in structuring and
populating databases, it needs effort in developing appropriate search processes for
identifying meaningful information from the natural language resources, and
transfer may be more difficult. In contrast, the database approach needs substantial
effort into prestructuring information and populating databases, so that search for
relevant information in the use phase becomes easier, and transfer less difficult.
A set of guidelines for biomimetic transfer has been suggested by Sartori et al.
(2010); various guidelines for composition of biological systems have been pro-
posed by Vattam et al. (2010a). A number of software tools have also been
developed. For example, Idea-Inspire is an interactive, biomimetic-inspiration tool
that uses the database approach, structures the entries using the SAPPhIRE model
of causality (Chakrabarti et al. 2005; Sarkar et al. 2008; Srinivasan et al. 2011),
and provides biological systems as stimuli for ideation. DANE is a tool, also based
on the database approach, which uses SBF model for structuring information
(Vattam et al. 2010b).
In summary, current research in biomimetic support is focused on providing
three types of outputs:
• stimuli for biomimetic inspiration (e.g. Chakrabarti et al. 2005).
• framework for systematic BID (e.g. Srinivasan and Chakrabarti 2010b).
• guidelines for biomimetic transfer (e.g. Sartori et al. 2010).
In this chapter, we provide an overview of our research into each of these.
8.3 BID Process and ‘Transfer’
According to Sartori et al. (2010), the BID process has four general steps:
1. formulate search objectives;
2. search for biological analogues;
3. analyse biological analogues;
4. Transfer relevant knowledge to the target domain.
Our research is focused on the following:
• provide biological inspiration—Step 2 in the BID process (Sect. 8.5);
• provide a systematic design process for supporting BID (Sect. 8.6);
• provide systematic guidelines for transfer—Step 4 in the BID process
(Sect. 8.7).
206 A. Chakrabarti
8.4 SAPPhIRE Model, Creativity, and an Integrated

Model of Designing
For understanding the work in the rest of the chapter, knowledge of the following
is essential and is therefore briefly described below.
8.4.1 Creativity
Since the focus of the work reported in this chapter is on understanding designing,
especially BID, in terms of its influence on the various aspects of creativity, an
understanding of creativity and its various aspects is crucial. Based on the
‘common definition’ of creativity (Sarkar and Chakrabarti 2007b, 2011), creativity
in design ‘occurs through a process by which an agent uses its ability to generate
ideas, products or solutions that are novel and useful’. This ‘common’ definition is
developed after analysing an extensive set of definitions of creativity from liter-
ature, taking into account all encompassing features that constitute these defini-
tions. In this definition, the two major aspects of creativity are ‘novelty’ and
‘usefulness’; novelty means ‘being recent and original to the society’, while
usefulness refers to the ‘utilitarian value to the society’.
Novelty is assessed using a measure that estimates how different the idea,
product, or solution under assessment is from existing ideas, products, or solutions.
The assessment is based on the following assumption: if the product is different
from existing products in the functionality itself, that is, invents a new function, it
is of the highest novelty. If that is not the case, and the product is structurally no
different from existing products, then there is no novelty. If, however, it is in
between, SAPPhIRE model of causality (see Sect. 8.4.2) can be used to assess how
different the product is from existing products (Sarkar and Chakrabarti 2011).
More recently, a variant of this measure has been developed based entirely on the
SAPPhIRE model (Srinivasan and Chakrabarti 2010a).
Usefulness is assessed based on the assumption that, the value of a technical
product lies in its usefulness to the society. Usefulness can be assessed using the
extent to which it is used, or likely to be used. In this sense, usefulness of a product
is reflected in the extent of use of the product. This is assessed using these sub-
measures: how beneficial the use of the product is (for instance, a life-saving drug
is more beneficial than a lifestyle drug), how long the product is used or its effect
lasts (for instance, a drug whose intended effects last a lifetime is more useful than
one that has to be taken once every day to remain effective), how many people use,
or benefit from the product (for instance, a car which runs the same distance with
all seats full is more useful than one that runs similar distance with fewer seats
occupied) (Sarkar and Chakrabarti 2011).
Finally, technical creativity is taken as a product of the degree of novelty and

degree of usefulness; a product with a high degree of both is considered the most
creative, a product lacking in substantial novelty or usefulness is less creative, and
a product lacking in both is the least creative (Sarkar and Chakrabarti 2011).
All the above measures have been empirically validated by comparing the
responses of experienced designers on the relative novelty, usefulness, and crea-
tivity of multiple series of products against those estimated using the measures.
The assumption, similar to that made by Amabile (1983, 1996) has been that,
experts are the best judges for creativity, and therefore, any measure of design
creativity should match the collective, intuitive opinions of experienced designers.
8.4.2 SAPPhIRE Model of Causality
As mentioned in Sect. 8.2, SAPPhIRE (acronymed from State change–Action–

Part–Phenomenon–Input–oRgan–Effect) model of causality (Chakrabarti et al.
2005) was developed for explaining functioning of an entity that uses physical
phenomena for attaining its functions. It explains how an entity works, through
causal descriptions of how part-level characteristics of the entity and its sur-
roundings trigger physical phenomena that change the state of the entity and its
surroundings. The constructs of SAPPhIRE model are (Ranjan et al. 2012):
Parts Physical components and interfaces that constitute the
entity of interest and its surroundings.
Physical phenomenon An interaction between the entity and its surroundings.
State A property of the entity (or its surroundings) that is
involved in an interaction.
Physical effect A principle of nature that underlies and governs an
interaction.
Organ A set of properties and conditions of the entity and its
surroundings required for an interaction between them.
Input A physical variable that crosses the boundary of the
entity and is essential for an interaction between the
entity and its surroundings.
Action An abstract description or high-level interpretation of an
interaction between the entity and its surroundings.
The relationships among these constructs are as follows: Parts (P) of an entity
and its surroundings create organs (R), which are the structural requirements for a
physical effect (E). A physical effect is activated by various inputs (I) on the
organs and creates a physical phenomenon (Ph), and changes the state (S) of the
entity. The changes of state are interpreted as actions (A), as new inputs, or as
changes that create or activate parts (see Fig. 8.1).
208 A. Chakrabarti
Fig. 8.1 SAPPhIRE model

of causality (Source
Chakrabarti et al. 2005)
The use of these constructs together and their links with functionality provides a
richer description of behaviour. Action, state change, and input constitute the
higher levels of abstraction. Physical phenomenon and effect comprise the inter-
mediate levels of abstraction. Organs and parts constitute the lower levels of
abstraction (Srinivasan and Chakrabarti 2010a). This model of causality explains
the functioning of a system as follows (Srinivasan and Chakrabarti 2009): a set of
components and interfaces that constitute a system and its environment (parts)
creates a set of properties and conditions of the system and its environment
(organs). When the system and its environment are not in equilibrium, there is a
transfer of a physical quantity in the form of material, energy, or signal (input)
across the system boundary. This physical quantity, in combination with a par-
ticular set of properties and conditions (organs), activates a principle (physical
effect). Activation of this principle creates an interaction between the system and
its environment (physical phenomenon). The interaction between the system and
its environment creates a change in property of the system (state change). The
change in property can be interpreted at a higher level of abstraction (action).
The example below is used for clarifying the model. Let us assume that some
water is kept in an electric kettle (parts). The relevant properties and conditions
that are available from the parts include the fluidic property of the water, the
electrical resistance properties of the coil, the heat transfer properties of the water
and the coil, etc. (organs). When the kettle is switched on, electric current flows
through the coil (input), and due to the resistance properties of the coil, activates
the resistive-heating effect (effect). This leads to exchange of heat within the coil
(phenomenon) which leads to an increase in its temperature (state change) that
may be interpreted as heating of the coil (action). The temperature difference
between the coil and the water (input), along with the heat transfer properties of
the water and the coil, activates heat conduction effect (effect), which leads to heat
exchange between the coil and the water (phenomenon) that leads to an increase in
the temperature of the water (state change), which is interpreted as heating of
water (action).
8.4.3 Integrated Model of Designing
While a model of causality explains how an existing entity satisfies its goals, that is,
its intended effects on the surroundings, designing is a process in which, starting
from its goals, descriptions for an appropriate entity have to be worked out.
Designing, therefore, should involve working out the details of the intended entity
at the various levels of abstraction that a model of causality requires, but in reverse
order, so as to end up with sufficient detail about its lowest—part level of
abstraction. Based on this assumption, a model of designing has been developed
that uses the levels of outcome abstraction, provided by the SAPPhIRE model, for
an entity to be designed. It is called the Integrated Model of designing (also called
the GEMS of SAPPhIRE as Req-Sol model), where GEMS(Generate–Evaluate–
Modify–Select/reject) activities are applied on SAPPhIRE levels of outcome which
evolve as Req-Sol (Requirements or Solutions), see (Srinivasan and Chakrabarti
2010b). The Integrated Model of designing has been empirically validated using
video–protocol studies of designing sessions that occurred well before the model
was proposed. This was to ensure that the way designers worked could not have
been biased by the knowledge of this model. A two-way comparison was used for
validation: whether all constructs of the model were present in the events that
constituted the designing sessions, and whether all events could be described using
the constructs of the model. The model was found to represent the natural processes
used by designers for conceptual design of technical products.
8.4.4 SAPPhIRE Abstraction Levels, Number of Ideas,

and Novelty
A series of designing sessions were analysed using the Integrated Model of

designing and the proposed measures of creativity (Srinivasan and Chakrabarti
2010a, b). The intention was to see what aspects of designing influenced which
aspects of creativity. In particular, two aspects of design outcomes were investi-
gated: the number of ideas explored and the levels of abstraction (in this case,
210 A. Chakrabarti
SAPPhIRE levels of outcome abstraction) at which these ideas belonged. Simi-

larly, two aspects of creativity were investigated: variety of solutions generated
and novelty of solutions generated. Variety is used to mean how different, on
average, the solutions generated by a design team or individual in a given
designing session have been from one another; novelty of a solution space is used
to mean how different on average the solutions generated by a designer or design
team have been from the solutions that already existed before the designing session
started.
The number of ideas generated had a significant correlation with both variety
and novelty of the solution space; this means that generating more ideas is likely to
lead to greater variety and novelty. It was also found that the number of ideas at a
higher level of outcome abstraction correlated more strongly to the variety and
novelty of the solution space; this means that being able to generate more ideas at
higher levels of outcome abstraction should have a more significant impact on
variety and novelty of solutions generated than with a large number of ideas at a
lower level of outcome abstraction (Srinivasan and Chakrabarti 2010a).
Analyses of the designing sessions led to another interesting finding. For the six
designing sessions that were analysed using the Integrated Model, the number of
ideas generated at the various levels of outcome abstraction, starting from action
and input (extreme left in the plot in Fig. 8.2), through the intermediate levels to
the part level (extreme right) was plotted; the six cases are marked as 1–6. For
every single designing session analysed, each of which involved a different
combination of design problem and designers, the distribution of the number of
ideas generated at the various levels of abstraction has been remarkably similar.
While the number of ideas at the action and part levels has been relatively high for
every design session, the number of ideas generated at the intermediate levels has
been substantially lower. Since for carrying out every action, multiple state
changes could be utilised; for every state change to be carried out, multiple
phenomena could be used; and so on for each level of abstraction, we expected that
180
160
140 [1]
120 [2]
100
[3]
80
[4]
60
40 [5]
20 [6]
0
a(i) s ph e r p
Fig. 8.2 ‘Bathtub’ curve (Source Srinivasan and Chakrabarti 2010a)

the number of ideas should have increased monotonically as the designers pro-
gressed from the higher levels to the lower levels of abstraction. However, this did
not happen. Why did this not happen, and how did that affect creativity?
We argue that there are two possible reasons as to why, instead of monotoni-
cally increasing as one progressed from higher to lower levels of abstraction, the
number of ideas generated followed a ‘bathtub’ curve as in the Fig. 8.2:
• Conjecture 1: The high number of action-level ideas comes from the design
brief. The high number of part-level ideas comes from designer competence in
mapping action-level ideas to part-level ideas. The relatively small number of
ideas at the remaining levels comes from a lack of designer competence at
mapping ideas at action or part levels to those at these intermediate levels.
• Conjecture 2: The high number of action-level ideas comes from the design
brief. The high number of part-level ideas comes from designer competence in
mapping action-level ideas to part-level ideas, and the belief that this would
produce less risky and more feasible solutions fast, which going into the fun-
damentals (via intermediate levels) might sacrifice.
If Conjecture 1 is true, improving designer competence would impact novelty

directly as there would be more number and variety of ideas generated. This would
impact usefulness indirectly, since more ideas with greater novelty would stand a
greater chance for stumbling onto solutions with substantially higher usefulness. If
Conjecture 2 is true, this designer attitude would have an indirect, incremental
impact on usefulness while sacrificing variety and novelty. Even if Conjecture 2 is
true, this designer attitude would be harder to change if there is lack in designer
competence in exploring ideas across the levels. Hence we take Conjecture 1 to be
more reasonable and develop support to improve designer competence to explore
ideas across all levels of outcome abstraction.
Supporting this requires two kinds of knowledge: product (or, domain) knowl-
edge—knowledge about how various entities work, and process knowledge—
knowledge with which domain knowledge can be modified (Srinivasan et al. 2011);
both are important for designing. In order to provide BID-related domain knowl-
edge, a tool called Idea-Inspire has been developed, which can be used to provide
biological and technical stimuli, see Sect. 8.5. In order to provide process knowl-
edge, a systematic approach is necessary that provides a process framework, within
which this domain knowledge can be provided in a situated manner. A framework
for this purpose has been developed and discussed in Sect. 8.6.
8.5 Supporting BID Using Stimuli: Idea-Inspire
Idea-Inspire (Chakrabarti et al. 2005) is a software tool that is used for supporting
ideation using knowledge about biological and technical systems as stimuli. Idea-
Inspire has three components.
212 A. Chakrabarti
The first component is a database of entries, each of which explains how an

entity achieves its goals; the information is provided using multiple modes: video/
animation, photographs or drawings, natural language description, and a SAP-
PhIRE model–based description, which is primarily used for computer-based
search for relevant entries (Chakrabarti et al. 2005).
The second component is an analogical search algorithm, which takes a user
description of the problem, described at one or a combination of SAPPhIRE levels
of outcome abstraction, and retrieves a series of entries that are relevant to the
problem description.
The third component is a graphical user interface, through which users interact
with the software, in terms of specifying the problem, receiving the list of relevant
entries, or looking at the details of these entries. Figures 8.3 and 8.4 show,
respectively, the GUI and one entry from Idea-Inspire.
Ideation effectiveness of Idea-Inspire has been evaluated in two stages. While
details have been given in Sarkar et al. (2008), an excerpt is provided below. In the
first stage of evaluation, three designers were asked to generate as many ideas as
possible to solve given design problems without any support; when they were
‘exhausted’ (according to them), they were asked to generate further ideas using
entries retrieved by Idea-Inspire as stimuli. On average, each designer produced
165 % additional ideas using Idea-Inspire entries as stimuli. This indicated
enhancement of ‘fluency’ of the designers, which is often used as a measure of
Fig. 8.3 Graphical user interface for idea-inspire

Fig. 8.4 An entry from idea-inspire (Source Chakrabarti et al. 2005)
novelty (e.g. Csikszentmihalyi 1997; Sternberg and Lubart 1999), which in turn is
a measure of creativity according to the ‘common’ definition of creativity.
More recently, influence of fluency on novelty in the context of conceptual
design has been empirically verified by Srinivasan and Chakrabarti (2010a); they
found that the number of ideas explored strongly correlates with the variety of the
solutions generated, which correlates with the novelty of the solutions generated. It
was also found, in the Idea-Inspire study, that about 42 % of the ideas generated by
the designers with stimuli from Idea-Inspire were selected by themselves as worth
developing further; this points to the likely usefulness of these ideas—the other
indicator for creativity according to the common definition.
In the second stage of evaluation, twelve masters-level design students with an
undergraduate degree in engineering and up to a year of industrial experience were
asked to use Idea-Inspire in their respective, mandatory design projects. Feedback
from these students, on the usefulness and usability of the software for their
purpose, indicated both an enhancement in the number of new ideas generated and
ease of use of the software.
214 A. Chakrabarti
8.6 Systematic Design Process: An Integrated Framework

of Designing
As discussed in Sect. 8.4.4., empirical studies of the conceptual design process

(Srinivasan and Chakrabarti 2010b) revealed that, while designers explored the
action and part levels of outcome abstraction extensively, they did not adequately
explore the intermediate levels of abstraction.
From earlier studies (e.g. Srinivasan and Chakrabarti 2010a), it was known that
exploration at higher levels of abstraction has a greater impact on novelty of
solutions generated. It therefore became clear that a systematic approach providing
process knowledge was necessary to encourage designers to explore ideas at all
levels of abstraction. The Integrated Framework for designing (also called the
GEMS of SAPPhIRE framework, see Srinivasan and Chakrabarti 2010b; Srini-
vasan et al. 2011) has been developed for this purpose. The framework prescribes
designers to apply GEMS to requirements at each SAPPhIRE level of outcome
abstraction (this stage is called requirement development stage or RDS) and apply
GEMS to solutions at each SAPPhIRE level of outcome abstraction (this stage is
called solution development stage or SDS), with the aim of exploring as many
ideas as possible for requirements and solutions at each level of outcome
abstraction (for details, see Srinivasan et al. 2011). Figure 8.5 depicts these two
stages: RDS followed by SDS, where G (generate), E (evaluate), M (modify), and
S (select/reject) activities are prescribed to be applied to requirements (shown as
‘req’) and solutions (shown as ‘sol’), respectively, at all levels of outcome
RDS G[req(a/s/ph/e/i/r/p/others)]
E[req(a/s…others)
contradiction, feasibility]
M[req(a/s…others)] S[req(a/s…others)]
SDS
req(a) sol(s) sol(e)
G[sol(a)] G[sol(ph)] G[sol(i/r)]
E[sol(a) req(a)] E[sol(ph) req(ph)] E[sol(i/r) req(i/r)]
M[sol(a)] S[sol(a)] M[sol(ph)] S[sol(ph)] M[sol(i/r)] S[sol(i/r)]
sol(a) sol(ph) sol(r)
G[sol(s)] G[sol(e)] G[sol(p)]
E[sol(s) req(s)] E[sol(e) req(e)] E[sol(p) req(p)]
M[sol(s)] S[sol(s)] M[sol(e)] S[sol(e)] M[sol(p)] S[sol(p)]
Fig. 8.5 Integrated Framework for designing (Source Srinivasan et al. 2011)
Table 8.1 Number of ideas generated when various types of support are used: d stands for
designer; f for Integrated Framework; i for Idea-Inspire; B for biologically inspired ideas; NB for
non-biologically inspired ideas
Mobility Steering Handling gradient Stability Overall
B NB total B NB total B NB total B NB total B NB total
d 0 0 0 0 0 0 1 4 5 0 2 2 1 6 7
f?d 0 7 7 0 5 5 0 0 0 0 3 3 0 15 15
i?d 0 0 0 0 0 0 2 0 2 0 0 0 2 0 2
f ? i + d 11 3 14 3 5 8 2 1 3 1 0 1 17 9 26
11 10 21 3 10 13 5 5 10 1 5 6 20 30 50
(Source Srinivasan et al. 2011)
abstraction, while going from action level (shown as (a)), through state change
(shown as (s)), phenomena (ph), effect (e), organ (r) levels, to part (p) level, in that
order. The framework proposes that designers generate, evaluate, modify as nec-
essary, and finally select appropriate requirements at all levels of SAPPhIRE
abstraction and then do the same for solutions, working gradually at more concrete
levels of the solution as they progress from action to part level (Fig. 8.5).
Two types of evaluations were used to test the effectiveness of the Integrated
Framework: comparative study in laboratory setting where various design teams
used various interventions (no support, framework only, and both framework and
Idea-Inspire), see Srinivasan (2011). On average, number, variety, and novelty
increased in the cases in which either the Integrated Framework was used alone, or
used along with Idea-Inspire, compared to when no support was used.
In the second study, a group of designers used the Integrated Framework and
Idea-Inspire during an industrial project (Srinivasan et al. 2011). Designers ini-
tially developed some ideas on their own and then started using the Integrated
Framework; they had the freedom to decide when they wanted to use Idea-Inspire
and when not, within the Integrated Framework. The results of evaluation are
detailed in Srinivasan et al. (2011) and summarised here (see Table 8.1) as follows:
Designers generated very few ideas on their own—most of them non-biological
ideas. Using the framework alone, this number doubled although still using only
non-biological ideas as inspiration. When they used Idea-Inspire only, few ideas
were generated—but all of these were inspired by biological stimuli. When the
Integrated Framework and Idea-Inspire were used together, the number quadru-
pled, with a dominance of biological ideas. This indicates that using the Integrated
Framework and Idea-Inspire together substantially improves ideation and balances
the generation of ideas that are inspired by biological as well as non-biological
stimuli.
216 A. Chakrabarti
8.7 Guidelines: SAPPhIRE-Based Guidelines

for ‘Transfer’
While various guidelines are provided in literature as to how various steps in BID
are (to be) carried out, relatively little is specified as to how transfer is carried out.
The definition of ‘transfer’, taken from Sartori et al. (2010) and based on Schmidt
(2005), is as follows: it is the reproduction of information from a model of a
biological system in a model or prototype for a technical system.
According to Schild et al. (2004), there are four possible levels of transfer:
transfer of technology to a new context, transfer of structure, partial transfer of
functional principles, and use of an analogy as idea stimulus. According to Vattam
et al. (2010a), transfer is guided by the sub-functions identified. However, liter-
ature provides little further detail as to what kind of knowledge is transferred
within each single sub-function, or what the steps of transfer are.
In order to understand the process of transfer is greater detail, we asked the
following questions (Sartori et al. 2010):
• What kind of knowledge is transferred in biomimetics?
• How can the step ’transfer’ in the biomimetic process be supported?
SAPPhIRE model of causality (Chakrabarti et al. 2005) is used to analyse
twenty existing cases of biomimetic transfer from literature, to understand the
types of knowledge used, and the generic process of transfer. The levels of
abstraction at which transfer took place are identified, and a guideline is developed
to enhance fluency of transfer. Based on this analysis, five types of transfers have
been identified to have taken place (for details, see Sartori et al. 2010):
• Transfer parts: This is direct mimicking of a biological system in a technical
system. The same materials are used and arranged in the technical system in the
same way as in biology. This is transferred at the lowest level of abstraction.
• Transfer organs: This involves developing a technical system with similar
organs as in its biological analogue.
• Transfer attributes: This involves developing a technical system with the same
or similar attributes as its biological analogue. Attributes are properties of the
biological analogue that are not clearly connected to any physical effects.
• Transfer state changes: This involves using a state change of a biological ana-
logue by the technical system in order to achieve an analogue action.
• Resulting (or incidental) transfer: This is the case where the technical system
resulting from an associated transfer provides adequate means for a new action
to be also supported.
Note that the above types mark the highest level of abstraction at which transfer
was found explicitly to occur; in many of these cases what was transferred,
implicitly or explicitly, was not just the knowledge for that level but a combination
of knowledge from multiple levels of abstraction including that level; in a broader
sense, these are similar to ‘design patterns’ of Alexander (1977), which were
proposed by Goel and Bhatta (2004) to be useful in analogy-based designing.
Transfers that were carried out in above the twenty cases are distributed as
follows. Most transfers took place at the part, attribute, and organ levels, with very
few at state change levels. Since variety at a higher level of outcome abstraction
has a greater impact on novelty, this leaves scope for improvement. Hence, a new,
SAPPhIRE-based set of guidelines for analogical transfer has been developed.
Two sets of guidelines have been developed (Sartori et al. 2010). The ‘Generic
Guideline’ has been developed to encapsulate the essential steps for carrying out
BID, and the recommendations specific to each of these steps as found from
existing literature. The Generic Guideline therefore encapsulates the current
wisdom as to how to carry out BID, as weaned out of existing literature.
The second, new ‘Guideline with SAPPhIRE’ has been proposed to follow the
same generic steps as in the Generic Guideline, but with specific recommendations
for using SAPPhIRE constructs within these steps. This guideline recommends
that the four classes of transfer identified be systematically used in the analysis and
transfer steps. This, it is argued, should increase the number of BID alternatives
generated, see details in Sartori et al. (2010).
A series of design studies have then been carried out to evaluate these guide-
lines, by comparing the performance of designers when using SAPPhIRE-based
guideline in carrying out BID, with that when using the Generic Guideline. The
study has been carried out in both India and Germany, using two different groups
of designers. In both the countries, the number of concepts produced when
SAPPhIRE guidelines were used was higher, by between 50 and 100 % from when
Generic Guideline was used. Not all of these solutions were biologically inspired
or feasible; for the concepts that were biologically inspired and feasible, the
average increase in the number of concepts produced was 60 % when SAPPhIRE-
based guidelines were used. More importantly, the percentage of transfer at the
higher levels of abstraction (from part and attribute to organ and state change)
improved steadily from the cases in literature to when the Generic Guideline was
used, and even further when the SAPPhIRE guidelines were used. This indicates
that there is potential for greater novelty in using these guidelines.
8.8 Summary and Conclusions
This chapter provides a brief overview of descriptive and prescriptive BID

research and then elaborates on descriptive and prescriptive BID research in the
research group of the author. The work produced measures for creativity and
models of causality and designing, which were used to understand how the out-
come levels of abstraction at which ideas are explored in current ways of designing
influence creativity, and at what levels of outcome abstraction do analogical
transfer currently take place. Using these as the bases, the following have been
developed: an Integrated Framework for designing, a set of guidelines for
218 A. Chakrabarti
supporting analogical transfer, and a database and associated software tool for
providing biological stimuli for ideation. Comparative, empirical studies using
these as interventions indicate substantial potential for their use as support in
enhancing creativity in developing technical products.
References
Alexander, C. (1977) A pattern language: towns, buildings, construction. Oxford University

Press, Oxford
Amabile TM (1996) Creativity in context. Westview Press, Colorado
Amabile TM (1983) The social psychology of creativity. Springer, New York
Benami O, Jin Y (2002) Creative stimulation in conceptual design. In: Proceedings of ASME
design engineering technical conferences and computers and information in engineering
conference (DETC/CIE), Montreal, DETC2002/DTM-34023
Blessing LTM, Chakrabarti A (2009) DRM, a design research methodology. Springer, London
Chakrabarti A, Bligh TP, Holden T (1992) Towards a decision-support framework for the
embodiment phase of mechanical design. AI in Eng 7(1):21–36
Chakrabarti A (2000) Increasing efficiency of compositional synthesis by improving the database
of its building blocks, AI EDAM, vol 14. Cambridge University Press, Cambridge,
pp 403–414
Chakrabarti A (2001) Improving efficiency of procedures for compositional synthesis using
bi-directional search. AI EDAM 15(1):67–80
Chakrabarti A (2004) A new approach to structure sharing. ASME JCISE 1(1):1–78
Chakrabarti A, Shu LH (2010) Guest editorial: biologically inspired design. AI EDAM
24:453–454
Chakrabarti A, Tang, MX (1996) Generating conceptual solutions on FuncSION: evolution of a
functional synthesiser. In: Gero S, Sudweeks F (eds) Artificial intelligencedesign’96. Kluwer
Academic Publishers, Dordrecht, pp 603–622
Chakrabarti A, Singh V (2007) A method for structure sharing to enhance resource-effectiveness.
J Eng Des 18(1):73–91
Chakrabarti A, Cagan J, Shea K, Campbell M, Stone R, Vergas-Hernandez N, Wood K (2011)
Computer based design synthesis research: an overview. ASME JCISE 11(2):021003-
1–021003-10
Chakrabarti A, Morgenstern S, Knaab H (2004) Identification and application of requirements
and their impact on the design process: a protocol study. Res Eng Design 15:22–39
aiding biomimetic and artificial inspiration of new ideas. AI EDAM 19(2):113–132
CheongH, ChiuI, Shu LH (2010) Extraction and transfer of biological analogies for creative
concept generation. In: Proceedings of ASME international design engineering technical
conferences and computers and information in engineering conference, Montreal, QC,
Canada, 15–18 Aug 2010, DETC2010-29006
domain information retrieval. AI EDAM 21:45–59
Csikszentmihalyi M (1997) Finding flow: the psychology of engagement with everyday life.
BasicBooks, New York
Fricke G (1996) Successful individual approaches in engineering design. Res Eng Design
8(3):151–165
Goel AK, Bhatta SR (2004) Use of design patterns in analogy-based design. Adv Eng Inform
18(2):85–94
Gordon WJJ (1961) Synectics. Harper & Row, NY
Gramann J (2004) Problemmodelle und Bionik als Methode. Dissertation, TU, Munich
Hacco E, Shu L (2002) Biomimetic concept generation applied to design for remanufacture. In:
Proceedings of ASME design engineering technical conferences and computers and
information in engineering conference, Montreal, 29 Sept to 2 Oct 2002, DETC2002/DFM-
34177
Helms ME, Vattam SS, Goel AK (2009) Biologically inspired design: process and products. Des
Stud 30(5):606–622
Hill B (1997) Innovationsquelle natur: naturorientierte innovationsstrategie für entwickler.
Konstrukteure und Designer, Shaker, Aachen, Germany
Hill, B. (2005) Goal setting through contradiction analysis in the bionics-oriented construction
process. CIM, 14, 19, 59–65. Blackwell Publishing Ltd, Oxford/USA, p 33
Hon KKB, Zeiner J (2004) Knowledge brokering for assisting the generation of automotive
product design. Annals CIRP 53(1):159–162
Mann D (2001) System operator tutorial—9—windows on the world. The TRIZ Journal. http://
www.triz-journal.com/archives/2001/09/c/index.htm as at 20 April 2010
Nagel JKS, Nagel RL, Stone RB, McAdams DA (2010) Function-based, biologically-inspired
concept generation. AI EDAM 24(4):521
Nidamarthi S, Chakrabarti A, Bligh TP (1997) The significance of co-evolving requirements and
solutions in the design process. In: International conference of engineering design pp 227–230
Ranjan BSC, Srinivasan V, Chakrabarti A (2012) An extended, Integrated Model of designing.
In: Horváth AA, Behrendt M, Rusák Z (eds.) TMCE 2012, Karlsruhe, Germany
Sarkar P, Chakrabarti A (2011) Assessing design creativity. Design stud 32(4): 348–383
from nature. ASME J Inf Sci Eng 8:3
Sarkar P, Chakrabarti A (2007a) Understanding search in design. In: International conference on
engineering design (ICED07), Paris, France
Sarkar P, Chakrabarti A (2007b) Development of a method for assessing design creativity. In:
International conference on engineering design (ICED07), Paris, France
design. AI EDAM 24:483–505
Schild K, Herstatt C, Lüthje C (2004) How to use analogies for breakthrough innovations.
Institute of Technology and Innovation Management, Technical University of Hamburg
Schmidt JC (2005) Bionik und Interdisziplinarität. Wege zu einer bionischen Zirkulationstheorie
der Interdisziplinarität. In: Rossmann T, Tropea C. (Hrsg.): Bionik. Aktuelle For-
schungsergebnisse aus Natur-, Ingenieur- und Geisteswissenschaften. Springer, Berlin,
pp 219–246
Shu LH (2010) A natural-language approach to biomimetic design. Artif Intell Eng Des Anal
Manuf 24:507–519
Srinivasan V, Chakrabarti A (2010a) Investigating novelty-outcome relationships in engineering
design. AI EDAM 24(2):161–178
Srinivasan V (2011) Supporting novelty in conceptual phase of engineering design, Ph D Thesis,
Indian Institute of Science, Bangalore, India
Srinivasan V, Chakrabarti A (2010b) An Integrated Model of designing. JCISE 10(3):031013
Srinivasan V, Chakrabarti A (2009) SAPPhIRE—an approach to analysis and synthesis. In:
Proceedings of international conference on engineering design [ICED09], Stanford, USA
Srinivasan V, Chakrabarti A, Pal U, Ranjan BSC, Ojha SP, Ranganath R (2011) Supporting
process and product knowledge in biomimetic design. Int J Design Eng 4(2):132–158
Sternberg RJIT, Lubart TI (1999) The concept of creativity: prospects and paradigms. In
Sternberg RJ (ed) Handbook of creativity. CUP, NY, pp 3–16
Terninko J, Zusman A, Zlotin B (1998) Systematic innovation. An introduction to TRIZ (Theory
of Inventive Problem Solving). St. Lucie Press, Boca Raton
stimulation of idea generation in design. Des Stud 29:203–221
Ulrich KT, Seering WP (1990) Function sharing in mechanical design. Des Stud 11(4):223–234
220 A. Chakrabarti
Vakili V, Chiu I, Shu LH, McAdams D, Stone R (2007) Including functional models of biological
phenomena as design stimuli. In: Proceedings of ASME international design eng. technical
conferences and computers and information in engineering conference, Las Vegas, NV, USA,
Sept 4–7 2007, DETC2007-35776 (DTM)
Vattam SS, Helms ME, Goel AK (2008): Compound analogical design: interaction between
problem decomposition and analogical transfer in biologically inspired design. In: Proceed-
ings of 3rd international conference on design computing and cognition, Atlanta, Springer,
Berlin, pp 377–396
Vattam S, Wiltgen B, Helms M, Goel AK, Yen J (2010a) DANE: fostering creativity in and
through biologically inspired design. In: Procedings of 1st International conference on design
creativity (ICDC2010), Kobe, Japan
Vattam SS, Helms ME, Goel AK (2010b) A content account of creative analogies in biologically
inspired design. AI EDAM 24:467–481
Vincent J, Mann D (2002) Systematic technology transfer from biology to engineering. Philos
Trans R Soc Phys Sci 360:159–173
Vincent JFV, Bogatyreva OA, Bogatyrev NR, Bowyer A, Pahl AK (2006) Biomimetics: its
Wilson J, Chang P, Yim S., Rosen D (2009) Developing a bio-inspired design repository using
ontologies. In: Proceedings of ASME design eng. technical conferences and computers and
information in engineering conference, IDETC/CIE, San Diego, DETC2009-87272
Wilson JO, Rosen D (2007) Systematic reverse engineering of biological systems. In:
Proceedings of ASME design engineering technical conferences and computers and
information in engineering conference (IDETC/CIE), Las Vegas, DETC2007/DTM-35395
Chapter 9
Overcoming Cognitive Challenges
in Bioinspired Design and Analogy
Julie S. Linsey and Vimal K. Viswanathan
Abstract Bioinspired design and analogy are powerful tools for innovation.
Engineers face many cognitive challenges when seeking to employ design by
analogy and bioinspired design. This chapter presents known difficulties engineers
must overcome for bioinspired design and summarizes the cognitive psychology,
multi-media learning and design evidence for the cognitive challenges. A number
of cognitive challenges block a designer from being effective when using design
by analogy. The challenges range from retrieving appropriate analogues based on
deep similarities to the challenge of seeing multiple solutions based on a single
analogue, to becoming fixated on initial solutions. Like any other idea generation
process, design fixation limits the solution space explored during design by
analogy and bioinspired design. There are empirically proven strategies for miti-
gating design fixation ranging from presenting uncommon examples to abstrac-
tions and categories of solutions. From research on multimedia learning and
design, additional heuristics applicable to the design of new bioinspired tools have
also been identified. These include annotations directly next to ambiguous or
unfamiliar representations to enhance communication and make learning easier.
Design heuristics and principles are presented after each section of the relevant
research. The chapter ends with the summary of the cognitive design heuristics for
bioinspired design methods and tools. This set of heuristics can be used as
guidelines for researchers developing new methods and support tools for bioin-
spired design.
Keywords Design by analogy Cognitive biases Fixation
J. S. Linsey (&) V. K. Viswanathan

Georgia Institute of Technology, Atlanta, USA
e-mail: julie.linsey@me.gatech.edu
V. K. Viswanathan
e-mail: vimkucv@gmail.com

222 J. S. Linsey and V. K. Viswanathan
9.1 Introduction
Bioinspired design and innovation through analogy are powerful tools for design,
as demonstrated by the anecdotal and empirical data (Casakin and Goldschmidt
1999; Leclercq and Heylighen 2002; Christensen and Schunn 2007; Basalla 1988).
Bioinspired design is a specific type of design by analogy which is based on
analogies from nature. Currently, significant work aims to create computer tools
and design methods to enhance and support this critical process (Hacco and Shu
2002; Chiu and Shu 2007; Chakrabarti et al. 2005a, b; McAdams and Wood 2002;
Cheong et al. 2011; Goel and Bhatta 2004; Linsey et al. 2012; Oriakhi et al. 2011;
Linsey et al. 2008a; Helms et al. 2009; Nagel et al. 2010; Vattam et al. 2010).
Designers who are attempting to implement bioinspired design and design by
analogy face a number of cognitive biases and challenges. Effective tools for
bioinspired design and analogy must be effectively designed to overcome these
cognitive biases and challenges. This chapter will describe the cognitive consid-
erations that design researchers must be aware of and then list a set of heuristics
for analogy tool design.
This chapter begins with a cognitive description of the design by analogy
process and the associated challenges including retrieving appropriate analogues,
creating multiple solutions based on a given analogue, and selecting design fea-
tures to copy. The next section then discusses the road blocks created by design
fixation and ways to mitigate its effects. The final section presents other consid-
erations and heuristics from research in multimedia learning and design. The
chapter ends with a summary of the cognitive challenges and the associated
heuristics and principles.
9.2 Analogy
The anecdotal and empirical evidence consistently demonstrates that professional

designers often use analogies and it is a powerful tool for innovation (Casakin and
Goldschmidt 1999; Leclercq and Heylighen 2002; Christensen and Schunn 2007;
Basalla 1988). Further, controlled experimental studies indicate that the natural-
istic studies are likely underestimating the actual frequency of analogy use, since
naturalistic protocol studies rely on either the designers reporting the use of
analogy or some direct indication of it (e.g., stating a plane is like a bird) (Linsey
2007; Linsey et al. 2008b).
9.2.1 Cognitive Process Model for Analogical Reasoning
Psychology has sought to understand the cognitive processes people use to create
and understand analogies (Falkenhainer et al. 1989; Gentner and Markman 1997;
Hummel and Holyoak 1997). Figure 9.1 shows the basic steps involved in
9 Overcoming Cognitive Challenges in Bioinspired Design and Analogy 223
Fig. 9.1 Steps in human

reasoning by analogy
processes
reasoning by analogy. Analogy is generally viewed as a comparison between two

products in which their relational or causal structure match (Holyoak and Thagard
1989; Gentner 1983; Falkenhainer et al. 1989). The problem domain is the target
of the analogy and the domain of prior knowledge that provides a potential
solution to the problem is called the source. Research on analogy suggests that
people first find a mapping between the relations in the source and the target. On
the basis of this mapping, aspects of the target may be re-represented to make them
more similar to the source. Furthermore, inferences about the target (i.e., potential
solutions) may be made based on the similarity of the target to the source. The
potential for creative problem solving is clearest when the two domains being
compared are very different on the surface, though the same process of comparison
can also be used for domains that share significant surface similarity (Gentner and
Markman 1997).
Designers face many challenges during the analogical reasoning process.
Retrieving an appropriate analogue from memory has often been shown to be the
most challenging step (Holyoak and Thagard 1989; Gentner 1983; Gentner and
Markman 1997; Falkenhainer et al. 1989; Holyoak and Koh 1987; Gick and
Holyoak 1980; Markman and Gentner 1993b). Search engines and other bioin-
spired design tools currently in development can assist with this (Hacco and Shu
2002; Chiu and Shu 2007; Chakrabarti et al. 2005a; Chakrabarti et al. 2005b;
McAdams and Wood 2002; Cheong et al. 2011; Goel and Bhatta 2004). People
also tend to focus too much on surface features instead of deep similarities
(Gentner and Landers 1985; Gick and Holyoak 1980; Helms et al. 2009). For
example, when using a bird as an analogue for human flight, the color of the bird is
a surface feature. Even once an effective analogue has been retrieved from either
memory or with computer tools, studies have shown that there is a difficulty in
selecting appropriate design features to map from the analogue to the problem
(Linsey et al. 2006; Linsey et al. 2007; Mak and Shu 2008; Cheong and Shu 2009;
Helms et al. 2009). Inaccurate models, both in scientific knowledge and the
designers’ knowledge can cause the wrong features to be selected and mapped.
Linsey et al. (Linsey et al. 2007) found that senior undergraduate mechanical
engineers who were required to use an airplane as an analogy and to map design
features that cause lift would occasionally select incorrect features such as the
smoothness of the plane or the propellers and try to map these features to the
solution rather than the shape of the wing for lift. This was in spite of the fact that
these students had taken and passed fluid dynamics where this material would have
been covered.
Heuristics
• Deep Connections—Encourage deep and functional connections between the
analogue and the problem.
• Matching Features—Highlight appropriate features that should be mapped from
the analogue to the problem.
9.2.2 Inaccurate Mental Models and Naïve Physics
Since at least da Vinci’s time, birds were recognized as good analogues for human
flight, but no one found a practical solution. The Wright brothers recognized that
birds’ twisting of their wings for control was an effective means to successful
control of their plane, rather than feathers or flapping being the key to success.
Often due to designers’ inaccurate mental models, an appropriate analogy will be
identified, but an effective solution will not result. People have mental models of
the world that are highly efficient for the tasks they frequently complete; but
generally inaccurate unless significant experience or training has occurred
(Markman 1999; McAfee and Proffitt 1991; Chi et al. 1981). The Wright brothers
had a rough theory of aerodynamics providing them with a better mental model of
flight and also guiding them to select appropriate features from birds.
Significant research in psychology has sought to understand how people reason
about the physical world and this directly impacts the concepts engineers develop
along with their evaluation. The areas of mental models and Naïve physics seek to
understand how people reason about the world around them (e.g., Gentner and
Stevens 1983; Forbus 1984; Kuipers 1994). Mental models of physical systems are
internal mental representations of external systems (Markman 1999). The area of
Naïve physics has produced some rather surprising findings about people’s mental
models which have direct implications for engineering design research. An
important finding is people’s mental models of physical phenomena, which they
observe frequently, can be surprisingly inaccurate. For example, when asked to
draw the water level of a glass tipped on its side, over 40 % will not be able to
indicate that the water line will be parallel to the horizontal plane (Markman 1999;
McAfee and Proffitt 1991). Similarly, inaccurate mental models of home-heating
systems have also been observed. Many people have the mental model of a home
thermostat like that of a car’s accelerator: the higher the thermostat is set, the
faster the house is heated (Kempton 1986). In reality, most home-heating and
home-cooling systems are either on or off. This mental model has a direct real-
world impact on people’s behavior and the environment. People with this mental
model tend to change the thermostat frequently throughout the day which is
inefficient and wastes energy. Limits in the capability of people’s mental models
have also been shown with highly trained scientists (Hutchins 1995).
Naïve physics’ implications for engineering design are that engineers’ mental
models are not likely to be highly accurate unless they have been repeatedly tested
through either extensive experience in situations where more accurate mental
models were required or through education. This is true for phenomena that
engineers may have observed many times. The challenge is even greater for
aspects of biology that engineers are not at all familiar with. This means that for
very innovative concepts, engineers are not likely to have accurate mental models
of behavior and will judge a concept inaccurately. Very viable and effective
concepts may be overlooked due to inaccurate mental models. Helms et al. (2009)
document one of the common errors with student teams is that they often miss the
significance of the underlying principles of the biological analogue and then
oversimplify the complex functions of the analogue.
Heuristic
• Supplement designers’ mental models—assist designers by supplementing their
erroneous mental models. Provide physical prototypes, virtual prototypes and
other external models that test and provide feedback to the designers on their
mental models. Provide other means to test and evaluate designers’ mental
models.
9.2.3 Distant Domain Analogies are Initially Bypassed
Another challenge engineers face is that distant domain analogies are ignored
unless the design problem is open (unsolved). Engineers tend to ignore distant
domain analogies unless they have spent time attempting to solve a problem and
are having difficulty, an ‘‘open’’ problem. If the analogous information is more
distantly related to the problem, it is more likely to be used to solve the problem if
it is presented when there is an open design problem to be completed rather than
before work has started on the problem (Tseng et al. 2008).
Heuristic
• Present distant domains analogues later—Present more distant domain ana-
logues after the designer has spent time working on the problem and it is still
unsolved.
• Encourage distant domain analogy use—Encourage designers to use distant
domain analogues by suggesting they create new solutions for every example
presented or by highlighting how distant domain analogues are related to the

problem.
• Provide several far-domain examples—Present more than one far-domain
example.
9.2.4 Single Inference Bias and Principles for Creating

Multiple Solutions
Another bias designers face is that it is often difficult to create many solutions
based on a single analogue (Krawczyk et al. 2005; Holyoak and Thagard 1989;
Gadwal and Linsey 2010a). For most tasks in everyday life, a single, most likely to
work solution is desired, for example, if you walk up to a door, you probably want
to try opening it the same way you did last time. The fundamental purpose of
analogy is to generate plausible and useful inferences. In order to obtain useful
inferences from analogical reasoning, analogical mappings have to be constrained,
otherwise too many inferences are possible (Krawczyk et al. 2005; Holyoak and
Thagard 1989). Very often multiple inferences based on a single analogue do exist.
For example, the analogy between a bird and human flight is a case where the
wrong inference was made for a long time. There have also been a number of wall
climbing devices and adhesives based on the gecko lizard which is able to quickly
scale walls (Gadwal and Linsey 2010a). In general, it is difficult to develop
multiple inferences (solutions) from a single analogue (Krawczyk et al. 2005;
Holyoak and Thagard 1989). Engineers can develop multiple inferences when
instructed to do so, but it is not an easy task (Gadwal and Linsey 2010a). On
average, participants were only able to create three solutions in 30 min, indicating
it is a difficult task and engineers need methods to assist them.
To assists engineers, Gadwal and Linsey created a set of principles from
existing examples of multiple solutions based on a single analogue (Fig. 9.2,
Gadwal and Linsey 2010b). They asked designers to generate ideas for a device
dispensing flour efficiently. The provided a child’s toy device that can hold and
release substances, as shown in Fig. 9.2, as an analogue for the problem. The set of
principles were derived based on the ideas generated by the participants. The
principles are as follows:
• Change the scale of the feature mappings
– Map at the same scale.
– Map at different scales: Substitute physical principles (e.g., magnetic instead
for van der Waals).
• Abstract the problem
– Change the abstraction of the design problem (e.g., ‘‘wall climbing’’ to the
more general case of ‘‘adhesion’’). The WordTree Method can facilitate this
process (see Linsey et al. 2008a, 2012 for more details). Figure 9.3 displays
Fig. 9.2 Above are two solutions for a device to sprinkle flour based on an analogue to a child’s
toy. These solutions illustrate the design principle for multiple solutions of repeating a design
feature (Gadwal and Linsey 2010b)
Fig. 9.3 Examples of

WordTree Change Shape
abstraction of the design
problem of folding laundry Abstraction of
the Problem Compact
Change Functional
Prepare for Basis:
Change Surface Storage Form
Fold Laundry
three different abstraction of the problem of designing a device to fold

laundry. Using the WordTree Method, this can be abstracted to ‘‘change
surface’’ and even more generally to just change. Fold could also be thought
of more abstractly as a problem to ‘‘prepare laundry for storage,’’ ‘‘compact’’
or most abstractly as ‘‘changing shape.’’ The functional basis (Otto and Wood
2001) results in an abstract to the function ‘‘form.’’
• Choose different features
– List properties and design features. Map various features into solutions. For
example, Fig. 9.4 shows the design problem of creating a lightweight travel
Design Problem:
Lightweight travel exercise Yoga mat
device
Inflatable yoga mat
Feature-Use
Analogue-Air Mattress substance available
at usage location Water Dumbbell
Punching bag
filled with sand
Fig. 9.4 Examples of multiple solutions created based on different features of the analogue and
by abstracting the analogue features
exercise device and the analogue of an inflatable mattress (lightweight and

easy to store). In Fig. 9.4, the design features of ‘‘rolls up’’ and ‘‘inflate/
deflate’’ are mapped to solutions for portable exercise equipment.)
• Abstract the features. The WordTree Method can facilitate this. For example,
‘‘inflate/deflate’’ can be abstracted to ‘‘increase in size’’, ‘‘use available sub-
stance at use location to make the device functional’’, ‘‘expand’’, etc.
• Repeat a design feature or take a repeated design feature and use it only once.
Figure 9.2 demonstrates this design principle through the design of device to
sprinkle flour based on an analogy to the child’s toy. The design in the top right
of the figure uses the design feature once to release the flour. The design in the
bottom right repeats the same design feature.
Heuristic
• Multiple Inferences—Assist users in creating multiple inferences. Use multiple
analogies design principles.
9.2.5 Alignable Differences
Data from cognitive psychology indicates that individuals systematically and

unintentionally ignore information that they believe to be important when it is non-
alignable with what they know about other options (Lindemann and Markman 1996).
For example, a bicycle and a motorcycle have many alignable differences (one uses a
motor and the other a human power, motorcycles travel longer distances and
are more expensive) whereas a car and a phone have very few alignable differences.
It is actually hard to say what is different about a car and a phone. When making a
selection and when finding similarities during the analogical reasoning process,
individuals will focus more on alignable differences than non-alignable ones even
though in most design situations both are likely equally important.
Heuristic
• Alignable and Non-alignable—Illustrate both alignable and important non-
alignable differences between the analogue and the problem
9.2.6 Biases and Challenges Still Requiring

Heuristics to Overcome Them
There are additional challenges that currently do not have tools to assist in
overcoming them, but are still important. Confirmation bias is the tendency to
accept information that is consistent with current beliefs while ignoring infor-
mation discredits or is contrary to current beliefs. Within design by analogy and
bioinspired design, (Hallihan et al. 2012) demonstrate that designers suffer con-
firmation bias during concept generation. This likely leads designers not to re-
evaluate the initial analogies they select, nor evaluate their initial analogical
mappings. Helms et al. (2009) observe design teams sticking to their first analogue
which is likely due to a confirmation bias.
Helms et al. (2009) have documented other common errors that student teams
encounter. Professionally likely face many of the same challenges. Many of these
errors have been referenced in the previous sections with heuristics for overcoming
them, but a few of the identified common errors have not been discussed. One error
is using the ‘‘off-the-shelf’’ biological solution, for example when designing a
device to shell peanuts, designers suggest training squirrels to shell the peanut
rather than using the principles a squirrel implements. Designers also often map
the wrong features from the analogue to the inspired solution frequently due to
superficial matches. Another issue is designers also transfer additional features
which are not required for the problem domain.
9.2.7 Approaches to Increasing Success in Design

by Analogy
While much work needs to be done in order to identify approaches for overcoming
the cognitive challenges during analogical reasoning, a few proven approaches
have been identified. Wherever possible, the high-level principle between the
analogue and the problem should be presented. Presenting both the abstract
principle and the analogue significantly increases success rates (Thompson et al.
2000; Namy and Gentner 2002b; Mak and Shu 2004). The presence of the
analogue helps designers in understanding the underlying principle and this

facilitates the analogical transfer of that principle instead of the surface features in
the analogue.
Unfortunately, the abstract principle is not often known prior to the solution
being created. Presenting at least two analogues also increases success rate. To
maximize success, these analogues should be diverse (Clark and Mayer 2008). The
examples should be different from each other to facilitate identification of the
principle and the unimportant surface features should be different. Prior work has
consistently shown that presenting two analogues rather than only one increases
the rate at which the problem will be solved (Namy and Gentner 2002a; Markman
and Gentner 1993a; Keane 1988; Gick and Holyoak 1980; Lopez 2011; Gadwal
and Linsey 2010a). What is unknown is how many examples combined with the
abstract principle is most effective, what the optimal number is in the presence of
extraneous information, or how different the examples should be.
Most of the prior work with the exception of Lopez et al. (Lopez 2011; Gadwal
and Linsey 2010a; Lopez et al. 2012) evaluated the effect in multiple analogues
with very little extraneous information that participants could focus on. In realistic
situations, designers must filter out the extraneous information to identify the
critical features which should be aligned and mapped between the analogues and
the solution. Said study did measure the effect of presenting participants with
multiple analogues both with and without significant extraneous products
(Fig. 9.5). They implemented an experiment with two factors: the number of
analogues (four levels: 1, 2, 3, or 5 products) and the level of extraneous products
(two levels: none or 3 extraneous products per analogue). Participants were pre-
sented with a set of products and then asked to solve a design problem. After
solving the design problem, participants were also asked to identify the principle
from the analogues that lead to the solution. Consistent with prior studies, without
extraneous products being present, two analogues were optimal to obtain success
in solving the design problem (Fig. 9.5). When a combination of both useful and
extraneous analogues was presented, three analogues were optimal. Consistent
with prior studies, multiple analogues that share the same principle are effective
for increasing analogical transfer.
Fig. 9.5 Percentage of

participants finding a solution
based on the analogues and
correctly identifying the
principle
Heuristics
• Abstract and Specific—Present both the abstract principle and specific analogue
examples to maximize analogical transfer.
• Diverse examples—Provide several diverse examples for far transfer.
• Multiple Analogues—Present two to three analogues that share the same
principle.
9.3 Design Fixation
Any tool or method to support and enhance the bioinspired design process must
minimize design fixation. The existing literature demonstrates the susceptibility of
both experts and novices to design fixation (Jansson and Smith 1991; Wiley 1998).
Design fixation causes the blind adherence of designers to presented examples or
their initial ideas (Jansson and Smith 1991). Fixation narrows the solution space
where designers conceive their ideas and decreases creativity. Design fixation is a
common problem for practicing engineers, design faculty, and engineering stu-
dents (Purcell and Gero 1996; Jansson and Smith 1991; Linsey et al. 2010;
Christensen and Schunn 2007).
There are a number of factors that likely influence design fixation including the
sunk cost effect, unusual solutions, abstraction and categories of solutions
(Table 9.1). Current and prior work indicate that sunk cost has great potential to
cause design fixation (Viswanathan and Linsey 2010; Viswanathan and Linsey in
review). Sunk cost basically indicates that once a decision is made or significant
effort is applied, people tend not to change the course of action (Kahneman and
Tversky 1979a; Arkes and Blumer 1985). This indicates that designers should be
more fixated when significant effort has gone into a project. In design, it has been
observed that once prototypes are built, designers make very few design changes
(Christensen and Schunn 2007). Sunk cost’s impact is further described in the next
section.
Two experimental studies attempting to identify the benefits of brainstorming
groups presented participants with uncommon examples (ideas participants rarely
Table 9.1 Factors that likely influence design fixation

Factors Examples
Increases Sunk cost Time-consuming and expensive prototypes
fixation
Decreases Unusual solutions Analogies to nature
fixation Abstraction Functions, design principles such as TRIZ or
design for X
Categories of solutions Energy sources, checklists
Focusing on ineffective Providing the causal reason why a design is
features ineffective
thought of spontaneously) and found that more ideas were generated compared to
those who received common examples, indicating that fixation is probably reduced
(Dugosh and Paulus 2005; Perttula and Sipila 2007). Providing uncommon
examples like biological analogues can reduce design fixation. Commonly iden-
tified biological analogues will likely not have this effect, but most biological
analogues would be uncommon examples.
Limited direct experimental work exists on identifying approaches for over-
coming design fixation, but two studies have directly identified approaches:
Chrysikou and Weisberg (2005), Linsey et al. (2010). First, Chrysikou and We-
isberg (2005) replicated a study by Jansson and Smith (1991) and also presented
participants with warnings on features they should not replicate. Chrysikou and
Weisberg (2005) did not observe participants fixing to the example like the
Jansson and Smith study did. The Jansson and Smith study also told the partici-
pants which features to not duplicate. A likely reason the Chrysikou and Weisberg
study did not see fixation is because participants in this study were forced to focus
more on what they should not do.
Linsey et al. (2010) reduced design fixation through providing participants with
a combination of unusual solutions (distant domain analogies), abstractions of the
design problem (functions of a device), and categories of solutions (categories of
natural energy sources). In the Linsey et al. study, an experimental group is pro-
vided with a fixating example (Fixation Group), whereas another group is provided
with the same example and the defixation materials (Defixation Group). A third
group did not receive an example (Control). Linsey et al. found that all the def-
ixation materials in combination reduced experts’ fixation; however, did not
identify which individual factors reduced design fixation. A follow-up study found
that fixation was only reduced for experts (design faculty) but not for more novice
engineers (senior mechanical undergraduates) (Viswanathan and Linsey 2012,
2013, Fig. 9.6). When a fixating example was present, the quantity of non-
redundant ideas generated by both faculty and novice designers was reduced,
indicating fixation. When the defixation materials were present, the quantity
increases for faculty designers, but not for novices. This indicates that defixation
Fig. 9.6 The defixation

materials only reduced
fixation for the experts (Error
bars ±1 S.E)
materials mitigated fixation for faculty designers, but those materials were not
equally effective on novices. Uncommon examples, abstractions, and categories
likely reduce design fixation.
Numerous design books also advocate the use of abstraction and categories in
design (Otto and Wood 2001; Antonsson and Cagan 2001; Pahl and Beitz 1996).
Suggested abstractions include, but are not limited to, functional, graph grammars,
black boxes, and bond graphs. Useful categories include the functional basis, idea
generators and other checklist methods, physical principles, and classifying
schemes (e.g., working geometry, and working motions). Idea generation methods
such as Mind Maps and Morph Matrices also highlight categories.
Much further, work is required to study these influences on design fixation and
to identify more ways to overcome design fixation. The effect sizes of each of
these factors also need to be measured, so that factors with greatest impact can be
focused on. The types of abstractions that reduce design fixation also need to be
determined. Ideally, bioinspired design tools would be able to help designers in
recognizing their design fixation and then provide stimuli to reduce that fixation.
For identifying design fixation in a design session, Gero (2011) has developed a
metric based on linkography from coding design protocol. Gero’s metric does not
measure fixation relative to a control group; so his approach, unlike the measures
often implemented in fixation studies, could be incorporated into a computational
tool to identify when a designer is fixated and then provide examples to reduce
fixation.
Heuristics
• Uncommon Examples—Common examples cause design fixation and uncom-
mon ones do not. Unusual example (analogues) should be provided whenever
possible.
• Causal Reason for Ineffectiveness—Explain why certain features are undesirable
in the design.
• Categories of Solutions—Categories of solutions can help designers identify
associated principles.
• Encourage Abstraction—Encourage thinking in terms of high-level principles
instead of surface features.
9.3.1 Sunk Cost Effect
Identified by behavioral economics, the sunk cost effect manifests a greater ten-
dency to continue in a selected path, after significant money, time, or effort is
invested in that path, even when an alternate path is more beneficial for the future
endeavors (Arkes and Blumer 1985). Good decisions should be based on the
expected costs of the choices in the future not past sunk costs (Keeney and Raiffa
1993; Holcomb and Evans 1987). However, in actual practice, sunk costs do affect
decisions, due to the sunk cost effect (Kahneman and Tversky 1979b). Some good
examples of this effect are portrayed by Thaler (1980). The resale prices of cars are
generally guided by the current market price, whereas the sellers always decide
based on the original buying price.
The sunk cost effect can fixate designers to their initial ideas, especially when
they spend more time or effect (costs) on those ideas. In engineering design, the
cost can be money, time, or effort that designers spend to solve a problem. Once
significant investment of these resources is made into a particular solution path,
designers tend to fixate on that path. In engineering design, the generation of
highly novel ideas is important and this requires ‘‘out-of-the box’’ thinking. The
adherence to one selected solution path can hinder this target. This can be espe-
cially true when designers build physical models of their ideas during idea gen-
eration. If this building process takes longer, the chances of fixation is also greater.
A controlled study conducted by the Viswanathan and Linsey (Viswanathan
and Linsey 2011; Viswanathan and Linsey in review) has shown evidence sup-
porting the presence of the sunk cost effect in engineering idea generation. In that
study, novice designers were instructed to generate as many ideas as possible for a
small object that could securely bind ten sheets of paper together without dam-
aging them. The participants were randomly assigned to five different conditions.
The first condition was sketching only condition, in which the participants sket-
ched their ideas. In the second condition, metal building, the participants sketched
their ideas and built those with steel wire. The third condition was a plastic
building condition, in which the participants sketched their ideas and molded those
out of plastic. Building ideas with plastic consumes more time as compared to
building with metal, making the associated sunk cost higher for plastic building.
The fourth and fifth conditions were metal constrained sketching and plastic
constrained sketching, respectively. In these conditions, the participants were told
that they would build their ideas in the second half of the experiment and were
instructed to sketch their ideas. These constrained sketching conditions isolated
Fig. 9.7 Example physical models built by the participants in the study
Fig. 9.8 Variation of average novelty across the conditions (Error bars ±1 S.E.)
any possible effects of implicit constraints imposed by the building materials and
processes. Figure 9.7 shows some example solutions generated by the participants.
The presence of design fixation in the ideas generated by participants was
measured using novelty and variety metrics (Shah et al. 2000; Linsey et al. 2005;
Linsey et al. 2011). Design fixation is believed to cause ideas to be less novel and
for participants to search a smaller portion of the design space. Novelty measured
how frequently a particular idea occurred. The ideas were sorted into bins of very
similar ideas and novelty was calculated as one minus the frequency of ideas in
each bin. Variety measures the span of the entire solution space that a participant’s
Fig. 9.9 Variation of average variety across the conditions (Error bars ±1 S.E.)
ideas cover. This is measured as the ratio of the number of bins that the partici-
pant’s ideas occupy to the total number of bins. Figures 9.8, 9.9 show the variation
of these measures across the five conditions. Participants who built their ideas with
metal generated ideas with a lower average novelty and variety as those in the
Sketching Only Condition, indicating design fixation. The data clearly showed a
further reduction in novelty of ideas when they built those ideas with plastic. As
building with plastic consumed more time as compared to building with metal, the
sunk cost effect could be causing this reduction. Comparing the Sketching Only
Condition with the constrained sketching conditions, there was no effect of
implicit constraints observed on novelty and variety.
Heuristic
• Minimize Sunk Cost: Minimize sunk cost (time, effort, money, etc.) to encourage
changes in the course of action and minimize fixation. Tools and methods
should require little time, money, and effort from the designer in order to be
effective.
9.4 Other Considerations
Much work has been done in the area of human computer interaction on creating
effective computer tools. Much more applicable work exists beyond the scope of
this chapter. Within this area, research on learning in multi-media environments
and from design research provides guidelines applicable for the development of
new methods and computer-based tools for bioinspired design. Biological phe-
nomena are generally outside a designer’s knowledge domain; therefore, before an
effective solution can be created based on the biological phenomena, the designer
must first learn about the biology to some degree. Engineering design research has
demonstrated that ambiguous or unfamiliar representations should be annotated
with short phrases to enhance communication (Hisarciklilar and Boujut 2009;
Linsey et al. 2011). Information, such as a graphic and annotation that must be
integrated, should not be separated. Images and labels should be placed close to
each other (Clark and Mayer 2008). Figure 9.10 demonstrates the principles of
annotation and contiguity. The images on the left show two examples that do not
follow these principles and the image on the right shows a clear description of the
flat spring outlined in the image (an ambiguous representation) placed directly
next to the image.
Researchers also must be aware that much depends on the knowledge and
experiences of the user (Clark and Mayer 2008). Certain representations will be
effective for different groups of people based on their prior experiences. A number
of studies in engineering design demonstrate cognitive differences between nov-
ices and experts in design (Ball et al. 2004; Casakin and Goldschmidt 1999;
Fig. 9.10 Illustration of the annotation and contiguity principles
Chi et al. 1981; Moss et al. 2006). Experimental findings often differ based on
expertise level and discipline expertise. Purcell and Gero (1996) found that experts
in mechanical engineering (senior undergraduates) would fixate on presented
examples while experts in industrial design (senior undergraduates) did not fixate.
The impact of intervention to assist designers can also be impacted by expertise.
Viswanathan and Linsey (2012, 2013) found that defixation materials such as
analogies, categories of solutions, and abstractions effectively reduced experts’
fixation but did not assist novices.
Further challenging the situation, users’ perceptions about the effectiveness of
an idea generation method are often inconsistent with the quantitative outcomes.
Users believe that brainstorming, where ideas are shared verbally with a group
then recorded, is more effective than brainwriting where individuals simulta-
neously write down their ideas and then exchange them (Paulus and Yang 2000).
In Paulus and Yang’s study, people generated more ideas during brainwriting than
brainstorming but believe that brainstorming is more effective for idea generation.
The inaccuracy of productivity perceptions has also been found with experts in
design. Linsey et al. (2010) in a study on design fixation presented experts with
poor design examples. The experts inaccurately believed that the example did not
hurt their idea generation process. The experts were somewhat accurate at per-
ceiving that the defixation materials alleviated their fixation.
Heuristics
• Annotation—Include annotations next to ambiguous/unfamiliar representations
to increase clarity.
• Contiguity—Avoid separating information that must be integrated (do not
separate images and labels and place them as close to each other as possible).
• User Dependence—A representation’s effect depends on the prior knowledge of
the user.
• User Perceptions—Measure both user perceptions and quantitative outcomes
when assessing a method’s effectiveness.
9.5 Summary
This chapter presents the known cognitive challenges that designers face when
attempting to implement design by analogy and bioinspired design. Cognitive
research in design, psychology, and multi-media learning has identified these
cognitive challenges. Table 9.2a, b summarizes the known cognitive challenges
that designers face along with providing a summary of the principles and heuristics
that methods and tool developers can implement to assist designers in overcoming
the cognitive challenges. Design by analogy is a powerful but not easy cognitive
process. Designers need assistance in retrieving appropriate analogues based on
deep and functional connections. Due often to inaccurate mental models of how
the analogue works or the non-alignable differences, designers may select the
wrong features from the analogue to implement to find a solution. In addition,
designers tend to initially unintentionally ignore distant domain analogues and
then are only able to find a few inferences (solutions) based on each analogue even
though numerous solutions may actually exist. Design fixation, initial examples or
solutions limiting creativity, further inhibits bioinspired design.
While significant challenges exist, there is much that developers of new design
methods and tools can do to assist designers. Computer-based designer tools can
retrieve analogues based on deep connections between the problem and analogue
such as functional or other relational similarity. Designers need to be encouraged
to identify multiple solutions (inferences) based on each analogue presented.
Presenting diverse examples, multiple analogues that share the same principle, a
diverse set of analogues and the abstract principles will also facilitate the design by
analogy process. To minimize design fixation, designers need uncommon solu-
tions, categories of solutions and abstractions. New tools also need to not require
significant sunk costs of time, money, or effort from designers in order to minimize
design fixation.
Table 9.2 Summary of cognitive challenges that designers face during design by analogy and heuristics to guide the design of tools to assist designers
Cognitive biases and challenges Principles and heuristics
a
Challenges in analogy Retrieving based on unimportant (surface) features Deep connections—encourage deep and functional
connections between the analogue and the
problem
Selecting the wrong features to map Matching features—highlight appropriate features
that should be mapped from the analogue to the
problem
Inaccurate mental models—individuals’ mental Supplement designers’ inaccurate mental models
models of the world are cognitively efficient but (e.g., external models, feedback, and testing)
often inaccurate
Distant domains are initially bypassed Present distant domains later. Encourage distant
domain use. Provide several far-domain
examples
Single inference—designers only infer a single Multiple inferences—assist users in creating
analogue for a problem multiple inferences. Use multiple solution design
principles
Alignable differences—individuals focus more on Alignable and non-alignable—illustrate both
alignable differences than non-alignable ones. alignable and important non-alignable
Individuals systematically ignore important differences between the analogue and the
information when it is non-alignable problem
Analogical inference is cognitively difficult Diverse examples—provide several diverse
examples for far transfer
9 Overcoming Cognitive Challenges in Bioinspired Design and Analogy
Multiple analogues—present two to three analogues

that share the same principle whenever possible
Abstract and specific—present both the abstract
principle and specific analogue examples to
maximize analogical transfer
(continued)
239
Table 9.2 (continued)
240
Cognitive biases and challenges Principles and heuristics

Confirmation bias
‘‘Off-the-shelf’’ solutions
Transferring extraneous features
b
Design fixation Designers fixate Uncommon examples—common examples cause design fixation
and uncommon ones do not
Causal reason for ineffectiveness—explain why certain features
are undesirable in the design
Categories of solutions—categories of solutions can help designers
identify associated principles
Encourage abstraction—encourage thinking in terms of high-level
principles instead of surface features
Sunk cost effect—designers fixate when Minimize sunk cost—minimize sunk cost (time, effort, money, etc.)
associated sunk cost is high to encourage changes in the course of action and minimize fixation.
Tools and methods should require little time, money, and effort from
the designer in order to be effective
Other considerations Other considerations for the design Annotation—include annotations next to ambiguous/unfamiliar
of tools and methods representations to increase clarity
Contiguity—avoid separating information that must be integrated
(do not separate images and labels and place as close to each
other as possible)
User dependence—a representation effects depend on the prior
knowledge of the user
Inaccurate user perceptions—user’s perceptions User perceptions—measure both user perceptions and
of the impact of methods are often inaccurate quantitative outcomes when assessing a method’s effectiveness
when compared with the quantitative outcomes
J. S. Linsey and V. K. Viswanathan
The chapter summarizes only the known cognitive biases and many more likely
exist. Most of the challenges and heuristics in the chapter are based on a limited
number of studies and much more research needs to be done. Almost nothing is
known about the relative effect sizes or the impact of multiple principles in
combination. There is little data on the influence of expertise in regards to
developing new tools or with the presented cognitive challenges and heuristics.
References
Antonsson EK, Cagan J (eds) (2001) Formal engineering design synthesis. Cambridge University
Press, Cambridge
Arkes HR, Blumer C (1985) The psychology of sunk cost. Organ Behav Hum Decis Process
35(1):124–140
Ball LJ, Ormerod TC, Morley NJ (2004) Spontaneous analogising in engineering design: a
comparative analysis of experts and novices. Des Stud 25(5):495–508
Basalla G (1988) The evolution of technology. Cambridge University Press, Cambridge
Casakin H, Goldschmidt G (1999) Expertise and the use of visual analogy: implications for
design education. Des Stud 20(2):153–175
Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju BS (2005a) A behavioural model for
representing biological and artificial systems for inspiring novel designs. Paper presented at
the international conference on engineering design, Melbourne
Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju BS (2005b) A functional representation for
aiding biomimetic and artificial inspiration of new ideas. AIEDAM 19(2):113–132
Cheong H, Shu LH (2009) Effective analogical transfer using biological descriptions retrieved
with functional and biologically meaningful keywords. Paper presented at the ASME
international design engineering technical conferences and computers and information in
engineering conference, San Diego
for functional terms of the functional basis. J Mech Des 133:021007. doi:http://dx.doi.org.lib-
ezproxy.tamu.edu:2048/10.1115/1.4003249
Chi MTH, Feltovich PJ, Glaser R (1981) Categorization and representation of physics problems
by experts and novices. Cogn Sci 5:121–152
domain information retrieval. AIEDAM 21(1):45–59
Christensen BT, Schunn C (2007) The relationship of analogical distance to analogical function
and pre-inventive structures: the case of engineering design. Mem Cogn 35(1):29–38
Chrysikou EG, Weisberg RW (2005) Following the wrong footsteps: fixation effects of pictorial
examples in a design problem-solving task. J Exp Psychol Learn Mem Cogn 31(5):1134–1148
Clark R, Mayer RE (2008) E-learning and the science of instruction: proven guidelines for
consumers and designers of multimedia learning, 2nd edn. Wiley, San Francisco
Dugosh KL, Paulus PB (2005) Cognitive and social comparison processes in brainstorming.
J Exp Soc Psychol 41:313–320
Falkenhainer BF, Forbus KD, Gentner D (1989) The structure mapping engine: algorithm and
examples. Artif Intell 41(1):1–63
Forbus KD (1984) Qualitative process theory. Artif Intell 24:85–168
Gadwal A, Linsey J (2010a) Exploring multiple solutions and multiple analogies to support
innovative design. Paper presented at the international conference on design, computing and
cognition, Stuttgart
Gadwal A, Linsey J (2010b) Inspiring multiple solutions from a single analogy. Paper presented
at the 13th AIAA/ISSMO multidisciplinary analysis optimization (MAO) conference, Fort
Worth
Gentner D (1983) Structure mapping—a theoretical framework. Cogn Sci 7:155–177
Gentner D, Landers R (1985) Analogical remindings: a good match is hard to find. Paper
presented at the international conference on systems, man, and cybernetics, Tucson
Gentner D, Markman AB (1997) Structure mapping in analogy and similarity. Am Psychol
52:45–56
Gentner D, Stevens AL (1983) Mental models Lawrence Erlbaum associates. Hillsdale, New
Jersey
Gero JS (2011) Fixation and commitment while designing and its measurement. J Creative Behav
45(2):108–115
Gick ML, Holyoak KJ (1980) Analogical problem solving. Cogn Psychol 12:306–355
Goel AK, Bhatta SR (2004) Use of design patterns in analogy-based design. Adv Eng Inform
18(2):85–94
Hacco E, Shu LH (2002) Biomimetic concept generation applied to design for remanufacture.
Paper presented at the ASME design engineering technical conferences and computer and
information in engineering conference, Montreal
Hallihan GM, Cheong H, Shu LH (2012) Confirmation and cognitive bias in design cognition. A
paper presented at ASME design engineering technical conferences, Chicago
Helms M, Vattam SS, Goel AK (2009) Biologically inspired design: process and products. Des
Stud 30(5):606–622
Hisarciklilar O, Boujut JF (2009) An annotation model to reduce ambiguity in design
communication. Res Eng Des 20:171–184
Holcomb JH, Evans DA (1987) The effect of sunk costs on uncertain decisions in experimental
markets. J Behav Econ 16(3):59–66
Holyoak KJ, Koh K (1987) Surface and structural similarity in analogical transfer. Mem Cogn
15(4):332–340
Holyoak KJ, Thagard P (1989) Analogical mapping by constraint satisfaction. Cogn Sci
13(3):295–355
Hummel JE, Holyoak KJ (1997) Distributed representations of structure: a theory of analogical
access and mapping. Psychol Rev 104(3):427–466
Hutchins E (1995) Cognition in the wild. MIT Press, Cambridge
Jansson D, Smith S (1991) Design fixation. Des Stud 12(1):3–11
Kahneman D, Tversky A (1979a) Prospect theory: an analysis of decision under risk.
Econometrica 47(2):263–291
Kahneman D, Tversky A (1979b) Prospect theory: an analysis of decision under risk.
Econometrica J Econometric Soc 47:263–291
Keane MT (1988) Analogical problem solving. Wiley, New York
Keeney RL, Raiffa H (1993) Decisions with multiple objectives: preferences and value tradeoffs.
Cambridge University Press, Cambridge
Kempton W (1986) Two theories of home heat control. Cognitive Sci 10(1):75–90
Krawczyk D, Holyoak K, Hummel J (2005) The one-to-one constraint in analogical mapping and
inference. Cogn Sci 29(5):797–806
Kuipers B (1994) Qualitative reasoning: modeling and simulation with incomplete knowledge.
The MIT Press, Cambridge
Leclercq P, Heylighen A (2002) 5, 8 Analogies per hour. In: Gero JS (ed) Artificial intelligence in
design ‘02, pp 285–303
Lindemann PG, Markman AB (1996) Alignability and attribute importance in choice. Paper
presented at the 18th annual meeting of the cognitive science society, Mahwah
Linsey J (2007) Design-by-analogy and representation in innovative engineering concept
generation. The University of Texas, Austin
Linsey J, Green MG, Murphy JT, Wood KL, Markman AB (2005) Collaborating to success: an
experimental study of group idea generation techniques. Paper presented at the ASME design
theory and methodology conference, Long Beach
Linsey J, Murphy JT, Wood KL, Markman AB, Kurtoglu T (2006) Representing analogies:
increasing the probability of success. Paper presented at the ASME design theory and
methodology conference, Philadelphia
Linsey J, Laux J, Clauss EF, Wood K, Markman A (2007) Increasing innovation: a trilogy of
experiments towards a design-by-analogy method paper presented at the asme design theory
and methodology conference, Las Vegas
Linsey J, Wood K, Markman A (2008a) Increasing innovation: presentation and evaluation of the
wordtree design-by-analogy method. Paper presented at the ASME IDETC design theory and
methodology conference, New York
Linsey J, Wood K, Markman A (2008b) Modality and representation in analogy. AIEDAM
22(2):85–100
Linsey J, Tseng I, Fu K, Cagan J, Wood K, Schunn C (2010) A study of design fixation, its
mitigation and perception in engineering design faculty. ASME J Mech Des 132 (4):041003-
041001-041012
Linsey J, Clauss EF, Kurtoglu T, Murphy JT, Wood KL, Markman AB (2011) An experimental
study of group idea generation techniques: understanding the roles of idea representation and
viewing methods. ASME J Mech Des 133 (3):031008-031001-031008-031015
Linsey J, Markman A, Wood K (2012) Design by analogy: a study of the wordtree method for
problem re-representation. ASME J Mech Des (in press
Lopez R (2011) Characterizing the effects of noise and domain distance in analogous design. MS
Thesis, Texas A&M University, College Station, TX
Lopez R, Smith S, Linsey J (2012) The effects of multiple analogies in the presence of extraneous
information. Working paper
53(1):1170120
Des 19(1):21–28. doi:10.1007/s00163-007-0041-y
Markman A (1999) Knowledge representation. Lawrence Erlbaum Associates, Mahwah
Markman AB, Gentner D (1993a) Structural alignment during similarity comparisions. Cogn
Psychol 25:431–467
Markman AB, Gentner D (1993b) Structural alignment during similarity comparisons. Cogn
Psychol 25(4):431–467
McAdams D, Wood K (2002) A quantitative similarity metric for design by analogy. ASME J
Mech Des 124(2):173–182
McAfee EA, Proffitt DR (1991) Understanding the surface orientation of liquids. Cogn Psychol
23:483–514
Moss J, Kotovsky K, Cagan J (2006) The role of functionality in the mental representations of
engineering students: some differences in the early stages of expertise. Cogn Sci 30(1):65–93
Nagel JKS, Stone RB, McAdams DA (2010) An engineering-to-biology thesaurus for
engineering design. Proceedings of 2010 ASME IDETC/CIE, Montreal
Namy L, Gentner D (2002a) Making a silk purse out of two sow’s ears: young children’s use of
comparison in category learning. Int J Exp Psychol Gen 131(1):5–15
Namy LL, Gentner D (2002b) Making a silk purse out of two sow’s ears: young children’s use of
comparison in category learning. J Exp Psychol Gen 131(1):5–15
Oriakhi E, Linsey J, Peng X (2011) Design-by-analogy using the wordtree method and an
automated wordtree generating tool. Paper presented at the international conference on
engineering design, Copenhagen
Otto K, Wood K (2001) Product design: techniques in reverse engineering and new product
development. Prentice Hall, Upper Saddle River
Pahl G, Beitz W (1996) Engineering design—a systematic approach, 2nd edn. Springer, New
York
Paulus PB, Yang HC (2000) Idea generation in groups: a basis for creativity in organizations.
Organ Behav Hum Decis Process 82(1):76–87
Perttula M, Sipila P (2007) The idea exposure paradigm in design idea generation. J Eng Des
18(1):93–102
Purcell AT, Gero JS (1996) Design and other types of fixation. Des Stud 17(4):363–383
Shah JJ, Kulkarni SV, Vargas-Hernandez N (2000) Evaluation of idea generation methods for
conceptual design: effectiveness metrics and design of experiments. ASME Trans J Mech Des
122(4):377–384
Thaler R (1980) Toward a positive theory of consumer choice. J Econ Behav Organ 1(1):39–60
Thompson L, Gentner D, Loewenstein J (2000) Avoiding missed opportunities in managerial life:
analogical training more powerful than individual case training. Organ Behav Hum Decis
Process 82(1):60–75
stimulation of idea generation in design. Des Stud 29(3):203–221
Vattam S, Wiltgen B, Helms M, Goel A, Yen J (2010) Dane: fostering creativity in and through
biologically inspired design. In: proceedings first international conference on design
creativity, pp 115–122
Viswanathan VK, Linsey JS (2010) Physical models in the idea generation process: hindrance or
help? Paper presented at the ASME international design engineering technical conferences,
Montreal
Viswanathan V, Linsey J (2011) Design fixation in physical modeling: an investigation on the
role of sunk cost. Paper presented at the ASME international design engineering technical
conferences, Washington D.C
Viswanathan VK, Linsey JS (2012) A study on the role of expertise in design fixation and its
mitigation. Paper presented at the ASME international design engineering technical
conferences, Chicago
Viswanathan V, Linsey J (2013) Design fixation and its mitigation: a study on the role of
expertise. ASME J Mech Des 135(5):051008-1–051008-15
Viswanathan V, Linsey JS The role of sunk cost in engineering idea generation: an experimental
investigation. ASME J Mech Des (in review)
Wiley J (1998) Expertise as mental set: the effects of domain knowledge in creative problem
solving. Mem Cogn 26(4):716–730
Chapter 10
An Engineering Approach to Utilizing
Bio-Inspiration in Robotics Applications
Saytandra K. Gupta, Wojciech Bejgerowski, John Gerdes,

James Hopkins, Lengfeng Lee, Madusudanan Sathia Narayanan,
Frank Mendel and Venkat Krovi
Abstract As an interdisciplinary field lying at the intersection of biology and

robotics, bio-inspired robotics has seen considerable interest from a research,
education, and ultimately application perspective. The scope of this interest has
ranged from creating and operating walking, crawling, and flying robots based on
biological counterparts; to evaluating biological algorithms for potential engi-
neering applications; to deconstructing the functioning of living organisms from
the macro- to the microlevels; and ultimately to constructing the next generation of
bio-inspired robots from the bottom up. Significant synergies are forthcoming from
such an approach, but numerous limitations still exist. We begin with the
S. K. Gupta (&) W. Bejgerowski J. Gerdes J. Hopkins

Department of Mechanical Engineering and Institute for Systems Research,
University of Maryland, College Park, USA
e-mail: skgupta@umd.edu
W. Bejgerowski
e-mail: wojbej@umd.edu
J. Gerdes
e-mail: jgerdes@umd.edu
J. Hopkins
e-mail: james.k.hopkins@navy.mil
L. Lee M. S. Narayanan V. Krovi
Department of Mechanical and Aerospace Engineering,
University at Buffalo (SUNY), Buffalo, NY 14260, USA
e-mail: llee3@buffalo.edu
M. S. Narayanan
e-mail: ms329@buffalo.edu
V. Krovi
e-mail: vkrovi@buffalo.edu
F. Mendel
Pathology and Anatomical Sciences, University at Buffalo (SUNY),
Buffalo, NY 14260, USA
e-mail: fcmendel@buffalo.edu

246 S. K. Gupta et al.
discussion of general principles behind taking inspiration from a biological source

and converting it into implementable engineering concepts using case studies of
bird-inspired robots, snake-inspired robots, and mastication robots. These case
studies describe how useful features of the biological creatures were selected and
simplified so that they can be implemented using the existing technologies.
Keywords Robotics Bio-mimetic systems Kinematics Robotic design

Parallel manipulators
10.1 Introduction
The problem of biology is not to stand aghast at the complexity but to conquer it—Sydney
Brenner.
Robot designers are often presented with challenging scenarios, where meeting
design requirements through traditional engineering approaches becomes difficult.
In such cases, robot designers may be interested in taking inspiration from biology;
the biological world is full of extraordinary systems that routinely solve extremely
complex manipulation and locomotion problems very effectively and efficiently.
For example, the musculoskeletal structures of animals have evolved over mil-
lennia to enable energy-efficient and effective mobility in a variety of terrains.
Locomotion research has long taken advantage of inspiration from these biological
analogs: using natural shapes for improved aero- and hydrodynamics, passive
dynamic systems for energy efficiency, neuromuscular synergies for simplifying
coordination- and control-challenges.
However, biological systems have evolved to exploit structural and organiza-
tional principles, spanning multiple physical scales and operational modes, in
order to realize functional performance gains in efficiency, passivity, decentral-
ization, and adaptability. Biological structures are well known for using flexible/
deformable structures, capable of adapting, bending, and twisting in response to
changing environmental conditions, instead of the typical rigid/non-compliant
engineering counterparts. Their underlying multi-scale and nonlinear nature makes
it a challenge to analyze their ultimate functional behavior using either a bottom-
up substructuring framework or a top-down subsystem-based approach.
Therefore, faithfully, mimicking biology alone does not appear to be a realistic
option for systematic engineering design of robots because of many existing lim-
itations including that (1) engineered materials are unable to match properties of
multi-functional living materials; (2) engineered subsystems cannot realize highly
multi-modal characteristics, for example, compliant joints present in animals;
(3) analyzing and simulating such multi-physics-based designs is a very difficult
task; (4) finally, manufacturing complex structures found in nature remains a great
challenge even today. Overcoming these challenges requires an interdisciplinary
research approach within a bio-inspired paradigm. The research approach must
10 An Engineering Approach to Utilizing Bio-Inspiration 247
apply novel engineering methodologies for controlled in vivo experimentation with

biological systems while simultaneously evaluating biological hypotheses and
observations to inform design and control of artificial engineered systems (Cardelli
2005).
In recent years, every scientific arena has benefited from the ubiquitous
availability of computational power and advanced computational tools. Computer
simulation can now be used to calculate the kinematic, dynamic, and finite element
analysis (FEA)-based responses of a prototype and visualize the results in a 3D
interactive virtual environment. By permitting designers to realistically, accu-
rately, and quantitatively prototype and test multiple intermediate models within a
virtual environment, virtual prototyping (VP), also known as simulation-based
design (SBD), has rapidly gained popularity and become a crucial part of most
engineering design processes (Bhatt et al. 2003). The usefulness of such a VP
exercise is limited only by the fidelity of the model and the accuracy of the results.
Oftentimes, there are many effects such as friction, contact that are very sim-
plistically modeled (for computational efficiency or for the lack of more accurate
models) and can only be accurately determined by physical testing. For such
situations, hardware-in-the-loop (HIL) frameworks permit a quick replacement of
the virtual model by the actual physical prototype to permit experimental testing.
The area of bio-inspired robotics has benefited significantly from the use of both
VP and HIL technologies.
While engineering-related fields have witnessed the greatest benefits, these
advances have percolated down far slower into other arenas. In particular, the lack
of significant and useful computational tools in traditional biological sciences such
as anatomy hinders the ability of scientists to effectively and rapidly test various
hypotheses in a rigorous and quantitative manner (Narayanan et al. 2008). Hence,
in this chapter, we adopt a pragmatic approach to using bio-inspiration for design
and control of robotic systems in a staged manner as
• Analysis of functional requirements and evaluation of feasibility of seeking
biological inspiration.
• Identification of the biological feature that offers an opportunity for significant
performance improvement.
• Engineering prototypical bio-inspired design incorporating parametric features
and exploiting rapid manufacturing.
• Analysis, refinement, and optimization of the parametric design which may lead
to significant natural feature exaggeration due to engineering constraints.
The above-described approach aims to identify a promising feature in nature
and then exploit the feature based on the engineering constraints and available
technology. The resulting solution may not at all resemble the original source of
inspiration, but this approach leads to vastly improved design of robots as noted in
the three illustrative case studies.
The first case study describes how compliant flapping mechanisms observed in
birds can be used to realize a highly maneuverable and quiet miniature air vehicle.
The second case study describes how rectilinear gaits used by snakes can be used
to realize robots with small cross-section, achieve significant forward velocity, and
operate on a wide variety of terrains. The final case study investigates vertebrate
mastication performance by using parallel manipulator architectures. In each of the
case studies, we will examine different aspects of (1) obtaining functional
requirements; (2) taking inspiration from nature; (3) manufacturing of the bio-
inspired design; (4) evaluating the performance for optimization and refinement;
and (5) lessons learned.
10.2 Bird-Inspired Miniature Air Vehicles
10.2.1 Functional Requirements
Miniature air vehicles (MAVs) are beginning to gain popularity in a variety of

applications. The combination of small size, portability, and ease of use provides
users with new opportunities to utilize MAVs in a variety of tasks. Traditionally,
MAVs use fixed wings or rotary wings to achieve flight. Each traditional style
offers unique advantages and disadvantages. For example, rotary-wing MAVs are
ideal for hovering and complex maneuvers, while fixed-wing MAVs provide
higher efficiency and speeds. Flapping-wing MAVs offer a compromise between
rotary- and fixed-wing designs while reducing the noise associated with a high-
speed propeller or rotor.
Our goal was to create a new MAV platform that was sufficiently maneuverable
to enable flight inside a building. We also wanted this platform to be quiet during
operation compared to rotary-wing platforms. The target flight weight for the
platform was under 100 g. The primary objective was to maximize payload
capacity. The secondary objective was to minimize the manufacturing cost.
The functional requirement of the drive mechanism was to generate flapping
motion. The functional requirement for the wings was to generate lift and thrust
forces using a flapping motion.
10.2.2 Taking Inspiration from Nature
Flapping motion used by birds tends to be quiet in operation and highly maneu-
verable, so we derived our biological inspiration from avian flight (Gerdes et al.
2012). For this case study, a variety of flying animals were observed that were in a
similar size range to establish common traits which could improve performance.
A well-known property of aerodynamics is the increasing importance of unsteady
aerodynamics as Reynolds number is reduced, which generally corresponds to
smaller flying animals (Dickinson and Gotz 1993). Due to the high frequency of
wing beats in smaller flying animals, kinetic energy recovery is required
to maintain efficient operation. To harness these effects, these animals employ

resonant bodily structures that store and release inertial energy with elastic flight
muscles, connecting tissues, and thorax structures. As the wings are displaced from
the equilibrium position, a body structure such as a thorax or a system of joints and
flight muscles is elastically deformed, storing potential energy. This technique
reduces the energy consumed by the flight muscles by partially eliminating inertial
loads (Madangopal et al. 2005; Madangopal et al. 2006).
Due to limitations of the existing actuator and battery technologies, we chose to
taking our inspiration from avian flight instead of inspect flight. Birds that pro-
vided inspiration included pigeons and doves. The benefits of elastic flapping
structures in nature clearly provide useful benefits for efficient operation. Flying
animals have also been shown to exhibit elastic storage of wing inertia to reduce
power requirements for flight (Ellington 1985). Hence, we decided to utilize
elastically deforming flapping wings in our design and use compliant drive
mechanism. Due to space restrictions, this case study will only focus on compliant
drive mechanism.
10.2.3 Description of Bio-Inspired Design
Compliant mechanisms, defined as flexible structures that elastically deform to

produce a desired force or displacement, have recently been integrated into many
mechanical systems to provide improved functionality, simpler manufacturing,
and greater reliability (Bejgerowski et al. 2009; Gouker et al. 2006; Howell 2001;
Mueller et al. 2009). Compliant mechanisms generally consist of rigid links with
compliant joints added in strategic locations in order to control the force or dis-
placement. Advantages over traditional designs with rigid-body-articulated joints
include (1) reduction in wear between joint members, (2) reduction in backlash,
and (3) potential energy storage in deflected members (Howell 2001). Using
compliant mechanisms also reduces the noise generated by mating components in
operation. Additionally, localized compliance in the mechanism can be used to
eliminate joints from the assembly, resulting in reduced part count and improved
system precision.
Our MAV design is powered by an electric motor; therefore, it uses a crank–
rocker mechanism to achieve flapping action. The mechanism uses localized
compliance, so it consists of rigid links connected by compliant joints. The
schematic diagram of the design is shown in Fig. 10.1.
An outer frame provides crash protection and houses the gearbox which reduces
the flapping speed and torque requirement on the motor to acceptable levels. The
rigid-body revolute joints used in the crank are steel pins due to the very large
oscillatory loads experienced during flapping. A prismatic joint maintains left–
right symmetry in the mechanism; otherwise, large torque from the motor is able to
cause a bulk deformation mode that results in asymmetric flapping and reduced
force output. The layers of the gearbox structure are held together with lightweight
Wing supports
Rigid links
Compliant joints Rocker
Rigid body
Crank
Revolute joints
Fig. 10.1 Schematic diagram of compliant mechanism used for flapping-wing action
Undercut for constraining the

Prismatic joint ensur-
prismatic joint to single DOF
ing symme trical
wing action
Multi-material
compliant structure
Frame for constraining

Bi-planar
rocker motion to a single plane
gearbox structure
Fig. 10.2 MAV drive mechanism design
carbon fiber rods, press fit into the holes distributed around the mechanism. The
design of the mechanism is shown in Fig. 10.2.
After identifying the basic shape of the mechanism, the next step was to
determine detailed dimensions of the mechanism. Considering the functional
requirements and constraints on the overall size of the MAV, it was important to
first identify the constrained and free dimensions of the mechanism design,
illustrated in Fig. 10.3. The constrained dimensions were identified from the
functional requirements of the MAV. The design of the mechanism required the
rocker operational envelope to be placed between the wing arm supports. For the
required flapping range of 65, the minimum separation between the supporting
members was constrained to 19 mm. Inherent stability was desired to simplify
flight control, so the relative angle on the wing arms was designed to be 15,
resulting in a stroke plane extending from 52.5 to -12.5 with an average dihedral
of 20. The length of the crank and the rocker was determined to be 4.1 and
Width =19 mm
Prismatic Joint
Length =66 mm Range =12 mm
Minimum Hinge
Thickness = 0.8 mm
Rocker Workspace Envelope
Fig. 10.3 Constrained dimensions of the design
45.7 mm, respectively. The range of flapping motion also determined the range for
the prismatic joint to be at least 12 mm to account for elastic deformations of the
structure in operation due to loading. The gear axis separation and the range of
motion for the prismatic joint determined the minimum length of the mechanism to
be 66 mm.
The next step was to optimize the design of the compliant hinges in the
mechanism. In miniature multi-material compliant mechanisms, compliant joints
are required to transfer relatively large loads while allowing for the required
degrees of freedom (DOF). The load transferred by the miniature compliant hinge
not only determines the sizing of the hinge cross-section, but also influences the
required level of bonding between the link and joint materials selected to create
rigid and flexible portions of the structure. Since many polymers do not chemically
bond during the molding stage (Gouker et al. 2006; Rotheiser 2004), restricting the
material choices to just chemically compatible pairs significantly reduces the
design space. Therefore, to expand the design space, it was necessary to consider
physical (i.e., mechanical) bonding through interlocking features to ensure a robust
interconnection between the materials.
A parameterized hinge model was used for a simulation-based design optimi-
zation using design variables and constraints that describe the geometry com-
pletely. To evaluate the hinge design, the parameterized geometry was represented
as a 3D solid model. The model was assigned material properties corresponding to
the polymer used to mold the hinge prototype, high-impact polypropylene (HIPP).
The simulation of wing forces’ propagation into the mechanism parts was per-
formed using a dynamics model, modeled using MSC Adams View R3 software.
Each iteration required incremental design changes and remeshing. The final
dimensions of the compliant hinges were chosen to minimize weight.
10.2.4 Manufacturing of Bio-Inspired Design
As described in the description of the bio-inspired design, multi-material injection

molding (MMM) was chosen for manufacturing and assembly. The choice of a
specific MMM method to realize compliant structures with miniature hinges is a
trade-off between the mold complexity and the cycle time. Several morphing
cavity methods have been developed (Priyadarshi et al. 2007; Bejgerowski et al.
2011), offering a high level of automation. The morphing cavity method needs to
satisfy the assembly and disassembly constraints imposed on the mold pieces; the
complexity of the part shape drastically affects the complexity of the molds and
hence significantly increases the tooling cost and lead times. To reduce the mold
complexity, while maintaining the generality of the manufacturing approach, the
cavity transfer method (Gouker et al. 2006) was employed in this design.
The manufacturing process sequence developed consists of two successive
injections of material. The compliant HIPP hinges were first molded and inserted
into the cavity for the second-stage mold. For the second stage, polyamide 6,6
(85 % volume) with short glass fibers (15 % volume) was used to provide the rigid
links in the mechanism. As the second-stage mold encapsulated the first stage,
assembly was automatically completed, leaving demolding and final assembly
with the rest of the MAV as the only remaining steps.
Figure 10.4 shows a photograph of the multi-material compliant MAV drive
frame molded using the cavity transfer method described. Careful visual inspec-
tion of the demolded part revealed no molding defects such as deep weld lines,
excessive warping, or displaced compliant joints. The differential shrinkage of the
part could not be completely eliminated. Nevertheless, since it did not affect the
functionality of the mechanism, we concluded that the shrinkage in the structure
was functionally acceptable.
After successfully accomplishing the molding of the multi-material compliant
drive frame, the MAV was assembled using the molded and machined parts along
with off-the-shelf components, including gears, rods, and motor. The body of the
MAV consisted of the drive mechanism with wings attached on the front and the
tail of servomotor at the back, both connected with two carbon fiber rods and a
foam body. The foam body housed the motor speed controller and remote control
Fig. 10.4 Multi-material

molding of MAV flapping Sprue and
mechanism Runner System
Multi-
Material
Drive Frame
Fig. 10.5 MAV final

assembly
receiver, as well as provided a rear anchoring point for the wings. The assembly of
the MAV is shown in Fig. 10.5.
10.2.5 Performance Results
Upon completion of final assembly, the MAV was evaluated in terms of flight
performance and functioning of mechanism. During the FEA and dynamic simu-
lations, it was determined that the most likely failure points were the left and right
interfaces between the rocker and the wing supports, as these joints supported the
majority of the aerodynamic and inertial loads. High-speed photographic analysis
showed minimal out-of-plane motion of these joints, and extended testing revealed
good fatigue life. Additionally, flight testing in gusty conditions, as well as some
crash landings, confirmed the robustness of these key mechanism points. Test
flights indoors and outdoors revealed good controllability, passive stability, and
excellent endurance. Previous versions of the MAV that used all rigid mechanism
Table 10.1 MAV test flight Parameter Unit Value

results
Overall max. weight g 72.5
Payload capacity g 31.0
Flapping frequency Hz 5.2
Wing area cm2 920.2
Wing span cm 63.5
Flight durationa Min 15
Flight velocity m/s 3.0
a
With 21.1 g 3 9 250 m Ah-cell li-po battery pack
links were only capable of about half the endurance exhibited by this model, thus
confirming the value of compliant mechanisms in boosting the efficiency of oper-
ation (Bejgerowski et al. 2011). Results are summarized in Table 10.1.
10.2.6 Lessons Learned
This case study illustrated that the use of compliance in mechanisms and wings
can be very beneficial from the performance point of view. The idea of using
compliance came by observing birds. But ultimately, the use of compliance in
MAVS was incorporated very differently from the original source of biological
inspiration. In the drive mechanism, it was implemented using a combination of
hard and soft polymer materials that led to compliant joints. In case of wings,
compliance was exaggerated significantly to produce large deformations during
the flapping cycle. This ultimately led to significantly improved performance in
terms of payload-carrying capacity.
The design of the drive mechanism required use of the in-mold assembly
process, an emerging manufacturing process. This process enabled reduction in the
number of assembly operations and use of lightweight polymer composite mate-
rials. This process also enabled realization of parts with small features. This
combination ultimately led to a lightweight drive mechanism design with
improved transmission efficiency. This process is also scalable in nature and can
be used during high-volume production runs to realize the design at a low cost.
10.3 Snake-Inspired Robot
Search and rescue applications often require use of robots that need to go through
very small spaces. Traditional robot platforms are not useful in such applications.
Our goal was to create a new robotic platform that has a cross-section of less than
5,000 mm2. We also wanted this robot to be able to reach a peak speed of at least
2 km/h. This robot had to be able to work on rugged terrains with significant
variations in coefficient of friction.
The main functional requirement for the robot was to be able to locomote over a
wide variety of terrains. It required the robot to be able to move forward, backward,
and turn. The robot also needed to be able to traverse through small passages. These
requirements eliminated wheeled and legged locomotion.
Because of their long, slender, limbless bodies, snakes possess the ability to tra-
verse small enclosed terrains, such as small holes, tunnels, and gaps, which would
prohibit most legged animals. Another natural advantage for snakes is an ability
known as terrainability, which is the ability of an animal to traverse rough or
difficult terrain. This ability allows a snake to crawl over rugged, non-level terrain
as effortlessly as it can traverse open, smooth terrains, and environments, allowing
much more maneuverability compared to conventional concepts. Because of these
two considerations, we derived our biological inspiration from snakes in designing
our new robot platform.
In snakes, rectilinear locomotion is described as the whole snake moving for-
ward along a straight line, sliding against the terrain as illustrated in Fig. 10.6
(Lissmann 1950). Among the various snake-inspired robot gaits, rectilinear-
gait-based motion has demonstrated favorable results through many useful fea-
tures. However, the majority of snake-inspired robot platforms which utilize
rectilinear motion are relatively slow in speed (Hopkins et al. 2009). Higher speeds
in snake-inspired robot platforms are possible, if rectilinear-gait-based motion was
applied in a straight line, as observed in natural snakes, and executed at high rates.
Hence, we decided to choose rectilinear gait for inspiration.
In order to achieve the functional requirements described above and take advan-
tage of the benefits of rectilinear motion observed in nature, the new robot plat-
form must be able to perform high-speed linear expansion/contraction and pivoting
Fig. 10.6 Rectilinear motion

(Lissmann 1950)
Fig. 10.7 Parallel

mechanism concept
motions between segments. This is achieved by developing a mechanism that

mimics the joint motion observed in Fig. 10.6 while providing the bending
articulation observed in traditional snake-inspired robot designs. A solution was
found in the form of a new parallel mechanism. Advantages over simply adding
independent prismatic joints to traditional snake-inspired robot architecture
include (1) high-speed linear motion, (2) compact design, and (3) scalable design.
Due to the wide variety of torque–speed ranges for servomotors, the speed and
torque capabilities for the 2-DOF mechanism are modifiable without impacting the
physical size of the mechanism.
The conceptual design of the new parallel mechanism, introduced by (Hopkins
and Gupta 2012) and illustrated in Fig. 10.7, couples the output from two planar,
independently powered scotch yoke–like mechanisms. This mechanism differs
from a standard parallel mechanism in that the input paths do not mirror one
another. The two input links are independently actuated by a powered revolute
joint mounted to the base link and are connected to the output link through a
passive prismatic joint and passive revolute joint in series. A third path between
the base link and the output link is established through an additional passive
prismatic and revolute joint set acting along the x-axis of the base link, permitting
pivoting motion for the output link while resisting motion along the base link’s
y-axis. A key feature in this mechanism is that the sliding axes of the passive
prismatic joints of the output link remain perpendicular to one another throughout
the full range of motion, as seen in Fig. 10.7. Due to this simple but unique
arrangement, the constraints imposed by the orientation of the prismatic joints
prevent any motion of the output link, while the powered revolute joints are held
stationary.
The detailed design of the parallel mechanism concept utilizes slotted holes and
sliding pin joints to replicate the functions of passive prismatic and revolute joints.
These features allow for fewer parts, fewer assemblies, and a more compact
design. Each parallel mechanism, pictured in Fig. 10.8, is composed of two
Fig. 10.8 CAD model of parallel mechanism
servomotors with servo arms attached to the output shafts acting as the input links
to the mechanism. Each servo arm is attached to the output link of the mechanism
(a U bracket) through a slotted hole and pin joint (see left image in Fig. 10.8).
Essentially, the complete robot is a serial collection of modules. Thus, a
modular structure was devised in which two identical parallel mechanisms were
assembled in a single module. The two mechanisms are assembled serially in a
single housing, with the mechanisms’ orientation offset 90 apart about the x-axis
(direction of the linear expansion) of the module. Both mechanisms contribute to
the total linear displacement of the adjacent module, while one mechanism is
capable of providing yawing motion and the other provides pitching motion. This
assembly provides the potential for full spatial motion for the robot through the
fact that the modules are able to lift as well as pivot horizontally. In addition, this
configuration allows all modules to contribute to the expansion–contraction
capability of the robot, significantly increasing its speed.
In addition to the modular design, the other important design aspect of this
snake-inspired robot architecture is a variable friction force concept used to
provide anchoring points on the terminal ends of the robot to enable locomotion.
The variable friction force concept is a simple yet effective method of anchoring
one end of the robot to the terrain to provide a counter to the reaction forces of the
powered joints of the modules during forward or turning gaits. In nature, the
anchoring is accomplished by redistributing more of the animal’s body weight
across the surface of the foot to increase the friction force between the foot and the
terrain. This concept adopts a similar approach. The surface of the friction anchor
is covered in a material with a much higher coefficient of friction than the rest of
the robot’s housing material. The friction anchor is placed in contact with the
terrain by the action of a powered revolute joint as illustrated in Fig. 10.9. The
friction force, a function of the normal force between the anchor and the terrain, is
increased or decreased by varying the angle, hFA, of the revolute joint which
changes the amount of the module’s weight being supported by the friction anchor.
Fig. 10.9 Kinematic

representation and prototype
anchor
Figure 10.9a depicts the friction anchor in its nominal position, with the anchor’s
high coefficient of friction surface not in contact with the terrain, allowing the
terminal end of the robot to freely slide. Figure 10.9b depicts the friction anchor
surface in contact with the terrain with only a slight change in hFA, useful in low
reaction force gaits. Figure 10.9c depicts a large change in hFA, useful in high
reaction force gaits.
The force normal between the friction anchor pad and the terrain may be
modeled as a function of hFA and the extension of robot using the solid model
illustrated in Fig. 10.10. In addition to hFA and the prismatic extension, the model
also incorporates a vertical joint at the location of the friction anchor pad. This
joint is not actuated throughout the analysis and only serves as analysis point to
compute the force normal at the pad location, using analytical dynamic techniques.
The Lagrangian formulation is then used to find the inverse dynamic equations of
motion for the closed loop model in Fig. 10.10 (Tsai 1999).
Fig. 10.10 Friction anchor

model
Fig. 10.11 FDM fabricated

snake-inspired robot
prototype
Additive manufacturing techniques are often used to automatically construct

physical objects from computer-aided design models. Prominent additive manu-
facturing techniques for thermoplastics include selective laser sintering (SLS), fused
deposition modeling (FDM), and stereolithography (SLA). FDM offered the best
compromise between cost, part size, accuracy, and part strength. Hence, it was used
to realize this design (Hopkins and Gupta 2012). The use of FDM for the snake-
inspired robot prototype reduces the number of parts needed for assembly and
thereby reducing the weight. It also enables manufacturing of true 3D features. The
robot prototype, pictured in Fig. 10.11, was fabricated using FDM manufacturing to
demonstrate the new drive mechanism design and friction anchor mechanism.
The prototype, pictured in Fig. 10.11, was made primarily from ABS plastic. The
robot has a 69.85 9 69.85 mm cross-section. The robot has a contracted length of
850.9 mm and a fully extended length of 1,143 mm. The total mass of the robot
is approximately 1.36 kg. The robot prototype consists of three modules. The
prototype also includes a friction anchor at both terminal ends. Each parallel
mechanism is capable of a 90 range of motion and 48.68 mm of extension. Each
mechanism consists of two standard-sized Hitec HS-985MG High Torque ser-
vomotors. They are capable of 12.40 kg-cm of maximum torque and a maximum
speed of 0.13 s/60. The prototype demonstrated a maximum forward velocity of
196.65 mm/s (0.71 km/h) and a maximum turning rate of 26.32/s. By the nature
of the snake-inspired robot’s modular design, the forward velocity of the robot
may be increased by the inclusion of additional modules. Therefore, while the
demonstrated forward velocity does not yet meet the design goals, the velocity of
the prototype can theoretically achieve the speed requirement with the inclusion of
six additional modules.
This case study illustrated how rectilinear gaits and anisotropic friction exhibited
by snakes can be utilized to develop a mobile robot. Both concepts were imple-
mented in the robot in a very different form compared to their natural counterparts.
The rectilinear motion was exaggerated dramatically to achieve the high-speed
locomotion. This was necessary due to the inherent limitations of the existing
actuator technologies and the desire to realize the design with a small number of
joints due to cost considerations. The exaggerated rectilinear gait led to very good
performance in terms of speed. Limitations of the existing engineering materials
required implementing anisotropic friction concept by using the friction anchors
that control friction force by applying pressure on the ground.
The final design had parts with complex shapes. So in order to minimize the
number of assembly operations, the design was realized using a layered manu-
facturing process (e.g., fused deposition modeling). Currently, the housings are
made of ABS plastic which meets the strength and durability necessary to withstand
the impact forces generated in the modules during expansions and contractions;
however, a second material must be added to the housing’s outer surface to reduce
the coefficient of friction between the robot and the terrain. In the future, the
housing material should be chosen with the friction as a requirement to eliminate
the need for additional materials.
10.4 Robotic Masticator Test Bed
The goal of this project was to develop an experimental test bed for analyzing the
masticatory motions of animals (including humans) and establishing the quanti-
tative relationship between relevant geometric parameters (tooth geometries,
numbers, and types) as well as regimen parameters (joint forces, motions).
Therefore, we seek to (1) design a general purpose mastication simulator based on
parallel architectures, capable of producing the motions and forces encountered in
mastications and (2) ensure its working with a range of test foods within controlled
and carefully monitored scenarios (with uniform/non-uniform mastication cycles).
Such a setup would enhance our understanding of observed mechano-bio-physi-
ological synergies from varied perspectives. From a biological science perspective,
it will aid in understanding the anatomical variability in mastication for intra- and
inter-species comparative evaluations. It will enable understanding of the cumu-
lative functioning of different muscle groups by studying how various animal
populations preprocess their food by chewing and biting and how certain breeds of
animals kill their prey with improved estimates of biting forces by performing
hypothesis testing of behavioral analyses (Signore et al. 2005). From an
Fig. 10.12 Bio-mimetic human masticatory simulators: a Waseda University’s Masticatory

Robot Series (Waseda-University, 2003) and b BioMouth’s Dental Simulator (University of
Massey) (Xu et al. 2008)
engineering standpoint, such a study on masticatory performance would allow us

to objectively assess ‘‘chewability or performance index’’ based on food proper-
ties, facilitate the parametric design of dentitions/prosthetic additions, and analyze
the complex spatial kinematics of temporo-mandibular joint (TMJ)—the joint
connecting the skull and jaw (Kapur and Soman 2006; Komagamine et al. 2011).
Early simulator systems that were designed to mimic the jaw motions focused
mainly on the ‘‘form’’ aspects (muscular architecture and geometry), with lesser
emphasis on actual functional behaviors. As a result, such systems were used to
reproduce masticatory motions of only a specific breed (humans or animals) and
were too specialized to account for diverse species. Two of the most notable and
relevant bio-mimetic systems developed for mastication are shown in Fig. 10.12.
Therefore, building a generic mastication test bed is useful to enable better
understanding of underlying muscular actuations across different species and to
overcome the limitations in quantitative analysis of masticatory performance. The
typical masticatory cycles, namely opening and closing, are found to be a non-
uniform combination of both spatial translations and rotations through the entire
chewing process (Koolstra 2002). The envelope of jaw motions corresponds to that
of a 6-DOF joint (3 translational DOF ? 3 rotational DOF), which implies that the
robotic system must be a spatial manipulator (Narayanan 2008). Moreover, the
biting force estimates cover a range of values (10–300 N) for small canines to
humans to huge vertebrae animals (Signore et al. 2005). Therefore, the masticator
should not only possess desirable workspace but also be capable of exerting high
end-effector forces while simulating jaw masticatory cycles.
Fig. 10.13 A four-staged

realization of our masticatory
test bed project flowchart
Such diverse objectives pose significant challenges which we addressed in a four-

staged realization approach summarized in Fig. 10.13. In our work, an optical
passive marker-based MoCap system was used to capture the masticatory motions
of animals. The typical setup used for such a study is shown in Fig. 10.14, and
further details of various motion case studies (for canines and humans) are
discussed elsewhere (Narayanan 2008; Kannan 2008; Signore 2006).
Fig. 10.14 a Animal motion capture setup; b Recorded frames with marker tracking
These motion studies provide an estimate of jaw motion envelope (Posselt

envelope (Posselt 1957)) that needs to be contained completely within the
workspace of our masticatory simulator. Parallel architectures are in general
advantageous for most applications owing to their inherent stability, self-balanced
structural characteristics and varying low-to-high stiffness, which can be closely
attributed to biological systems. Kineto-static and quasi-static analyzes were then
performed to estimate optimal geometric parameters as well as the actuator
capacities required for candidate architectures. Thus, the sizing dimensions and
geometry of various components of the parallel platform were obtained taking
this into account which is detailed in (Narayanan et al. 2010). This enables us to
realize our objective of functional performance matching (motion envelopes and
bite forces that occur in a typical mastication cycle) rather than form-based
(anatomical/structural/musculoskeletal) similarities with the corresponding
biological systems (humans and animals).
Multiple spatial parallel manipulator (SPM) architectures considered for our

masticatory simulator include conventional Stewart platform (Kannan 2008),
6-R-U-S (revolute–universal–spherical with active revolute joints), and 6-P-U-S
(prismatic–universal–spherical with active prismatic joints) manipulators. How-
ever, due to the inherent advantages of locating actuators at the base (ground) in
terms of actuator efforts and structural rigidity as well as benefits of using pris-
matic sliders compared to revolute actuators (Narayanan et al. 2010), the 6-P-U-S
system was considered superior than others.
A detailed 3D CAD model of the corresponding robotic platform is visualized
in Fig. 10.15a. A comprehensive parametric study, based on maximizing work-
space Fig. 10.15b, manipulability at various points in the workspace, and actuation
limits, was undertaken to optimize design parameters for the 6-P-U-S manipulator
case (Shah et al. 2010). We then sought not only to accurately imitate the mas-
tication cycle (motion and food resistance) but also to reproduce dentitions of
different animals. For this purpose, CT scans and 3D laser scans of canine skulls
Fig. 10.15 6-DOF masticatory simulator system: a 3D CAD model; b Workspace; and
c Physical prototype
and mandible were used to obtain highly detailed 3D models of canine jaws which
were then fabricated into rapid prototypes and metal casts. Figure 10.16a shows
the resulting dentitions for labrador and bulldog specimens. These physical
prototypes were then mounted on the platform with the force transducer for
conducting our mastication experiments.
The robotic simulator designed for this purpose, shown in Fig. 10.15c, is capable
of measuring the bite forces during chewing cycle experiments using a 6-DOF
force–torque transducer (ATI Delta) mounted under dentition prototypes
(in Fig. 10.16). In the complete masticator setup (Fig. 10.17a), the platform was
(a) (b)
Canine Dentition
6-DOFATI Delta
Force Transducer
Fig. 10.16 a Rapid prototypes and metal casts of dentitions (upper and lower mandibles) and
b Masticatory simulator mounted with bulldog jaw and force transducer
Fig. 10.17 Raw MoCap data a HIL testing; b Z-coordinate; and c Bite forces of canine chewing
cycle
driven using canine MoCap trajectories for different types of food (hard to soft)
with varied canine dentitions and resulting bite force estimates were measured in
real time. Representative measured trajectory motions and bite forces, shown in
Fig. 10.17b, c, now offer a quantitative data streams to evaluate masticatory
performance/efficiency across a variety of test foods and dentitions.
A parallel architecture mastication simulator was designed and prototyped and

ultimately tested by hardware-in-the-loop jaw motion-tracking experiments. The
motion capture trajectories of various breeds of canines, from bulldogs to golden
retrievers, chewing various forms of food were recorded. The mastication
simulator proved to be capable of reproducing mastication cycles of varied breeds
of canines, with their varied ranges of end-effector motions (Narayanan 2008;
Kannan 2008). Further, by instrumenting the physical simulator with a
force–torque transducer allowed direct estimation of bite forces while performing
mastication experiments. While offering a good initial platform, this work has the
potential to be extended in many directions. For example, the concept of devel-
oping a ‘‘chewability index’’ requires a more comprehensive motion analysis
across different forms of animal foods as well as various breeds which will be
pursued in the future. Further, our test bed now offers an automated means of
performing both parametric sweep studies and careful design-of-experiment
studies to help characterize these aspects. Finally, using a more robust motion-
tracking method (electromagnetic) can overcome some of the challenges of
marker-based motion capture (marker occlusion, non-adhesion, etc.).
10.5 Discussion
In this chapter, we summarized different approaches to successfully realize

bio-inspired design and fabrication of robotic systems from an engineering per-
spective. In general, the approaches discussed here aimed to identify a promising
feature in the nature and then analyze it using the engineering principles and
scientific tools. Therefore, the resulting solution may not at all have resembled the
original source of inspiration. However, we demonstrated by means of three
specific bio-inspiration-based case studies that this approach can lead to novel
engineered systems.
The first case study on MAV highlighted issues of design complexity—in
achieving the challenging requirements with limited actuation and cost-effective
manufacturing methods. The second case study involving snake-inspired robot
explained critical issues in modular design, extended dynamic analysis and
physical prototyping inherent in biological organisms. The final case study
demonstrated the benefits of simulation-based design and its systematic applica-

tion to analyze a bio-mechanical subsystem of the jaw masticator.
Significant progress has been made in taking inspiration from biology and
incorporating these ideas into robotics. But numerous limitations still exist—on
the one hand, the multi-scale irregularities, inhomogeneities, and nonlinearities
inherent to biological systems still pose considerable technical challenges to
complete characterization and understanding; on the other hand, the relationships
between the distinct research vocabularies and approaches of biological and
robotics researchers persist in creating roadblocks. Hence, there is a critical need
not only to foster innovative research and technology but also to promote training
of researchers and engagement of a community well-versed in biology, design, and
robotics.
References
Bejgerowski W, Ananthanarayanan A, Mueller D, Gupta SK (2009) Integrated product and

process design for a flapping wing drive mechanism. J Mech Des 131:6
Bejgerowski W, Gerdes JW, Gupta SK, Bruck HA (2011) Design and fabrication of miniature
compliant hinges for multi-material compliant mechanisms. Int J Adv Manuf Technol
57(5):437–452
Bhatt RM, Tang CP, Lee L-F, Krovi V (2003) Web-based self-paced virtual prototyping tutorials.
In: Proceedings of 2003 ASME design engineering technical conferences and computers and
information in engineering conference (DETC/CIE), DETC2003/CIE-48201, Chicago,
Illinois USA, Sep 2–6 2003
Cardelli L (2005) Biological systems as complex system. 7th framework progress information
society technologies/future emerging technologies (position paper)
Dickinson M, Gotz K (1993) Unsteady aerodynamic performance of model wings at low reynolds
numbers. J Exp Biol 174:45–64
Ellington CP (1985) Power and efficiency of insect flight muscle. J Exp Biol 115:293–304
Gerdes JW, Gupta SK, Wilkerson S (2012) A review of bird-inspired flapping wing miniature air
vehicle designs. J Mech Robot 4:2
Gouker RM, Gupta SK, Bruck HA, Holzschuh T (2006) Manufacturing of multi-material
compliant mechanisms using multi-material molding. Int J Adv Manuf Technol 30(11-
12):1049–1075
Hopkins JK, Gupta SK (2012) Characterization of Forward Rectilinear-Gait Performance for a
Snake-Inspired Robot. In: Proceedings of 2012 performance metrics for intelligent systems
(PerMIS’12) workshop, College Park, MD
Hopkins JK, Spranklin BW, Gupta SK (2009) A survey of snake-inspired robot designs.
Bioinspiration & Biomimetics 4:2
Howell LL (2001) Compliant mechanisms. Wiley-Interscience, Hoboken
Kannan S (2008) Quantitative analysis of masticatory performance in vertebrates. M.S. Thesis,
University at Buffalo (SUNY)
Kapur KK, Soman SD (2006) Masticatory performance and efficiency in denture wearers.
J Prosthet Dent 95(6):407–411
Komagamine Y, Kanazawa M, Minakuchi S, Uchida T, Sasaki Y (2011) Association between
masticatory performance using a colour-changeable chewing gum and jaw movement. J Oral
Rehabil 38(8):555–563
Koolstra JH (2002) Dynamics of the human masticatory motion. Crit Rev Oral Biol Med 13(4):11
Lissmann HW (1950) Rectilinear locomotion in a snake (Boa occidentalis). J Exp Biol

26(4):368–379
Madangopal R, Khan Z, Agrawal S (2005) Biologically inspired design of small flapping wing air
vehicles using four-bar mechanisms and quasi-steady aerodynamics. J Mech Des
127(4):809–817
Madangopal R, Khan ZA, Agrawal SK (2006) Energetics-based design of small flapping-wing
micro air vehicles. Ieee-Asme Trans Mechatronics 11(4):433–438
Mueller D, Gerdes J, Gupta SK (2009) Incorporation of passive wing folding in flapping wing
miniature air vehicles. In: Proceedings of ASME mechanism and robotics conference, San
Diego, CA, 2009
Narayanan MS (2008) Analysis of parallel manipulator architectures for use in masticatory
studies, M. S. Thesis, University at Buffalo (SUNY)
Narayanan MS, Kannan S, Lee L-F, Mendel FC, Krovi V (2008) Virtual musculoskeletal
scenario-testing case studies. In: Proceedings of 2008 IEEE virtual rehabilitation (VR),
Vancouver, Canada: IEEE, Aug 25–27 2008
Narayanan MS, Chakravarty S, Shah H, Krovi V (2010) Kinematic-, workspace- and static-
analysis of A 6-P-U-S parallel manipulator. In: Proceedings of ASME international design
engineering technical conferences (DETC2010-28978), Montreal, Quebec, Canada, Aug
15–18 2010
Posselt U (1957) An analyzer for mandibular positions. J Prosthet Dent 7(3):368–374
Priyadarshi AK, Gupta SK, Gouker R, Krebs F, Shroeder M, Warth S (2007) Manufacturing
multi-material articulated plastic products using in-mold assembly. Int J Adv Manuf Technol
32(3-4):350–365
Rotheiser J (2004) Joining of plastics, handbook for designers and engineers. Hanser Gardner,
Cincinnati
Shah HL, Narayanan MS, Lee L, Krovi V (2010) CAD-enhanced workspace optimization for
parallel manipulators: a case study. In: Proceedings of IEEE 2010 conference on automation
science and engineering, Toronto, Ontario, Canada, Aug 21–24 2010
Signore MD (2006) A screw-theoretic framework for musculoskeletal system modeling and
analysis. M.S. Thesis, University at Buffalo (SUNY)
Signore MJD, Krovi VN, Mendel FC (2005) Virtual prototyping and hardware-in-the-loop testing
for musculoskeletal system analysis. In: Proceedings of IEEE international conference on
mechatronics and automation, Niagara Falls, Canada, 394–9 2005
Tsai L-W (1999) Robot analysis: the mechanics of serial and parallel manipulators. A Wiley-
Interscience Publication-Wiley, Hoboken
Waseda-University (2003) Dental robotics group [Online]. Available: http://www.takanishi.mech
. waseda.ac.jp/research
Xu WL, Torrence JD, Chen BQ, Potgieter J, Bronlund JE, Pap JS (2008) Kinematics and
experiments of a life sized masticatory robot for characterizing food texture. IEEE Trans Ind
Electron 55(5):12
Chapter 11
An Ontology of Biomimetics
Julian F. V. Vincent
Abstract Statistical analysis of the mechanisms and processes in biological

organisms (derived from published, peer-reviewed, research papers) reveals that
there are ‘design’ rules which could be used to facilitate technical design, thus
producing biologically inspired design without the necessity for the designer using
such a system to invoke biology or biological expertise since this has already been
done when the rules were extracted. Even so, this is not a necessary and sufficient
condition for good design. Four principles derived from the Russian system TRIZ
(widely used in technology as an objective system for solving problems inven-
tively) are highlighted and summarised as Local Quality; Consolidation or
Merging; Dynamics; Prior Cushioning. More design rules, derived in the same
way, are needed to expand the importance of information (sensu lato) and mate-
rials, two aspects that the TRIZ system currently does not deal with adequately.

Keywords TRIZ Inventive principle Cluster analysis Ontology Local quality

Consolidation or merging Dynamics Prior cushioning Information Hierarchy
11.1 Introduction
There are several databases of ‘biological design’ of which the best known is
‘AskNature’, produced online by the Biomimicry Institute. Although beautifully
presented, many of its sources are casual and not peer reviewed. Although it can be
searched for specific words or ideas, it cannot derive relationships that might be
implicit in the data. The European Space Agency had a database of biomimetics,
but this seems not to be accessible. It was based on work done by the late Centre
J. F. V. Vincent (&)
University of Bath, Bath, UK
e-mail: j.f.v.vincent@bath.ac.uk

270 J. F. V. Vincent
for Biomimetic and Natural Technologies in Bath but suffered from being small,
complex and inconsistent. Other databases are available but suffer from similar
problems. Difficulties with a database are that it has to be designed before it can be
used and it is difficult to expand its scope during its use; more importantly, a
database can only record history and cannot deduce new relationships. In reality,
these databases deal largely with the physical and chemical properties of organ-
isms and biomimetically derived designs represent a small proportion. There is
very little indication of how one can take a concept from its biological context and
transfer it to an engineering or technical context, which can be very different and is
the most difficult part of the design process.
There is a more basic problem. If the aim is to produce a bridge between engi-
neering and biology, a comparison has to be made between a largely analytical topic
and a purely descriptive topic (most modern engineering is reliant on mathematics;
the differences between organisms can be transformed into numbers—leading to
cladism—but this is a totally artificial transform). It is probably necessary to regard
technology from a descriptive point of view, which slews the approach towards
design but requires careful definition of what the technology is actually doing.
Used with care, lexicography enables new ideas and relationships to be derived
from biological texts, but in its current form, it does not create a generally
accessible database, being more of a search algorithm used in individual cases.
Even so, with careful choice of words and their interrelations, using a thesaurus
and a biological text, this approach produces good and unexpected results (Mak
and Shu 2008). On its own, it lacks data storage—the methods and results need to
be programmed into a search method. It has the advantage that it does not rely on
specifically chosen information and so is much more general, choosing examples
without any underlying assumptions.
11.2 A First Approximation
In the Centre for Natural and Biomimetic Technologies at the University of Bath,
UK, we initially tried a very broad approach based on solving problems: given a
certain problem, what change provides a solution? For instance, an insect or snake
sheds its old skin easily by having a structural adaptation (a line of thinner
material along the back which can split more easily). Or for a safety jacket to
provide flotation, it can increase its volume (thus occupying more space) by
inflation with gas. This is a very crude classification, and inevitably examples use
more than one of these sources of change. For instance, the inflating jacket requires
energy to compress the gas, but this can be said to occur at the size level of the gas
molecules, whereas the inflating jacket is nearer a metre across. Many more
examples (about 5, 000 from technology and 2, 500 from biology) were gleaned
from the Internet and biological publications, and the solutions to the problems
classified as changes in structure, substance (things), energy, information (do
things), in space and time (somewhere). The examples were then plotted on to
11 An Ontology of Biomimetics 271
two graphs, one for technology and one for biology (Fig. 11.1a, b) in which the
vertical axis represents all problems and the horizontal axis shows how these
problems are distributed according to size (Vincent et al. 2006).
It is important to realise that these two graphs do NOT indicate the quantity of
any of these six categories that is required to solve a problem—it is the CHANGE
that is important. These are CONTROLLING variables, and their presence is in
proportion to their frequency of use for producing change. There are many dif-
ferences between these two graphs, but the most salient is at the lm level, which
can be equated with materials processing. Technology uses energy as the main
control variable and requires substance (i.e. material resources). Biology is very
different, with energy the least important control parameter across the entire size
scale—again note that this is importance in control. It is not to say that energy is
not important for a living organism, just that it is not used very much when it
Fig. 11.1 a, b Different use of substance, structure, energy, space, time and information in
solving problems in biology and technology
Fig. 11.2 Craig’s roof

insulation. a Short-wave
reflector, long-wave
transmitter. b Oriented
honeycomb insulation.
c Black layer for radiation.
d Concrete heat sink
comes to managing change or adaptation (which may itself have energy con-
straints). There is also far less emphasis on input of raw material—in fact, nature
uses many fewer raw materials than technology, probably in part because it
recycles so much—and information (derived ultimately from DNA) is much more
important. Although it is difficult to understand quite how information is expres-
sed, the implication is that it informs the structure of the material leading to the
wide range of properties which, when density is accounted for, provides much the
same range of mechanical properties (or at least strength and stiffness) as tech-
nology. In other words, to a first approximation, the replacement of energy and
material with structure and information implies a biomimetic transformation of
materials processing, one of the most basic parts of engineering. This is important
because it frees technology from believing that all such transformations neces-
sarily require energy and material resource. Biology shows that careful assembly
of molecular components (e.g. liquid crystalline structures) can produce acceptable
results. This is crucial for the credibility of biomimetic materials processing, but
still leaves open the means of effecting such a transformation, although for
designers it is a definite direction in which to move.
An example of this transfer in thinking at a more general level is given by a
novel form of roof insulation, developed with the precept that the clear night sky
represents a heat dump at absolute zero (Craig et al. 2008). The normal solution for
cooling a building in an area of extreme solar heating is to use an air conditioner—
control by use of energy. Craig replaced energy with information and structure in
his design and produced a novel form of insulation that keeps direct short-wave
radiation from a concrete slab that forms the roof of the building, but allows the
heat stored in the slab to be radiated away through an orientated honeycomb which
forms the main part of the insulating layer (Fig. 11.2). Temperature differences in
the building between day and night can be as large as 13C. The information is in
the wavelengths and the orientation of the honeycomb and the structure is provided
by the honeycomb and the organisation of the layers.
11.3 Ontologies
A more sophisticated and general, yet detailed, approach uses programming

techniques developed for the Semantic Web. These are based on the resource
description framework (RDF), a formal language for describing structured
information. RDF allows exchange of information on the Web and so lends itself
to a communal resource. Information on RDF is readily available (Hitzler et al.
2010; Robinson and Bauer 2011).
The most recent development of RDF uses OWL2. OWL stands for Web
Ontology Language (don’t ask…). The public domain editor developed by Stan-
ford and Manchester Universities, ProtégéOWL, makes it relatively easy to
organise and analyse information. A brief description of OWL2 serves to illustrate
its advantage. Information is structured into a hierarchy of classes and subclasses
that you define. A class can contain any number of subclasses at the same level,
and the hierarchy of sub-, sub–sub-, etc., classes can go to great depth, although
this is not always desirable. Classes are given suitable names, and their members
are related to each other by a range of properties that you define. These properties
relate objects to each other (Object Properties), or they relate specific data
quantities to objects (Data Properties). Additions and changes can be made on the
fly, so that the size, range and complexity of the ontology are increased or
decreased as required. This, of course, is very different from a conventional
database. Within broad limits an ontology can undergo great modification at all
stages of its use and development. This is particularly convenient since the
temptation is always to include too much information at the outset since one is
never sure what information is necessary, sufficient or significant in any other way.
With Protégé, the redundant information can easily be deleted or reformulated.
The heart of the ontology is the establishment of relationships between the
classes and their members. Crucially, the reasoning is based on an ‘open world’
model as opposed to ‘closed world’ which is the case with simple databases. In a
closed world assumption, if something is not stated to exist, then it is assumed not
to exist. However, in an open world model, if something is not stated to exist, then
it is impossible to say whether or not it does exist. A clear and explicit statement as
to its existence or non-existence has therefore to be made. Large parts of ontology
rely upon set theory. Classes can be defined as disjunct (i.e. non-overlapping) or
intersecting (overlapping sectors). A Property can be functional, relating an item to
only one other (e.g. Sally hasBirthMother Liz) or inverse functional (Liz is-
BirthMotherOf Sally). If we say that Liz isBirthMotherOf Sally and that Betty is
Birth MotherOf Sally, we deduce that Liz and Betty are the same person.
A Property can be transitive (any friend of yours is a friend of mine), symmetrical
(I’m your friend and you are my friend), antisymmetrical (your friend is my friend
but I’m not your friend) or reflexive (I know you and I know myself). All these
properties can have their domain and range specified. This can ensure that rela-
tionships are clearly restricted and false inferences are difficult to make. In a
hierarchical system, such as a living organism the ability to make Object Prop-
erties transitive is particularly useful, since it means that the properties of an
organism can be retrieved from within any level of the hierarchy. This allows the
establishment of a chain such as (amino acid)—(protein)—(collagen)—(bone)—
(femur)—(endoskeleton). An example is shown in Figure 11.3, where collagen is
explicitly stated as being part of a number of materials (in squares) but remaining
reliance on collagen is implied within the rest of the relationships. Thus, using a
Fig. 11.3 Collagen can be inferred, by ontological reasoning, to be a component of many

systems. The rectangles surrounding the central one represent the only materials stated to contain
collagen. Topics at the ends of the chains are case studies of possible biomimetic significance.
The example of lightweight skeleton is mentioned in the text
range of rules of relatedness, a Web of interrelations and properties is generated.

These properties relate classes by means of restrictions ‘some’ (meaning ‘at least
one’—an existential restriction) and ‘only’ (a universal restriction).
There are more rules of relationship, for instance distinguishing between a
necessary condition for an object to be a member of a class (which is termed
primitive) and a necessary and sufficient condition, which makes the object a
member of a defined class. Say we have a primitive class A for which membership
is defined by the possession of characteristic A(c). This is not good enough for us
to say that any object that has A(c) is a member although it could be. But if class A
is made Definitive, then possession of A(c) can also be made a sufficient condition.
This difference is useful if we have a member of class B that has the characteristic
A(c), then we know that class B is subsumed by class A.
Developing from this, a number of restrictions can be chained together using
Description Logics to interrogate the ontology, with the relations ‘some’, ‘only’,
‘and’, ‘or’ and ‘not’. In this way, a concept can be derived which is implicit in the
available data but which may not be obvious.
In this ontology, the information is organised using a technique from TRIZ, a
Russian system for solving technical problems. TRIZ is a collection of interrelated
methods; the method around which the ontology is structured is probably the
simplest of all of them. It owes much to a part of Hegelian Philosophy (which is
taught early in the Russian education system) that is in turn based on the ideas of
Plato and Heraclitus (thus establishing its credentials). A problem can often be
characterised as due to the interaction of a pair of factors, one of which is the factor
that is to be improved (Hegel’s thesis) and the other, any perceived disadvantage
of implementing the improvement (Hegel’s antithesis). Resolution of the problem
is achieved by making a change in the system, producing Hegel’s synthesis.
Genrich Altshuller, the inventor of TRIZ, used this construct to describe the action
of patents. He drew up a list of factors that contribute the thesis and antithesis
(currently, there are 39) and of the changes (Inventive Principles. Currently, there
are 40) that produce the synthesis. Since each factor can be a thesis or an antithesis
it is possible to arrange them along two sides of a matrix, with the appropriate
synthetic Inventive Principles within the body of the matrix. We thus obtain a
matrix (called the Contradiction or Conflict Matrix) containing 1,521 cells, most of
them populated with up to four Inventive Principles and representing the resolution
of a problem whose definition and solution were obtained from published patents.
The matrix thus represents best practice of engineering (and related technologies)
and so is a convenient baseline for comparisons.
The introduction of biology into this system is managed by considering the
problems solved by living organisms in the same way that a patent is considered: an
organism has a problem, what is its solution? There is a difficulty. Although, given a
problem, it is logically simple to show that the solution has been derived from the
definition of the problem using Hegel’s approach, it is logically impossible to prove
in the other direction. For instance, given the answer ‘42’ it is impossible to state
with certainty the question that led to that particular answer. Indeed, there is an
infinity of questions. An organism is a compendium of solutions to biological
problems that may have been solved millions of years ago by an organism living
under conditions very different from the present. That the solution worked is evi-
denced by the survival of the organism; the nature of the problem has to be inferred,
making this the most difficult part of generating a believable ontology. Ultimately,
believability comes from the recognition of internal patterns that result from the
logic of the ontology (which comes from the individual and independent decisions
made during the assembly of the data) rather than some subliminal desire to impose a
pattern on the data. I have tried very hard to make all judgements as objective and
independent as possible (see next paragraph). Ultimately, it needs a small team of
people who can introduce such independence with ease.
The starting point is the abstract of a published, peer-reviewed, scientific paper
chosen from as wide a field of biology as possible, covering molecular biochemistry
to ecosystems and behaviour. The only proviso is that, like a patent, the paper
presents the analysis and solution of a biological problem. Consider the problem of
making a skeleton from bone. Calcified bone is approximately twice as dense as
other tissues, so it is important to minimise the size of the skeleton, but this implies
increasing stress on the bones and the potential for fatigue fracture. If muscle mass
increases in linear proportion to bone mass, extending a bone’s fatigue life by
increasing its cross-sectional dimensions may not be effective because the inertia of
bigger bones would require larger muscles with increased skeletal loads. Thus, bone
remodelling to remove fatigue damage may be essential for the existence of rela-
tively large, long-lived vertebrates. The problem can be defined as the desirable
strength and stiffness of bone (Feature number 14) but the undesirable force needed
to move it because of its weight (Feature number 1). Relevant Inventive Principles
(IPs) are then allocated by making judgements on their applicability. The choice has
been made considerably easier by allocating typical functions to each of the IPs;
Table 11.1 shows those developed for IPs 25 and 40, both relevant to the solution of
the problem of the skeleton. Some 450 such functions have been identified, each of
them unique to a single IP. Other IPs for this problem are IP31 (use porous material;
introduce lightness) and IP34 (discarding and recovering material—for example,
remodelling the bone). These IPs are then compared with the groups of IPs in the
Conflict Matrix, each group defined by a pair of conflicting features (thesis and
antithesis) looking for a minimum of 3 out of 4 of the Inventive Principles which
biology would use to resolve the strength/weight duality. This is a check to see
Table 11.1 Relevant entry in the ontology

IP25: self-service
Remodel
Self-assemble
Self-clean
Self-regulate
Self-repair
Use passive feedback
Use proprioception
Use waste
IP31: porous material

Fill pores or cells
Filtration system
Introduce emptiness
Introduce pores or cells
Use cellular material
IP34: discarding and recovering

Discard and recycle
Discard objects
Recycle
Regenerate
Resharpen
Restore used or consumed objects
IP40: composite materials

Introduce chemical heterogeneity
Increase number of layers
Make composite
Make heterogeneous
Use fibrous composite
Use particulate composite
Use structural heterogeneity
Use a composite system
Fig. 11.4 The ontology entry for a bony skeleton and the effort required to move it
whether TRIZ has the same group of IPs as biology, but representing a different pair
of Features. If the Matrix reveals such extra groups, the Features proposed as defi-
nition of the problem are reviewed and may be changed. Otherwise, the original pair
of Features is retained, and the IPs which they define are used in the ontology. These
two sets of IPs (from biology and from TRIZ) represent the degree to which biology
and TRIZ are comparable. In this instance, there is a big difference—it is possible in
biology to repair the skeleton while it is being used. This is not normally the case in
an engineering context. The relevant entry in the ontology is shown in Fig. 11.4.
It is now possible to ask more interesting questions of the ontology. For

instance, if the problem is to do with weight (which might be a positive or negative
aspect) then the Description Logics feature of Protégé enables you to ask for the
relevant examples to be listed. More interesting is to ask at the same time which
examples in the ontology have the biological solution to weight as the introduction
of porosity. Unsurprisingly, these turn out to be lightweight skeletons and bird
bones. But if the question is raised—which of those examples that have weight as a
problem would have used porosity had the problem been solved by a more
engineering approach, the examples flagged up are shark tail and sycamore seed.
Shark tail is more concerned with hydrodynamics, since the tail contributes to an
upward force as the fish swims. However, the sycamore seed reduces its rate of
descent (and hence a functional reduction in density as measured by its apparent
response to the force of gravity) by autorotation rather than by being feather-light.
While this is a mechanism present in the autogyro and, to a limited extent, in the
helicopter, it is not necessarily the first solution that comes to mind.
Using the TRIZ system to define solutions to problems is probably useful in
many realms, especially engineering. But biology is as much descriptive as any-
thing, and it turns out that this aspect can be included in the ontology, using the
Inventive Principles and their underlying Functions. The comparison with engi-
neering is not then available, since there is no thesis/antithesis pair generated, and
the ontology becomes more like a database of effects. This aspect has the
advantage that features of biological materials which are common to them all (and
which therefore will be less likely to be used as a medium for change), such as that
they are (as far as I know) all composites and all hierarchical, can be catalogued.
At the time of writing, this addition to the ontology is only just beginning, and the
analysis presented here is based on the thesis/antithesis/synthesis data.
11.4 Analysis
Over a period, I have accumulated such results for rather more than 260 examples
of biological problems. The resulting set of descriptions forms a pool of data that
can be mined to discover general principles and trends that differentiate biological
and technical solutions to problems, thus generating a set of design rules for
biomimetics that can be used in the absence of biological knowledge. This analysis
is preliminary and I present here only some basic simple statistics and cluster
analysis. However, the differences and similarities are rational and continue the
direction of results that the graphs shown in Fig. 11.1 have suggested. The current
version of the ontology of biomimetics, together with limited instructions and
comments, is available at http://wiki.bath.ac.uk/display/OOB/.
The similarity between the technical and biological solutions to problems can
be calculated from the coincidence between the TRIZ and biological solutions.
This ranges from zero to 100 (total identity) and is shown as a cumulative total
(Fig. 11.5). The straight line is not necessarily significant. The proportion of
Fig. 11.5 The similarity

between biology and
technology in terms of the
Principles used to solve
biological problems
biological problems whose solution is not provided by TRIZ is about 60 %. About

5 % are totally identical to the TRIZ solution. This range covers the estimate of
12 % similarity (made from a 6 9 6 matrix generated in an independent study,
using the six factors of Fig. 11.1) (Vincent et al. 2006) and 18 % (made in an
independent study of the adaptations of insect cuticle) (Vincent 2005). Clearly, the
degree of difference between biology and technology depends on the method of
measurement, which should not be surprising. And seemingly the greater the level
of detail at which the comparisons are made, the less the difference becomes.
The novelty which biology can introduce to technology seems to be evenly
distributed across the size range (Fig. 11.6), which shows the novelty at the nine
levels of hierarchy used in this study. This is comforting for those practising
biomimetics, since it suggests that the concepts can be transferred to a wide range
of technologies, from molecules to road systems. However, the data are still rather
sparse and would benefit from more biological examples.
Things become more interesting when we look at the detail.
The Inventive Principles are fractionated into the six classes of function out-
lined above, primarily to allow an independent test of the results of Fig. 11.1. The
decisions as to which class the Principles were placed, made by Olga and Nikolay
Bogatyrev, were based on the original definitions expressed in Russian by Alts-
huller. The results are presented (Figs. 11.7, 11.8, and 11.9) as distributions
showing the relative number of Inventive Principles used in the solution of
problems in biology and technology. The number has been normalised against the
total number of Principles used. Thus, a line plotted from (0,0) to (100,100) or its
equivalent divides those Principles mainly used by technology from those mainly
used by biology. This is shown on all the graphs.
Fig. 11.6 The proportion of

novel solutions is
independent of the level of
hierarchy
Fig. 11.7 The relative

occurrence of energy-based
Principles in the biological
and technological solutions to
biological problems. 8 Anti-
weight; 12 Equipotentiality;
18 Mechanical Vibration; 28
Mechanics substitution; 29
Pneumatics and Hydraulics;
37 Expansion; 38 Strong
oxidants; 39 Inert atmosphere

occurrence of structure-based
biological problems. 1
Segmentation; 2 Extraction;
3 Local quality; 5
Consolidation; 6
Universality; 7 Nesting and
Hierarchy; 16 Partial or
Excessive Action; 24
Mediator; 34 Discarding and
Recovering; 40 Composites

occurrence of time-based
biological problems. 9 Prior
Counteraction; 10 Prior
action; 11 Prior Cushioning;
15 Dynamics; 19 Periodic
Action; 20 Continuity; 27
Designed Obsolescence
I have chosen only three of the possible six subfactors of Fig. 11.1 to illustrate
the differences between biology and technology that this study has generated.
These differences, because the Inventive Principles are suggestions for the changes
Fig. 11.10 Cluster analysis of the TRIZ Conflict Matrix
required to solve a problem, represent the changes that would move the design
process towards biology and hence achieve apparently counter-intuitive goal of
allowing designers to be biomimeticists without knowing anything about biology.
Considering Energy (Fig. 11.7), the only IP used more by biology is ‘Anti-
weight’. The suggestion that TRIZ makes in the definition of this IP is ‘merge a
heavy object with one that provides lift or use aerodynamic, hydrodynamic,
buoyant and other forces’. More important is that this result places seven of the
Energy Principles in technology and only one in the biology, mirroring the dif-
ferential in importance of energy as a control parameter shown in Fig. 11.1. Thus,
this result has been confirmed using an independent metric.
Considering Structure (Fig. 11.8), the balance between technology and biology
shown in Fig. 11.1 is once again reflected, with three Principles in technology and
7 in biology, once again confirming, in more detail, the general picture presented
in Fig. 11.1. Note that the Principle ‘Composites’ (IP40) is commoner in tech-
nology, possibly because composite materials are ubiquitous in biology and so do
not constitute a novelty that could introduce change. The same could be said of
‘Nesting and Hierarchy’ (IP7) which, although in the biology half of the diagram,
is not outstanding. Nearly, all biological materials and structures are hierarchical,
being formed from the molecule up. Local Quality (IP3) makes up for some of this
by suggesting heterogeneity, zonation, etc. Remember, however, that the inference
taken from Fig. 11.1b is that Information is directly responsible for Structure.
Unfortunately, as indicated by Fig. 11.1a, information is almost literally trashed by
technology, being squozen out at the level of materials processing. This is reflected
in the poor showing of information in the list of IPs—it contributes only four of
Fig. 11.11 Cluster analysis of biological examples with no TRIZ equivalent
them, whereas, statistically, it should contribute six or seven. This point will be
returned to. Meanwhile, the other two important structural factors are Segmenta-
tion and Consolidation.
Time is marginally more important in biological than technological problem
solving (compare Fig. 11.1a, b), and more detail (Fig. 11.9) shows that Dynamics
and Prior Cushioning are important. In fact, ‘Prior’ is an important factor in the
more biological Principles, suggesting that biology designs its structures and
mechanisms in a way that we might regard as introducing redundancy. In biology,
this preparedness for action, and the ability to adapt which ‘Dynamics’, shows that
biological organisms are far more flexible and adaptable in their response to their
surroundings and possible requirements.
11.4.1 Cluster Analysis
The questions to be asked are not just which are the commonest IPs in use, but do
they show any sort of relation with each other. The four or so IPs associated with
each solution (the crossing points in the Conflict Matrix of the pairs of thesis and
antithesis) are compared as a group. Principles which are related, and which will
tend to occur together in the synthesis of a problem, will be classified similarly and
belong to the same cluster. There are several methods for calculating and pre-
senting clusters—the one used here produces a tree. The analyses were performed
using the public domain language R. Unfortunately, I have not yet been able to
analyse and compare these results statistically, so what follows is a preliminary
inspection.
Figure 11.10 is derived from the entire TRIZ Conflict Matrix, which is taken as
providing the technology baseline. The order in which the IPs are numbered bears
no relation to their recommendation for the design or to the principles by which
they work. The first four Principles occur separately on the left edge of the cluster,
remote from the rest of the tree. This shows that they have only very weak
similarity to the rest of the Principles or to each other. It also seems reasonable that
the most obvious Principles would be the first to be identified, so we can say with
some certainty that these isolated Principles are mutually independent and strong.
To the right of the tree, IP10 is also rather isolated, so we can say something
similar about that one. IP35 (which in effect says ‘If there is anything you can
change, then change it’!) is by far the commonest in the TRIZ Matrix and is
obviously the least related to any other Principle. It is also the Principle that has
the most associated Functions (3 or 4 times more than any other Principle). It is
well accepted amongst the TRIZ community that IP35 is a dump for a large
number of solution strategies that could not be shoehorned into any other Prin-
ciple. The last five Principles in the Matrix (making up the total of 40) were
probably added at a much later date. This brief and rather casual analysis (which
will be expanded and improved) exposes one of the basic problems with the TRIZ
system—the Inventive Principles are not equally weighted; IP35 is grossly
overweight.
However, so long as all the analyses are performed from the same starting
position with the same assumptions, comparisons are still valid.
Figure 11.11 is the cluster tree for 117 biological examples from the ontology
whose deduced Inventive Principles have no counterpart in the TRIZ matrix. There
are a few similarities with the TRIZ Conflict Matrix cluster—notably, the position
of IP1. IP3 is the dominant one (the same result came from a study on insect
cuticle) (Vincent 2005), and IP35 has, in effect, been lost. Principles 5, 11 and 15
are also very important. These four Principles are the strongest indicators of
biologically inspired design.
The thought occurs that, even though there is a significant number of examples
whose Inventive Principles are more or less what TRIZ would predict (given the
thesis/antithesis definition of the problem), the biological solution has nonetheless
selected a set of results which does not resemble the cluster results for the entire
TRIZ Conflict Matrix. This turns out to be true (Fig. 11.12) in that IP35 has been
greatly reduced in its significance. Otherwise, this cluster seems to be not too
different from the basic TRIZ cluster.
11.5 Systematic Omissions
There are some clear gaps when the clusters are compared with ideas at the
beginning of this chapter. Most notably, there is no explicit mention of Informa-
tion, which Fig. 11.1b shows clearly is a major component in the biological
solution of problems. The answer appears in Fig. 11.1a—in our technology, we
Fig. 11.12 Cluster analysis of biological examples with strong TRIZ identity
have very little use for information—at least, this was true during the period, about
30 years ago, when TRIZ was being formulated. The result is that information is
not considered as an important part of the solution of problems. Unfortunately,
many people treat TRIZ as a set of ultimate truths, so this lack tends not to be
perceived by the TRIZniks who are most expert in this field. The very least that
Fig. 11.1 shows is that technology is definitely not an assemblage of techniques of
Best Practise. Biology (the most sustainable and arguably, therefore, effective way
of doing things on this planet) differs in most respects. These two graphs represent
the first real challenge to technology in that they not only show where it is wanting,
but also show what is needed. And the recognition of the importance and use of
information in modern technology is crucial and must be admitted into the model.
To be fair, information at the molecular level is being developed in the science of
nanotechnology. In order to see how it should be used within the process of bio-
logically inspired design, as developed here, we need to understand information
from the molecular level. This brings in topics such as thermodynamics, liquid
crystals and self-assembly, water as an assembly medium. This is a distinct lack and
should go straight to the top of the biomimeticist’s wish list. What other ways do we
have of manipulating information that can be expressed in such a way that they can
be incorporated into TRIZ system (and thus into the system presented here)?
A second clear gap is the materials used. This chapter has addressed problems
of how to design structures and mechanisms but has said little about how they are
to be realised. The assembly process is probably covered by information, but the
materials are integral to this. Biological systems are assembled in a watery
environment that provides the context for the types of interaction that drive the
assembly process. Hydrophobicity is the driving force for molecular interaction—
its outcomes are summarised in the design processes suggested by Principles 3 and
5—Local Quality and Consolidation. The advantage of the biological system is
that the materials are easily recycled, but this is merely another way of saying that
all the molecular processes of biology are provided by a relatively small range of
chemical processes that are thus easily integrated. The recycling of metal is an
equally well-closed loop, except (of course) that it requires large amounts of
energy. Unfortunately, commercial pressures, commercial secrecy, the profit
motive, even self-aggrandisement, all militate against the generation of a fully
integrated system of materials. Metals are generally recyclable only because there
is not an awful lot you can do with a material which has to keep in more or less its
elemental state in order to be useful. Plastics come under commercial pressures
and so are highly varied and generally incompatible and so difficult to recycle.
Perhaps therefore this TRIZ-based system needs some design principles to aid the
choice of materials.
A third apparent omission is that the pervasive appearance of composite
materials and hierarchy in biology is not salient in this study. Presumably, this is
because all biological materials and structures are made using these principles.
Therefore, they cannot be introduced, although they can be fine-tuned. Even so,
they appear in the Principles of Local Quality and Consolidation.
References
Craig S, Harrison D, Cripps A, Knott D (2008) BioTRIZ suggests radiative cooling of buildings
can be done passively by changing the structure of roof insulation to let longwave infrared
pass. J Bionic Eng 5:55–66
Hitzler P, Krötzsch M, Rudolph S (2010) Foundations of semantic web technologies. CRC Press,
Boca Raton
Des 19:21–28
Robinson PN, Bauer S (2011) Introduction to bio-ontologies. CRC Press, Taylor and Francis
Group, London
Vincent JFV (2005) Deconstructing the design of a biological material. J Theor Biol 236:73–78
Vincent JFV, Bogatyreva OA, Bogatyrev NR, Bowyer A, Pahl A-K (2006) Biomimetics—its
Chapter 12
Evolution and Bio-Inspired Design:
Natural Limitations
Frank E. Fish and John T. Beneski
Abstract Biomimetics is the incorporation of novel structures and mechanisms

from nature into the design and function of engineered systems. Promotion of
biomimicry has been justified on the basis that evolution has modified structures
and functions in organisms to achieve optimal solutions and maximize perfor-
mance. Such justifications reflect an incomplete understanding of evolution and
constraints imposed on biology. Evolution is not a conscious or predictive process
and does not drive toward perfection. Organisms are not optimal with regard to
any one specific function. Where a biological feature will out-perform available
technologies, these features can be targeted for assimilation into bio-inspired
designs. For engineers and entrepreneurial investors interested in a biomimetic
approach, an understanding of evolution and the limitations and constraints that
have shaped biological organisms are necessary to avoid unsupportable and
overzealous claims.

Keywords Evolution Darwin Natural selection Ffitness Adaptation

Phenotype Genetic variation Design Reproduction and replication MantaBot
Autonomous underwater vehicle (AUV) Muscle
12.1 Introduction
The field of biomimetics and bio-inspired design has become an important source
of innovative ideas. Biomimetics attempts to produce engineered systems that
possess characteristics, resemble, or function like living systems (Vogel 1998).
The biomimetic approach seeks technological advancement through a transfer of
F. E. Fish (&) J. T. Beneski

Department of Biology, West Chester University, West Chester, PA 19383, USA
e-mail: ffish@wcupa.edu

DOI: 10.1007/978-1-4471-5248-4_12, Ó Springer-Verlag London 2014
288 F. E. Fish and J. T. Beneski
innovation from natural to engineered systems. By emulating these biological

characteristics in those instances where the performance of living organisms is
superior to manufactured devices, the performance of engineered systems may be
improved through biomimetics.
It has been a long-standing idea that new technologies can be developed from
nature (Fish 1998; Vogel 1998; Lu 2004; Bar-Cohen 2006). Animals and plants
have served as the inspiration for various technological developments. Copying
biological organisms by the biomimetic approach attempts to seek common
solutions from engineering and biology for increased efficiency and specialization
(Vincent 1990; Ralston and Swain 2009). Bio-inspiration extends biomimicry by
expanding and improving the original biological concept (Ralston and Swain
2009). Engineers that target the diverse specializations exhibited by organisms for
technology transfer can effectively reduce the time for the development of inno-
vative technological solutions.
Biologists have turned to physics and engineering for the answers when seeking
explanations for the function of various adaptations involved with physiology and
structural anatomy. The biology of adaptation is thus inferred through reverse-
engineering. However, with biomimetics, the engineers must look to the biologists
to provide working examples and mechanisms of physical phenomena. Novel
approaches by living organisms, therefore, form the basis of the development of
new technologies.
Biological organisms are in their own right machines. They are fully autono-
mous, self-powered, and self-repairing. The component structures that comprise
these entities are adapted for particular functions that allow the organism to
occupy a specialized environment space, or niche. The spectacular diversity of
life-forms has produced a variety of mechanical and physiological solutions for
interacting with both the biotic and abiotic environment. Inspired by such solu-
tions, engineers are attempting to build machines that mimic the functions of
organisms, but these machines and structures are limited and are nowhere near as
sophisticated or versatile as real biological forms (Denny and McFadzean 2011).
Because biological designs result from the evolutionary process of ‘‘natural
selection’’ (Darwin 1859), biological organisms are considered to have already
performed the ‘‘cost-benefit-analysis,’’ optimizing particular designs for specific
functions. In this sense, biology has provided a design prototype (Allen 2010).
Over the course of millions of years, different lineages of organisms have, in
effect, experimented with various combinations of morphologies, physiologies,
and behaviors to enhance performance. The planet is thus considered to be an
enormous natural laboratory, where an infinite number of experiments have been
attempted over the eons (Lu 2004; Bar-Cohen 2012). The results of these exper-
iments are the diverse assemblages of animals, plants, fungi, protozoans, and
bacteria that inhabit the Earth today. They are considered the evolutionary win-
ners. The losers are the organisms that have gone extinct or never came into
existence. They were assigned to an evolutionary trash heap of maladaptation. It is
‘‘survival of the fittest.’’ However, such an appreciation of evolution is patently
false. This erroneous rationale is too often employed as a natural justification for
12 Evolution and Bio-Inspired Design: Natural Limitations 289
the perfection of nature and the power of the biomimetic approach. Many human-
contrived inventions are thought to be the result of biomimicry (i.e., net spider
web, submarine dolphin, wheel rolling round fruit; Lu 2004), although they may
only be analogies with no cause and effect.
The purpose of this chapter is to provide a fundamental understanding of how
evolution works and its importance with respect to biomimetics and bio-inspired
design. The chapter has been written primarily for engineers as the mechanism of
evolutionary change is generally known to biologists. However, biologists can also
benefit from this chapter in understanding the inherent differences in the disci-
plines and seeing how a productive biomimetic design may be developed through a
synergy of biology and engineering. By examining the relationship between
evolution and biomimicry, limitations and constraints can be elucidated for nature-
based technologies. It will then be possible to better target biological designs for
technology transfer and more quickly bring biomimetic products to market.
12.2 Evolutionary Mechanics
Although the general principles of organic evolution are widely known, its
implications are often misinterpreted or misdirected. This is unfortunate because
evolution is arguably one of the most profound tenets of modern biology in that it
provides a solid, unifying concept for all of biology’s disciplines and subdisci-
plines as well as a common thread for continued investigation. At its core, evo-
lution is a conceptual framework for the process by which living systems change
over time. Because the time frame for evolutionary change is geological, it is
difficult if not impossible to observe these changes directly. Instead, evolutionary
biologists must rely on the results (or products) of evolution, which include all
living systems past and present.
The genesis of modern evolutionary theory has a rich and colorful history
involving numerous personalities. However, its formalization, development, and
ultimate acceptance within the scientific community are generally attributed to
Charles Darwin as documented in his seminal work ‘‘The Origin of Species’’
(Darwin 1859). In developing his theory, Darwin sought to explain the bewildering
diversity of life by answering two basic questions: (1) How do organisms change
over time? and (2) how do new types of organisms originate? Darwin’s concept of
evolution can be summarized as a series of observations and deductions
(Fig. 12.1).
In contrast to his contemporaries, Darwin’s extensive observations of nature led
him to conclude that evolution operated at the level of the population rather than
the individual. Darwin further concluded that populations evolved (changed over
time) by differential reproductive success; that is to say, individuals with favorable
traits leave more offspring than individuals with less favorable traits, thereby
increasing the frequency of the favorable traits in the next generation. The measure
of an individual’s ability to pass its traits into the next generation is termed fitness.
Fig. 12.1 The observations and deductions of Darwin (1859) that lead to the idea of evolution by
natural selection
Individuals with a high fitness pass more of their traits into the next generation
than individuals with a lower fitness. The process by which favorable traits are
preferentially passed from generation to generation is termed natural selection.
Although Darwin’s theory provided an elegant framework for the process by
which populations change over time (natural selection), it lacked an adequate
mechanism. Darwin’s theory of natural selection relied on the passing of favorable
traits from parent to offspring. However, the work of Gregor Mendel and therefore
the genetic basis of inheritance were unknown to Darwin, and he therefore was
unable to adequately explain how the traits of parents could be transferred to their
offspring. R. A. Fisher, J. B. Haldane, Sewell Wright, and T. H. Huxley finally wed
Darwinian evolution with the principles of genetic inheritance. The resulting
elaboration of Darwinian evolution is variously referred to as neo-Darwinism, the
Synthetic Theory of Evolution, or the New Synthesis. One of the first outcomes of
this union was the field of population genetics.
According to the New Synthesis, evolution can be defined as the heritable
changes that occur in a gene pool over time due to differential reproduction (i.e.,
ability to produce more offspring due to a heritable trait). This simple definition
encompasses three key concepts: change, time, and populations. In studying
evolution, it is important to distinguish between statements that apply to popula-
tions and statements that apply to individuals. Evolution occurs at the level of the
population (populations evolve, individuals do not evolve). By contrast, natural
selection occurs at the level of the individual by favoring the reproductive success
of the fittest phenotypes. Phenotypic traits are characteristics of individuals that are
coded for by specific genes (e.g., hair color, body size, wings). Therefore, if the
frequency of a gene within a population changes, the frequency of the corre-
sponding phenotypic trait in that population will also change. As such, evolu-
tionary change does not occur within individuals, but occurs in a population over
time as a consequence of which genes are passed from generation to the next
generation. Evolutionary fitness is a measure of an individual’s ability to pass
genes on to the next generation. The more genes an individual passes on to the next
generation, the higher the fitness of that individual.
It is important to note that by definition fitness is relative (there is no absolute
fitness scale) and to recognize that fitness is not differential survival. To measure
fitness, we compare the reproductive success of all phenotypes and then assign the
phenotype with the highest reproductive success a fitness value of (1.0). For all other
phenotypes, we assign a fitness value that reflects their reproductive success relative
to the phenotype with the highest reproductive success. For example, if a second
phenotype produced 80 % as many offspring as the fittest phenotype, it would be
assigned a fitness value of (0.8); if a third phenotype produced 50 % as many
offspring as the fittest phenotype, it would be assigned a fitness value of (0.5).
It is also important to note that selection operates on phenotypes, not the
underlying genetics in that different combinations of genes or alleles can yield the
same phenotype. Phenotypic traits include any morphological, physiological,
behavioral, or other definable characteristic of an individual. Natural selection
determines which phenotypic traits are favored and which are not. To be favored, a
trait must increase the fitness (reproductive success) of an individual with that trait
over an individual without it. Favored traits are often described as adaptive traits,
so that adaptation is the process of acquiring favorable traits.
As a result of evolution, populations change through time (as traits change in
response to changing gene frequencies) with the direction of change determined by
natural selection. Evolution only works by selecting between alternative pheno-
types already present in the population. This makes variation the raw material for
evolutionary change. Yet, the process of evolution itself works by eliminating
variation through natural selection. Continued evolution is therefore dependent on
additional mechanisms, which add variation to a population.
Genetic variation is added to populations by several mechanisms. New genes
can arise through mutations, genetic recombinations, or can be introduced (or
reintroduced) by gene flow from neighboring populations (gene pools). In addition,
the distribution and availability of genes within a population can be altered by
various mechanisms or by chance (mating systems, genetic drift, genetic bottle-
necks, and founder effects). For example, mating systems driven by female choice
can lead to males and females being phenotypically different. Such species are
referred to as being sexually dimorphic and the process leading to sexual dimor-
phism is termed sexual selection. Darwin recognized sexual selection is a special
case of natural selection in which males and females are under different selective
pressures. Because each of the different mechanisms for altering genetic compo-
sition can differ from population to population, each population of a species can be
on a different evolutionary pathway. If these pathways diverge sufficiently, the
original species can split into two new species.
As indicated by the fossil record, the rate and direction of evolutionary change
are highly varied by both over time and taxonomic group. This variability is driven
by multiple factors including changes in the range and scope of genetic variation,
shifts in the adaptive landscape, and shifts in selective pressures. Various com-
binations of these factors over geological time have produced a wide range of
evolutionary change that span from minor phenotypic adjustments to the apparent
sudden appearance of new phenotypes (or disappearance of existing phenotypes).
Because these factors are in constant flux, evolution is a continuous process and
populations never stop changing.
One of the complications to studying evolution is that evolutionary change is
not always evident. Although we define evolution as heritable changes in a gene
pool, not all genetic change is translated into phenotypic change. In addition, the
phenotypic consequences of genetic change are rarely predictable. Whereas some
genetic mutations will have little if any effect on the phenotype, others can have
profound effects depending, in part, on whether or not a gene codes for a functional
or non-functional protein, whether the protein is a structural or regulatory protein,
and the timing of protein expression.
Evolutionary change is limited or constrained by historical, developmental, and
logistical factors. From a historical perspective, evolution implies a continuity of
genetic information through time in the form of ancestor–descendant relationships.
This ancestor–descendant relationship constrains evolution (descendants are lim-
ited by their ancestors). However, this relationship also provides a test for evo-
lution in that character evolution should map to natural classifications, known as
phylogenies, which in turn should reveal evolutionary trends. A phylogeny is the
evolutionary history of an organism that shows its relationship to its ancestors and
related species. Examination of evolutionary trends indicates that character evo-
lution can be progressive or retrogressive and that generalized characters can
provide a template for the radiation of more specialized characters across different
groups of organisms.
Morphological evolution has strong ties to developmental programs. The spe-
cific area of biology that investigates this relationship is referred to as evo–devo
(Arthur 2002; Gilbert 2003). According to evo–devo, developmental programs
consist of a mosaic of interacting modules (Kuratani 2009; Breuker et al. 2006).
Collectively, these modules add up to a specific body plan with a specific set of
interdependent characteristics. The development of each module is influenced by
the development of each other module both spatially and temporally. Modularity
preserves the integrity and cohesiveness of each morphological unit while
allowing for adjustments throughout the course of development in response to the
interactions between the modules. When they are required for proper development,
the interactions between modules can conserve existing morphological expression,
thereby constraining the direction and magnitude of morphological change.
However, temporal alterations in the expression of modules (heterochrony) can
produce large, coordinated changes in morphology in a relatively short period of
time. Therefore, developmental modularity can either canalize morphological
change through spatial interactions or accelerate morphological change through
temporal changes in expression or duplication of modules.
From a logistical perspective, not all evolutionary change is feasible. Each
organism is a mosaic of interacting, interdependent characteristics. As such,
individual characteristics do not evolve in isolation, but evolve in concert with all
other characteristics. Whereas evolutionary modification to any particular char-
acter may have positive fitness effects on some characteristics, it may also have
negative fitness effects on other characteristics. In these situations, evolution must
settle for compromise rather than an optimized solution for each affected char-
acteristic. Compromise selection can be driven by additional factors including
seasonal variation, habitat variation, and ontogeny (development and life history).
Each of these factors may provide competing selective pressures that vary across
time, location, and age and thus further constrain the reach of evolution.
Despite the continuous march of evolutionary change and the numerous factors
that influence, drive, and alter its direction and timing, the process of evolution is not
restricted to greater and greater phenotypic divergence. It is not unusual for multiple
evolutionary pathways to independently arrive at a common solution. These com-
mon solutions may be shared by unrelated groups (convergent evolution) or related
groups (parallel evolution). Convergence not only demonstrates that there is more
than one pathway to the same end point, but also demonstrates the power of natural
selection to find a favorable solution from a variety of starting points.
12.3 Culture Clash
Evolution is the cornerstone of modern biological thought. It has propelled biology

from a mere description of nature to a predictive science. The products of evo-
lution are the adaptations. Adaptations are incorporated into the design of
organisms to solve problems and to provide functions involved with survival and
reproduction. In this case, nature has developed a technology through evolution.
Biomimetics unites the natural technology of biology with physical technology of
engineering. By incorporating both biology and engineering, the biomimetic
approach requires that evolution be fully considered.
Effective transfer of technologies from biology to engineering requires the
union of researchers working in these different fields. As simple as this may seem,
there are differences between the two fields and their approaches to problem
solving. The different approaches manifest themselves as cultural differences
between the disciplines. Biologists and engineers often work apart. Whereas
college biology curricula include courses in physics, engineering students are not
necessarily required to take biology courses. Engineers may have little knowledge
of biology beyond everyday experiences and evolution is often outside these
experiences. Similarly, biologists have little appreciation for engineering issues as
biological examples were not emphasized in physics classes and much of modern
biology is chemistry-oriented (e.g., genetics, molecular biology) rather than
physics-oriented (e.g., biomechanics).
A major difference between biologists and engineers is the concept of ‘‘design.’’
For the biologist, ‘‘design’’ refers to a description of the physical structure of
component or a whole organism in relation to the environment that it must interact.
In this sense, ‘‘design’’ implies only a functionally proficient arrangement of the

parts composing an organism, which are the result of natural selection (Vogel
1988). For example, the gazelle has a design for running swiftly through open
terrain, and the dolphin has a design for efficient movement in the ocean. The
engineer’s concept of ‘‘design’’ not only encompasses the structure of a system,
but also includes the process by which the system is conceptualized for a particular
function. In this sense, design is a human endeavor that implies anticipation and
purpose (Vogel 1988). The design process involves discovery, planning, devel-
opment, construction, evaluation, and invention (Vogel 1998). The difference in
how biologists and engineers view ‘‘design’’ leads to misunderstandings in the
interpretations of the origin of biological structure and function.
Without a combined understanding of biology and engineering, oversimplified
assumptions can retard the development of viable biomimetic applications. For
example, the development of the airplane, although inspired by birds, required
abandonment of a bird model. Airplanes do not flap their wings like birds to
simultaneously produce lift and thrust. Such a mechanism is impractical in modern
aircraft due to limitations from scaling and the high speeds necessary to remain
aloft by commercial and military jets. Lift generation at a size and speed scale that
is sufficiently large to carry a human requires steady-state aerodynamics rather
than the unsteady flapping of a bird wing (Jakalb 1990; Harris 1989). As a result,
the design of aircraft has advanced beyond the size and capabilities of birds for
level flight.
Harris (1989) argued that slavish adherence to the bird as a model system for
early airplanes held back design improvements through the early 1900s. Birds did
serve as the inspiration for flight and the early development of wing design (Li-
lienthal 1911; Jakalb 1990). However, large aircraft do not emulate the design of
the wings and their control for birds and other powered flyers, such as bats and
insects. Today, interest has focused on the agility of birds to perform complex
aerial maneuvers. Flying with the agility of birds is not presently possible as there
is an absence of detailed information on sensory input and control of the complex
mechanical linkages, which are associated with the motion of the wings. Indeed,
even the sensing and regulation of airflow over the wing and control surfaces is not
understood. In regard to maneuverability and agility, birds continue to demonstrate
superior performance to manufactured aircraft.
If biomimetic products are to be produced, the difference in cultures between
biologists and engineers needs to be recognized. Each group works with systems
that are parameterized in fundamentally different ways (Vogel 1998; Fish 2006).
The development of machines which have been the focus of engineered systems
is relatively large in size compared to their biological counterparts. Skyscrapers,
jumbo jets, and ships are of a scale that dwarfs termite mounds, birds, and whales,
respectively. Conversely, microscopic cells and viruses are smaller than machines.
These biological entities are not only capable of performing particular functions
but also have the capacity to change their programming for new or modified
functionality. The new approach of nanotechnology seeks to work at a level
smaller than the cell to produce biomolecular machines, and nanomaterials for
structural components and surface textures (Ummat et al. 2006; Zhang et al. 2006).
Differences in size between current engineered and biological structures affect the
forces experienced by these systems. While large vehicles traveling through a fluid
medium experience an environment that is dominated by inertial and gravitational
fluid forces, small organisms may be more affected by viscous forces. Water
striders (Gerridae) can cruise across a water surface supported by molecular
cohesion and surface tension (Hu et al. 2003), whereas ships operate at the water
surface by buoyancy from the mass of water displaced by the hull. Grappling with
such disparate criteria requires a designer to possess a working knowledge of both
the relevant biomechanics and the engineering issues associated with adapting the
biology to a machine for some targeted application.
The forces related to the scale of the systems dictate the construction materials
used. Engineered systems are composed of rigid materials. These materials include
metals, ceramics, and hard plastics. Even where compliance is required, hard
materials are used (e.g., spring steel). The choice of these rigid materials is a
matter of practicality for simple durability considerations as non-living systems do
not possess the capacity for self-repair. Biological systems are generally con-
structed from materials based on organic molecules. Metals are of limited avail-
ability as construction materials in biological systems. Iron, copper, and
magnesium are broadly used, but only in respiratory processes and only as one
atomic component of a significantly larger molecule. Iron and copper are part of
the molecules hemoglobin and hemocyanin. These molecules are essential to
capture oxygen for transport through the bodies of animals. Magnesium is part of
the basic structure of chlorophyll and functions to absorb radiant energy for plants.
Magnetite (Fe3O4) is found capping the small teeth of the grinding radula in
mollusks (Gordon and Joester 2011). A class of organic compounds, called sid-
erophores, is used by marine bacteria to complex with iron and collect the metal
from an environment, where this material is limited (Martinez et al. 2003).
Proteins, carbohydrates, and lipids form structures and components of simple
cells to complex plants and animals (Wainwright et al. 1976; Vincent 1990). These
materials are created and often function in an environment that is wet. Although
biological skeletons can be formed from ceramics (e.g., vertebrate bones, mollusk
shell, sponge spicules), these structures are composites with varying amounts of
organic molecules (Vogel 1988). Generally, organisms are primarily composed of
biological materials that are compliant (Wainwright et al. 1976; Vogel 1988).
These compliant structures are part of a design that permits the body to bend
(Wainwright 1988). Where motion is restricted due to a rigid skeletal framework,
the compliant materials permit flexibility at joints. Furthermore, compliant
materials allow for the storage of elastic energy when stressed. Release of the
elastic tension can be used for energy recycling in repetitive motion. The springing
ligament in the horse’s leg aids in reaccelerating the hoof when running. The
Achilles tendon in the kangaroo is stretched when hopping and recycles enough
energy in each step to maintain a nearly constant metabolic effort as speed
increases (Alexander 1988).
Engineered systems generally use rotational motors. The energy to power these
devices is derived from radiant energy, chemical transformation, or thermal and
nuclear sources. The energy for movement and manufacturing in biology is
restricted to chemical catabolism, although ultimately radiant energy from the sun
is used to produce the chemical compounds. High-energy chemical bonds are
broken to release energy. This energy is then transferred to intermediate com-
pounds (e.g., ATP, creatine phosphate), which can be transported around the cell.
With the exception of bacteria, rotational motors have not evolved in biological
systems (Fig. 12.2). Movement in plants is powered by growth and fluid pressure.
Animals utilize translational movements that are activated by a chemical motor,
that is, muscle. Muscles contract while exerting a force to do work. As muscles
cannot lengthen on their own, muscles are typically arranged as antagonistic pairs.
As one muscle contracts, the other is lengthened either passively or while exerting
a force. This results in oscillatory motions of the body of an animal or its
appendages, but not rotary motion. Some energy is lost by cyclical accelerations.
Complex neural networks with multiple sensory inputs control animal systems,
whereas engineered machines are controlled by simple computational systems
with limited sensory feedback. A human brain of approximately 1.5 kg is com-
posed of 100 billion cells with one quadrillion synaptic connections (Denny and
McFadzean 2011). The large number of nerve cells (neurons) and neural con-
nections is associated with a large range of behavioral responses. Whereas
autonomous machines must be pre-programmed to produce an appropriate
response to a particular known stimulus, animals can be plastic in their response.
Biological organisms are functionally multifaceted (i.e., they move, feed,
remove wastes, and reproduce) and must compromise optimal solutions for spe-
cialized functions to perform adequately rather than maximally (Katz and Jordan
1997; Webb 1997). A machine is constructed with a particular and defined purpose
or a mission that it was designed to fulfill (Denny and McFadzean 2011). Having
an engineered system with a single purpose increases the maximal efficiency of the
targeted operation. As biological organisms are multitasking, they must balance
any one function with a number of other functions that compete for energetic
resources within the body, but are necessary to maintain life (Fish 2006). Ulti-
mately, all biological organisms are driven by three primary motivations of
obtaining sustenance (i.e., food and water), security (i.e., avoid being killed, self-
preservation), and sex (i.e., reproduction) (Denny and McFadzean 2011). These
are criteria that are seldom included in engineering schematics.
Perhaps the greatest difference between biological and engineered systems is
that biological systems are capable of reproduction. Mechanical systems can be
manufactured in large lots with strict control for exact duplication. However, for
organisms, the ability to reproduce is not confined to merely making copies of
individual units. Indeed, new mutations and recombinations of genetic material
increase variation, which is the raw material of evolution. Reproduction allows for
changes that can lead to new evolutionary solutions or improvements in functional
efficiency. Machines cannot autonomously replicate themselves. Improvements in
Fig. 12.2 Improbable and real fish propulsion systems. A mechanical fish model with a
rotational screw propeller (top), based on the art of Dr. Seuss on display at Universal Orlando.
Lateral view of a fish exhibiting fins for propulsion and stability (middle). Dorsal view of a shark
swimming by undulations of the body and caudal fin (bottom)
machine design and function only come about from tinkering with established
designs or scrapping the old design for a completely new approach.
Internal reproduction of the cells of the body affords organisms the ability to
grow and self-repair. Growth permits change in a body. This change in size may
afford the reproducing adult the capability of maximizing the production of young
by increasing the space available for storage of gametes or housing and protecting
the young prior to parturition. Over the course of growth and maturation of the
young, their small initial size will allow them to feed on different foods and occupy
different microenvironments, which fully mature individuals would not be able to
exploit. The assemblage of different species at varying stages of development and
size would enhance the stability of a given ecosystem.
The ability of organisms to self-repair is not merely associated with wound
healing. It is also associated with immunity in fighting disease, invasion by foreign
bodies, and with coagulation in limiting damage to the body. Organisms can
prevent damage to delicate tissue by generating new cells to replace those cells,
which have been damaged or abraded away. Another mechanism is to coat an
irritant, rendering it harmless. For example, mollusks such as bivalves (e.g., clams,
oysters) can secrete a smooth nacre material around a grain of sand to produce a
pearl. The pearl prevents abrasion to the soft tissue (mantle) responsible for
secreting the animal’s shell. Literally, sand in the gears of a machine would
generate enough friction to stop operation. Filters and lubrication must be designed
into machines to reduce contaminants from wear or external sources.
Engineered systems are built with an assessment of potential for failure and
with a plan for maintenance that is based on failure of prototypes. In the devel-
opment of an engineered design, such considerations are known as failure criteria
and dictate the physical limits for the form, function and materials used (Petroski
1996). Autonomic healing is only recently being attempted with battery technol-
ogy (Blaiszik et al. 2012).
Besides the inherent differences between biological and engineered systems, a
cultural disparity in the way engineers and biologists view the data in their
respective disciplines. Engineers strive to limit the number of variables of any
mechanical system, especially in the design of structures or devices with targeted
functionality. Biologists consider a large number of variables. Furthermore,
engineers analyze all the errors associated with a system in an attempt to control
and reduce variation. Biologists study variation for each of the parameters that
control a system. Indeed, variation in biological systems is the foundation of the
evolutionary process, which is at the core of modern biological thinking.
12.4 MantaBot: Example of Biomimetics
The biomimetic approach demands careful observation of the whole biological

system to identify the principles and attributes of the system that are appropriate to
the function that is to be emulated. Thus, major limitations and constraints of any
biological design must be defined before translation to an engineered system. As
biological systems are the product of evolutionary mechanisms with their limita-
tions, it is possible to improve on the design where the biology is constrained.
In the area of autonomous underwater vehicles (AUV), there is a need for
AUVs that can be deployed quickly and can be adapted for a variety of missions
(e.g., surveillance, search and rescue, sentry duty, logistics support, and chemical
or biological agent detection) (Bandyopadhyay 2005; Colgate and Lynch 2004;

Fish et al. 2003, 2012; Low 2011; Moored et al. 2011a). Parameters that are
important for the next generation of AUV include efficiency, maneuverability,
stability in high-energy environments, operation in the littoral zone and open
ocean, station holding, and ability to follow bottom terrain. Design considerations
include a rigid hull and minimum of control and propulsive surfaces. The animal
kingdom has a number of organisms that meet these qualifications and provide a
viable solution. Because of these performance and design characteristics, partic-
ular attention has focused on the development of a biomimetic autonomous
undersea vehicle (BAUV) fin propulsor that mimics the biological principles and
kinematics of the myliobatoid rays with particular interest in the manta (Manta
birostris, Order Myliobatiformes, Family Mobulidae). The biological role of the
manta and other myliobatoid rays conforms to the design space of a BAUV.
Batoid fishes (skates and stingrays) represent a group of over 500 elasmobranch
species that have evolved dorsoventrally flattened bodies with reduced or whip-
like tails and expanded pectoral fins that are fused to the head to form a broad flat
disk (Rosenberger 2001; Douady et al. 2003). This deviation from the typical
torpedo-shaped bodies of bony fish and sharks is an adaptation to living on the
ocean bottom. The batoids swim solely by movements of their two greatly
expanded pectoral fins (Breder 1926; Klausewitz 1964; Heine 1992; Rosenberger
2001). The pectoral fins have triangular, wing-like planforms with an aspect ratio
(the ratio of span to chord) of 3.5 (Fish et al. 2012). The cross-sectional geometry
of myliobatoid has a streamlined appearance. The lateral pectoral fins display
symmetrical cross-sectional profiles reminiscent of engineered foils (Abbott and
von Doenhoff 1959).
The manta ray (Manta birostris) is phylogenetically one of the most highly
derived species of batoid fishes. Manta rays inhabit tropical seas of the world.
They are adapted to a pelagic life in the open ocean, but may live inshore, and are
found along reef fringes near deep water (Deacon et al. 1997). Mantas are filter
feeders, preying on crustaceans and small fish. Other rays feed on benthic prey.
The manta and other highly derived pelagic rays (e.g., Myliobatis, Rhinoptera,
Aetobatus, Manta) swim by oscillatory locomotion (mobuliform mode). This
mode of swimming consists of a small undulatory component (the wavelength of
the undulation is greater than the chord length of the pectoral fin), and the pectoral
fins are flapped dorsoventrally, analogous to the flight of birds (Breder 1926, 1964;
Klausewitz Heine 1992; Rosenberger 2001).
The geometry of the fins and their kinematics for thrust production indicate a
high-efficiency propulsive system. The streamlined shape of the body and fins of
myliobatoids indicates a low drag profile (Fig. 12.3; Webb 1975; Vogel 1994; Fish
and Lauder 2006). Both spanwise and chordwise flexibility are apparent as the
pectoral fins are oscillated (Fig. 12.3; Klausewitz 1964; Heine 1992; Rosenberger
2001; Schaefer and Summers 2005; Fish et al. 2012). Spanwise and chordwise
flexibility are associated with enhancing propulsive efficiency and thrust produc-
tion. Spanwise flexibility prevents the total loss of thrust at the reversal of an
oscillatory stroke (Liu and Bose 1997). Chordwise flexibility at the trailing edge of
Fig. 12.3 Flexibility of the propulsive fins of batoids (top, manta; bottom, cownose ray)
the fin could increase the efficiency by up to 20 % with only a moderate decrease
in the overall thrust (Katz and Weihs 1978, 1979; Bose and Lien 1989; Bose 1995;
Prempraneerach et al. 2003). Actively swimming rays with flexible fins may have
higher propulsive efficiencies compared to values predicted for models of rigid
lifting surfaces. Hydrodynamic computations performed by Heine (1992) showed
efficiencies of 0.7–0.9 for swimming rays. Efficiencies in this range are considered
high, because few engineered propellers achieve efficiencies higher than 0.7
(Larrabee 1980; Liu and Bose 1993). In addition, oscillating biological hydrofoils
with flexibility maintain high efficiency over an extended operational range (Fish
and Lauder 2006). Standard fixed-pitch marine propellers have a maximum pro-
pulsive efficiency in only a very narrow range of operational speeds. Thus,
oscillating hydrofoil propulsion as demonstrated by batoids is in keeping with the
requirements for a BAUV.
The size of myliobatoid rays is appropriate as a model system for a BAUV. The
manta is reported to be over 6 m wide and weigh over 1,580 kg, although other
related species, such as the cownose ray (Rhinoptera bonasus), are smaller (Deacon
et al. 1997; Compagno 1999). Scaling issues are inconsequential as the size and
speed of these rays correspond to design and operation of AUVs. Thus, the hydro-
dynamics of the biological and engineered systems are equivalent as defined by
important variables, the Reynolds number, Re, and the Strouhal number, St.
The Reynolds number represents the ratio of inertial forces to frictional forces
and defines the flow conditions (laminar or turbulent) around a submerged body. A
manta ray swimming at one body length/s would have a value of Re over three
million, and data for the cownose ray from Heine (1992) indicate a Re of
approximately 90,000. These high values indicate that inertial forces dominate.
Propulsion is provided by production of a caudally directed momentum jet of
fluid. The momentum jet is generated by oscillatory motions of the enlarged
pectoral fins (Fig. 12.4), leaving a wake of staggered alternately rotating vortices.
The rate at which these vortices are shed and the efficiency of the propulsive
movements are related to St. St is the ratio of inertial forces from local acceleration
to inertial forces from convective acceleration and represents the degree of
unsteadiness in the flow. St is the product of the propulsive frequency and vertical
excursion of the fin divided by the velocity of the animal (Rohr and Fish 2004).
The cownose ray has St = 0.19 (Heine 1992). This value is close to the Strouhal
number range of 0.2–0.4, which is stated to be where the propulsive efficiency is
maximal (Triantafyllou et al. 1993, 2000; Triantafyllou and Triantafyllou 1995;
Rohr and Fish 2004).
There have been a number of attempts to develop a bio-inspired batoid AUV
(Moored et al. 2011a, b). Robots have been constructed to mimic the oscillatory
swimming of rays by motors (Gao et al. 2007; Yang et al. 2009; Zhou and Low 2010),
Fig. 12.4 Bioinspired robots

(MantaBots) based on the
design and swimming
kinematics of rays. The
mechanical rays were built by
Princeton University (top)
and the University of Virginia
(bottom)
Fig. 12.5 Kinematic traces of the body and propulsive pectoral fins of a swimming manta (top)
and MantaBot (bottom). The large oscillations are from the pectoral fins, and the small
oscillations are indicated for the anterior end of the manta and the MantaBot. The manta shows
asymmetrical movements of the fin strokes at lower frequency than for the MantaBot. Vertical
motions of the anterior end were relatively greater for the MantaBot than the manta
pneumatic pectoral fins (Brower 2006; Suzumori et al. 2007; Cai et al. 2010), fluidic
muscles (Festo 2008), and ionic polymer–metal composites (Takagi et al. 2006).
Recently, two prototypes (MantaBots) were developed and tested (Fig. 12.5;
Pennisi 2011). The MantaBots were based on biologically derived data, which
detailed the design geometry and kinematics of living batoids. Each MantaBot
used different mechanisms to actuate the pectoral fins. In one version produced at
Princeton University, four metal rods were actuated with a servomotor. The rods
heaved a flexible plastic fin, producing varying degrees of spanwise and chordwise
curvature (Moored et al. 2011b; Pennisi 2011). A second MantaBot was con-
structed at the University of Virginia and used a mobile tensegrity structure to
move a propulsive elastomer fin (Pennisi 2011). The tensegrity structures consisted
of truss-like structures, which acted like a skeleton tendon internal framework
(Moored et al. 2011a, b; Fish et al. 2012). The rigid elements of the tensegrity
structure were articulated. Two cable elements could generate tension to give
integrity to the structure to support large loads and actuate movement.
While both MantaBots were capable of swimming and could perform elaborate
maneuvers (Pennisi 2011), the propulsive movements of the pectoral fins were
similar but not an exact duplication of a batoid (Fig. 12.4). The kinematics of the
MantaBot pectoral fins was more symmetrical on the up- and down-strokes with
respect to the longitudinal axis of the body than for live batoids. Living rays can
swim faster and are more agile and maneuverable than the MantaBots (Parson et al.
2011). The MantaBots swam at just over 0.5 body lengths/s, and manta rays can
swim at 0.8 body lengths/s. The amplitude of heave at the rostrum was 27 % of the
fin tip amplitude of the MantaBot, whereas the heave amplitude was 12 % for
manta (Fig. 12.4). St were 0.6 and 1.6 for the Princeton University and
University of Virginia MantaBots, respectively. Such values were outside the range
for maximum efficiency and above the value of St for the cownose ray (Heine 1992;
Triantafyllou et al. 1993, 2000; Triantafyllou and Triantafyllou 1995).
For the MantaBots, engineering has derived separate solutions that provide
similar motions, but are still limited compared to the actual rays. The skeletal,
neural, sensory, and muscular systems of living organisms are still more complex
than robotic systems. This complexity permits organisms to work within a large set
of behavioral and performance responses. As further refinement to robotic systems
is built into the design, emulation of biological systems will converge with the
engineering. However, there are limitations to biological systems from evolu-
tionary constraints that may reduce performance. Emulation of biological systems
should only be taken to a point in which performance is maximized, but then can
be enhanced through engineering of a bio-inspired design.
12.5 Evolutionary Constraints
Caution must be exercised when using evolution as a justification for the devel-
opment of biomimetic products. Nature has served as inspiration for various
devices for centuries, but the theory of evolution, as currently understood, has only
been around since the mid-1800s. With the arrival of biomimetics and bio-inspired
design as fields that could produce products and increase funding and investment,
there has been an attempt to validate these fields with a natural justification.
Evolution is seen as an iterative process that arrives at the best design. Therefore,
the evolutionary process becomes imperative to highlight the advantages of natural
designs. However, evolution does not hold all the answers or solutions to problems
that affect our daily lives.
Evolution is a process that over the history of the Earth has produced a mul-
titude of differing species that have developed solutions (adaptations) to local
environments. In some cases, evolution has lead to different but analogous solu-
tions to similar environments in different phylogenetic groupings. Consider
swimming by a fish and a squid. The fish wags (undulates) its caudal fin from side
to side to accelerate a mass of water into its wake, and thereby gain forward
momentum (Fig. 12.2). The squid, however, uses jet propulsion to push the animal
backward. Both animals can move rapidly through the water, but the mechanics of
the propulsive systems differs. Flexibility of the vertebral column of the fish
permits an undulatory wave to move posteriorly down the body to laterally flex the
caudal fin. Because the body of the squid is held rigid by a non-flexing internal
skeleton (i.e., pen), such movements are not possible in the squid, necessitating
jetting. These differing solutions to the problem of rapid movement through water
have further consequences. The fish can push a large volume of water at slow
speed, thereby moving with a high efficiency, whereas the squid pulses a jet of
fluid at a much higher speed to gain an equivalent momentum, resulting in a lower
propulsive efficiency (O’Dor and Webber 1986).
For a non-biologist using evolution to justify biomimicry, the greatest mistake
is inferring that evolution has an optimal design goal. Evolution is neither con-
scious nor predictive. Evolution is not visionary. Construction and organization of
an organism’s features have not evolved toward a specific goal, and evolution does
not drive toward perfection. Evolutionary change by natural selection does not
provide a ‘‘perfecting’’ principle, only a ‘‘better than’’ principle (Luria et al. 1981).
Organisms evolve features not to be optimal or perfect, but merely to perform
adequately (Katz and Jordan 1997; Webb 1997).
Genetic algorithms are used to mimic the process of evolution by natural
selection. Genetic algorithms start with a population of randomly generated
individuals [i.e., primordial ooze of hundreds or thousands of computer programs;
Koza (1994)]. The individuals are evaluated iteratively by a fitness function, which
represents a predetermined solution set. The solution set is encoded as a finite-
length string that is composed of elements with a finite number of possibilities
(Whitley 1994; Barrett 2002). By iteration and selection, a singular, predetermined
optimal solution can be produced. However, biology is not driven by a predeter-
mined optimal solution. The evolution of living organisms is shaped by the
interactions of an organism with its environment, its phylogenetic history, and the
genetic mechanisms that promote genetic diversity (e.g., mutation, recombina-
tion). Furthermore, evolution can work in the opposite direction from genetic
algorithms to produce a diversity of solutions from a singular ancestral type. Each
solution or species divides up the environment into various ecological niches.
Biological organisms are multitasking entities. An organism is a mosaic of
integrated structures and functions to achieve evolutionary success (i.e., survive
and reproduce). Some of these components may be at odds with other features of
an organism. As a result, organisms must compromise optimal solutions for the
necessity of having an integrated system that can perform a number of simulta-
neous functions. The integrated parts of an organism must share the limited
metabolic energy available for maintenance and function. Increased allocation of
limited resources to one component of a body may improve function, but be to the
detriment of another component. Natural selection acts on an entire organism and
not its individual parts (Luria et al. 1981). Despite the contrary argument (Bar-
Cohen 2006; Allen 2010), evolution rarely leads to solutions with a maximal
performance and with an economy of resources.
An example of how optimal design is lacking in biological organisms can be
found in the structure and performance of the propulsive mechanics of fishes. Fish
carry with them a large amount of muscle. Most of the muscle mass is not used
during routine swimming, such as when a fish is cruising or migrating. Routine
swimming is accomplished using muscle composed of slow oxidative (red) fibers
(Alexander and Goldspink 1977). These fibers use aerobic means of generating
energy for muscle contraction and are highly efficient. As long as oxygen is
available, red fibers can repeatedly contract over an extended period of time, but at
a relatively slow rate. However, the bulk of the propulsive muscle mass is com-
posed of fast glycolytic (white) fibers. These fibers contract faster and more
powerfully than red fibers, but cannot sustain repeated contractions over a long
time and have a very low efficiency. For fish that migrate long distances at low-to-
moderate speeds, it makes more sense for these fish to have a muscle mass
composed of a higher proportion of red fibers than white fibers. Indeed, carrying a
large mass of white muscle seems detrimental. The inactive white fibers continue
to metabolize nutrients, and the extra mass encumbers additional energy costs to
move the body. Why then is so much of the musculature composed of white fibers?
While not utilized all the time, the white fibers are advantageous during those brief
instances when life and death are on the line. When a fish has to chase prey for
food or to escape being preyed itself, it is advantageous to have a large mass of
white fibers to generate the forces required to accelerate quickly. At these times,
efficiency is not as important as rapid acceleration to close the distance on prey or
increase the distance away from a predator. In addition, the fish must carry other
organs that add to its mass, thereby impeding performance. Carrying large gonads
and reproductive products can reduce swimming speed and survival, although
reproductive organs are necessary for evolutionary fitness.
Not all possible structures and processes are available to organisms. As men-
tioned previously, biological organisms do not use metal as a framework for a
physical support. The structures of organisms can only work within the constraints
of materials based on organic molecules. The formation of these molecules is
directed through recipes encoded on the DNA molecule and manufactured by
living cells. Metabolic processes are only possible within a narrow range of
temperatures for the formation of complex molecules. Above critical temperatures,
proteins denature and cellular systems fail. Alternatively, synthetic manufacturing
systems can use high temperatures and pressures to produce new molecular con-
figurations or meld materials together in construction.
The wheel may be considered one of the greatest inventions by humans. It
allows for a reduction in the energy cost for movement by reducing friction and
eliminating oscillatory motions. The wheel is free to rotate continuously around an
unattached central axis. Although the wheel is ubiquitous in engineered
mechanical systems, it is rare in natural systems. With the exception of bacteria,
multicellular organisms have not evolved wheels. Rotation movements translated
over 360° are found in whole body rolling maneuvers by organisms, such as
tumbleweeds, caterpillars, stomatopods, and desert spiders (Full et al. 1993;
Armour and Vincent 2006). The energy for rolling is derived from air and water
currents or from gravity. Self-actuating movements by animals are powered by
muscles, which only allow a limited degree of rotation. A physical connection is
made by muscles and stabilizing ligaments across a rotational joint. It is this
construction of rotational joints that makes the evolution of wheels in animals
impossible. As continuous rotation is not possible in the musculoskeletal system of
animals, locomotion requires oscillatory movements. These oscillations are char-

acterized by reversals in the direction of body and appendage movements. The
energetics is impacted by operating in such a fashion. Periodic accelerations occur
over a propulsive cycle. Thus, kinetic energy varies greatly and the efficiency is
decreased (Alexander 1983).
Another constraint imposed by evolution is that the environment is in a non-
equilibrium state, which places design criteria in a state of constant flux. Organ-
isms must be plastic enough to deal with the constant changes in the environment.
For any geographic location, the physical parameters of temperature, pressure,
humidity will vary with time, season of the year, and across geological eons.
Additional factors include the availability of food and shelter, distribution of
predators, and prevalence of diseases. Changes in these factors can be unpre-
dictable. This state of constant flux means that no design ever lasts indefinitely
(Van Valen 1973). What is good today may not be good tomorrow. Furthermore,
what is observed today may be on the way out. All species eventually go extinct.
The dinosaurs represent a successful group of medium to large reptiles that
were the dominant land vertebrates on the planet for approximately 150 million
years. However, in a geologically short period of time, the entire lineage went
extinct. The only remnant of the dinosaurs is the offshoot that gave rise to the
birds. Although dinosaurs might have evolved adaptations that solved a number of
problems, their total extinction was due to a sudden catastrophic environmental
change that overwhelmed the genetic capabilities of these animals to evolve to
meet the ecological insult. The environmental changes that occurred afterward
may have made it impossible for dinosaurs to reestablish themselves. Furthermore,
the mammals usurped the ecological niches formerly occupied by the dinosaurs or
formed new niches.
Organisms themselves can change the very environment in which they live, and
not always for the better. Expectations of harmonious living between individuals
and species (Benyus 1997) may be merely wishful thinking. Whereas harmonious
symbiotic relationships have occurred such as in the origin of eukaryotic cells
(Margulis and Sagan 1997) and organisms like lichens and corals, competition is
fierce between individuals and species. Natural selection works at the level of the
individual. Because there is an inherently selfish interest, life does not work for the
best interests of the community. Adages such as ‘‘kill or be killed’’ and ‘‘Nature,
red in tooth and claw’’ have an element of truism to them. Individual organisms
can degrade the physical environment to their own benefit and the detriment of
others. Certain plants will release secondary compounds that negatively affect the
growth of neighboring plants. This strategy is referred to as allelopathy. The black
walnut (Juglans nigra) releases a phenolic compound from its leaves, stems,
branches, and roots. The toxic effect extends up to 27 m from the trunk of the tree
(Lambers et al. 1998). Even the beaver (Castor canadensis) will modify its
environment, so that various species are negatively impacted. By damming small
streams, the creation of a pond will flood areas to the detriment of various ter-
restrial plants and animals and organisms that inhabit faster flowing water.
Lastly, in regard to the constraints imposed on evolutionary systems, design is

constrained by evolutionary history. Species are merely terminal points on bran-
ches of an evolutionary bush. Organisms evolved along lines of common descent
with shared genetically regulated developmental patterns, confining the design
space. For example, the insulation of birds is fixed to a morphology that uses
feathers, whereas mammals utilize fur. Both hair and feathers function similarly
entrap an insulative air layer against the skin and maintain an elevated body
temperature. These structures are derived from the keratinized scales of reptilian
ancestors, although from different parts of the scales and with different molecular
conformations (Vincent 1990). Once taking separate evolutionary pathways,
species are limited with respect to developmental options. Although originally
used for insulation, the feathers became co-opted for flight in the structure of the
wings. Mammals cannot grow feathers. Therefore, the evolution of flight in
mammals required a wing constructed of a flexible skin membrane, rather than a
feathered wing. Whereas feathers have sufficient strength to maintain the wing
surface and deal with aerodynamic forces generated by the flapping wing, the skin
membrane in bats must be reinforced by elongation of the bony digits.
Phylogenetic history can impede performance by a species. Most fish breath
solely using gills, which allows them to remain submerged and away from the
surface of the water. This means that the drag on a swimming fish can be mini-
mized. Moving in the proximity of the water surface incurs additional drag due to
the formation of waves (Fish et al. 1991). Dolphins and whales (i.e., cetaceans)
secondarily returned to the sea. These marine mammals evolved a body design
analogous to fish for reduced drag (Fish and Hui 1991). However, dolphins and
whales are restricted to the mammalian body plan, which uses lungs for gaseous
respiratory exchange. Although it would be more efficient to possess gills and
remain submerged, cetaceans must come to the surface to breathe. As the addi-
tional drag at the surface can be as high as five times the submerged drag (Fish and
Hui 1991), the phylogenetic history of cetaceans can impose severe energetic
penalties. Cetaceans have had to develop behavioral and physiological adaptations,
such as porpoising and prolonged apnea, to avoid surface effects.
12.6 Final Comments
It is tantalizing to consider the development of new and superior technological

designs for enhanced performance based on biological systems. However, such
innovations have been elusive (Fish 1998, 2006). The commercial production of
biomimetic products has been rare (Vogel 1998). There has been greater success
from bio-inspiration. Strict adherence to biological designs in biomimicry rarely
produces any practical results and in some cases can impede the development of
engineered systems (Vogel 1994, 1998; Fish 1998). Yet, the fields of biomimetics
and bio-inspired design are generating excitement. Biology can provide fresh
solutions to more conventional approaches and even suggest new avenues of

research. For engineering, the designs, structures, and materials of biology rep-
resent potentially untapped resources. Entrepreneurs, venture capitalists, and
corporations speculate on the future of new designs (Petroski 1992; Vogel 1998).
Each is hedging on getting a lead in the marketplace. The number of biomimetic-
based patents has increased faster than the total of all US patents between 1985
and 2005 (Bonser 2006). ‘‘Biomimicry,’’ ‘‘biomimetics,’’ and ‘‘bio-inspired
design’’ have become terms that draw attention to an emerging field of study and a
source of investment capital. However, if these terms are not to be merely buzz-
words that focus attention without comprehension, then clearer definitions and a
conceptual biological framework are required.
As this chapter has emphasized, it is necessary to understand evolution with its
inherent limitations to all possible designs. The technology that nature has evolved
is not always ahead of the technology of human ingenuity (Vogel 1998). Only by
understanding evolution and how organisms have adapted to their present and past
environments can one avoid the pitfalls of overstatement regarding biomimicry.
Besides an understanding of evolutionary mechanisms, practitioners of biomim-
icry must be aware of the limitations of biology for transition to engineered
systems, and differences in the culture of biologists and engineers.
How can the process of bio-inspired design be accomplished to live up to the
expectations for the development of new products? To address this question, one
must first ask the question, ‘‘What needs to be improved?’’ Engineering is a goal-
directed, applied science. The technological problem must be outlined to deter-
mine the direction of design for improvement in functionality. Next, biological
structures or processes need to be identified that perform better than the current
technology or solve the problem from a direction different from the current design
path. The input of a biologist, who has knowledge of the workings of the system, is
requisite. Simplified explanations of biological phenomenon from the popular
literature and general media are inadequate to base development of new designs.
This step is followed by research on the mechanism of action by the biological
structure or process. In this analysis, considerations of the limitations of the
biology are necessary. Any limitations due to energy efficiency, construction of
materials and morphology must be evaluated with respect to a cost-benefit anal-
ysis. For example, engineered systems can be economically constructed as the
performance of the materials used and the forces to be encountered are highly
predictable. Alternatively, biological systems require increased safety factors due
to the unpredictability of the environment and variation in the structural design of
the organism (Alexander 1998), where the ‘‘safety factor’’ is the ratio of a com-
ponent’s strength or performance to the maximum expected load during operation
(Diamond 1998). The cost to introduce the biological advancement or replicate it
may be uneconomical. Finally, the biomimetic approach requires a coordinated
effort of biologists, engineers, industrial designers, and business people to produce
future bio-inspired products.
Acknowledgments We would like to thank Dr. Anthony Nicastro and Janet Fontanella for their
comments on the manuscript. We also great appreciate the cooperation of Drs. Hilary Bart-Smith,
Keith Moored, Hossein Haj-Hariri, Tetsuya Iwasaki, and Alexander Smits on the robotic manta
project. This chapter is based in part on research performed with support from the Office of Naval
Research grant no. N000140810642 to FEF.
References
Abbott IH, von Doenhoff AE (1959) Theory of wing sections. Dover, New York
Alexander R McN (1983) Animal mechanics. Blackwell Scientific Publications, Oxford
Alexander R McN (1988) Elastic mechanisms in animal movement. Cambridge University Press,
Cambridge
Alexander R McN (1998) Symmorphosis and safety factors. In: Weibel ER, Taylor CR, Bolis L
(eds) Principles of animal design: the optimization and symmorphosis, Cambridge University
Press, Cambridge, pp 28–35
Alexander RMN, Goldspink G (1977) Mechanics and energetics of animal locomotion. Chapman
& Hall, London
Allen R (2010) Bulletproof feathers. University of Chicago Press, Chicago
Armour RH, Vincent JFV (2006) Rolling in nature and robotics: a review. J Bionic Eng
3:195–208
Arthur W (2002) The emerging conceptual framework of evolutionary developmental biology.
Nature 415:757–764
Bandyopadhyay PR (2005) Trends in biorobotic autonomous undersea vehicles. IEEE J Ocean
Eng 29:1–32
Bar-Cohen Y (2006) Biomimetics: biologically inspired technology. CRC, Boca Raton
Bar-Cohen Y (2012) Biomimetics: nature-based innovation. CRC, Boca Raton
Barrett DS (2002) Optimization of swimming locomotion by genetic algorithm. In: Ayers J,
Davis JL, Rudolph A (eds) Neurotechnology for biomimetic robots. MIT Press, Cambridge,
pp 207–221
Benyus JM (1997) Biomimicry: innovation inspired by nature. Perennial, New York
Blaiszik BJ, Kramer SLB, Grady ME, McIlroy DA, Moore JS, Sottos NR, White SR (2012)
Autonomic restoration of electrical conductivity. Adv Mater 24:398–401
Bonser RHC (2006) Patented biologically-inspired technological innovations: a twenty year
view. J Bionic Eng 3:039–041
Bose N (1995) Performance of chordwise flexible oscillating propulsors using a time-domain
panel method. Int Shipbuild Progr 42:281–294
Bose N, Lien J (1989) Propulsion of a fin whale (Balaenoptera physalus): why the fin whale is a
fast swimmer. Proc Roy Soc Lond B 237:175–200
Breder CM Jr (1926) The locomotion of fishes. Zoologica 4:159–297
Breuker CJ, Debat V, Klingenber CP (2006) Functional evo-devo. Trends Ecol Evol 21:488–492
Brower TPL (2006) Design of a manta ray inspired underwater propulsive mechanism for long
range, low power operation. Master’s thesis, Tufts University
Cai Y, Bi S, Zheng L (2010) Design and experiments of a robotic fish imitating cow-nosed ray.
J Bionic Eng 7:120–126
Colgate JE, Lynch KM (2004) Mechanics and control of swimming: a review. IEEE J Ocean Eng
29:660–673
Compagno LJV (1999) Systematics and body form. In: Hamlette WC (ed) Sharks, skates, and
rays: the biology of elasmobranch fishes. The John Hopkins University Press, Baltimore,
pp 1–42
Darwin C (1859) On the origin of species by means of natural selection. John Murray, London
Deacon K, Last P, McCosker JE, Taylor L, Tricas TC, Walker TI (1997) Sharks and rays. Fog
City Press, San Francisco
Denny M, McFadzean A (2011) Engineering animals: how life works. Belknap Press, Cambridge
Diamond JM (1998) Evolution of biological safety factors: a cost/benefit analysis. In: Weibel ER,
Taylor CR, Bolis L (eds) Principles of animal design: the optimization and symmorphosis.
Cambridge University Press, Cambridge, pp 21–27
Douady CJ, Dosay M, Shivji MS, Stanhope MJ (2003) Molecular phylogenetic evidence refuting
the hypothesis of Batoidea (rays and skates) as derived sharks. Mol Phylogenet Evol
26:215–221
Festo M (2008) Fluidic muscle brochure. http://www/festo.com
Fish FE (1998) Imaginative solutions by marine organisms for drag reduction. In: Meng JCS (ed)
Proceedings of the international symposium on seawater drag reduction, Newport, Rhode
Island, pp 443–450
Fish FE (2006) Limits of nature and advances of technology in marine systems: what does
biomimetics have to offer to aquatic robots? Appl Bionics Biomech 3:49–60
Fish FE, Blood BR, Clark BD (1991) Hydrodynamics of the feet of fish-catching bats: influence
of the water surface on drag and morphological design. J Exp Zool 258:164–173
Fish FE, Hui CA (1991) Dolphin swimming—a review. Mamm Rev 21:181–195
Fish FE, Lauder GV, Mittal R, Techet AH, Triantafyllou MS, Walker JA, Webb PW (2003)
Conceptual design for the construction of a biorobotic AUV based on biological
hydrodynamics. In: Proceedings of the 13th international symposium on unmanned untethered
submersible technology. Autonomous Undersea Systems Institute, Durham New Hampshire
Fish FE, Lauder GV (2006) Passive and active flow control by swimming fishes and mammals.
Ann Rev Fluid Mech 38:193–224
Fish FE, Smits AJ, Haj-Hariri H, Bart-Smith H, Iwasaki T (2012) Biomimetic swimmer inspired
by the manta ray. In: Bar-Cohen Y (ed) Biomimetics: nature-based innovation. CRC, Boca
Raton, pp 495–523
Full R, Earls K, Wong M, Caldwell R (1993) Locomotion like a wheel? Nature 365:495
Gao J, Bi S, Xu Y, Liu C (2007) Development and design of a robotic manta ray featuring flexible
pectoral fins. Robotics and Biomimetics, ROBIO 2007. IEEE International Conference,
pp 19–523
Gilbert SF (2003) The morphogenesis of evolutionary developmental biology. Int J Dev Biol
47:467–477
Gordon LM, Joestner D (2011) Nanoscale chemical tomography of buried organic-inorganic
interfaces in the chiton tooth. Nature 469:194–197
Harris JS (1989) An airplane is not a bird. Invent Tech 5:18–22
Heine C (1992) Mechanics of flapping fin locomotion in the cownose ray, Rhinoptera bonasus
(Elasmobranchii: Myliobatidae). Ph.D. Dissertation, Duke University
Hu DL, Chan B, Bush JWM (2003) The hydrodynamics of water strider locomotion. Nature
424:663–666
Jakab PL (1990) Visions of a flying machine. Smithsonian Institution Press, Washington
Katz J, Weihs D (1978) Hydrodynamic propulsion by large amplitude oscillation of an airfoil
with chordwise flexibility. J Fluid Mech 88:485–497
Katz J, Weihs D (1979) Large amplitude unsteady motion of a flexible slender propulsor. J Fluid
Mech 90:713–723
Katz SL, Jordan CE (1997) A case for building integrated models of aquatic locomotion that
couple internal and external forces. In: Tenth international symposium unmanned untethered
submersible technology: proceedings of the special session on bio-engineering research
related to autonomous underwater, Durham, NH, pp 135–152
Klausewitz W (1964) Der lokomotionsmodus der Flugelrochen (Myliobatoidei) Zool Anz
173:111–120
Koza JR (1994) Genetic programming as a means for programming computers by natural

selection. Stat Comp 4:87–112
Kuratani S (2009) Modularity, comparative embryology and evo-devo: Developmental dissection
of evolving body plans. Dev Biol 332:61–69
Lambers H, Chapin FS III, Pons TL (1998) Plant physiological ecology. Springer, New York
Larrabee EE (1980) The screw propeller. Sci Am 243:134–148
Lilienthal O (1911) Birdflight as the basis for aviation. Longmans, Green, London
Liu P, Bose N (1993) Propulsive performance of three naturally occurring oscillating propeller
planforms. Ocean Eng 20:57–75
Liu P, Bose N (1997) Propulsive performance from oscillating propulsors with spanwise
flexibility. Proc R Soc A 453:1763–1770
Low KH (2011) Current and future trends of biologically inspired underwater vehicles. Def Sci
Res Conf Expo. doi:10.1109/dsr.2011.6026887
Lu Y (2004) Significance and progress of bionics. J Bionics Eng 1:1–3
Luria SE, Gould SJ, Singer S (1981) A view of life. Benjamin/Cummings, Menlo Park
Margulis L, Sagan D (1997) What is sex?. Simon and Schuster, New York
Martinez JS, Carter-Franklin JN, Mann EL, Martin JD, Haygood MG, Butler A (2003) Structure
and membrane affinity of a suite of amphiphilic siderophores produced by a marine bacterium.
Proc Nat Acad Sci 100:3754–3759
Moored KW, Fish FE, Kemp TH, Bart-Smith H (2011a) Batoid fishes: inspiration for the next
generation of underwater robots. Mar Tech Soc J 45:99–109
Moored KW, Dewy PA, Leftwich MC, Bart-Smith H, Smits AJ (2011b) Bioinspired propulsion
mechanisms based on manta ray locomotion. Mar Tech Soc J 45:110–118
O’Dor RK, Webber DM (1986) The constraints on cephalopods: why squid aren’t fish. Can J
Zool 64:1591–1605
Parson JM, Fish FE, Nicastro AJ (2011) Turning performance of batoids: limitations of a rigid
body. J Exp Mar Biol Ecol 402:12–18
Pennisi E (2011) Bio-inspired engineering: manta machines. Science 332:1028–1029
Petroski H (1992) The evolution of useful things. Alfred A Knapf, New York
Petroski H (1996) Invention by design. Harvard University Press, Cambridge
Prempraneerach P, Hover FS, Triantafyllou MS (2003) The effect of chordwise flexibility on the
thrust and efficiency of a flapping foil. In: Proceedings of the Thirteenth international
symposium on unmanned untethered submersible technology: proceedings of the special
session on bio-engineering research related to autonomous underwater vehicles, Lee, New
Hampshire, Autonomous Undersea Systems Institute
Ralston E, Swain G (2009) Bioinspiration-the solution for biofouling control? Bioinsp Biomim
4:015007. doi:10.1088/1748-3182/4/1/015007
Rohr JJ, Fish FE (2004) Strouhal numbers and optimization of swimming by odontocete
cetaceans. J Exp Biol 207:1633–1642
Rosenberger LJ (2001) Pectoral fin locomotion in batoid fishes—undulation versus oscillation.
J Exp Biol 204:379–394
Schaefer JT, Summers AP (2005) Batoid wing skeletal structure: novel morphologies,
mechanical implications, and phylogenetic patterns. J Morph 264:298–313
Suzumori K, Endo S, Kanda T, Kato N, Suzuki HA (2007) A bending pneumatic rubber actuator
realizing soft-bodied manta swimming robot. In: Robotics and Automation, IEEE Interna-
tional Conference on 2007, pp 4975–4980
Takagi K, Yamamura M, Luo Z, Onishi M, Hirano S, Asaka K, Hayakawa Y (2006)
Development of a rajiform swimming robot using ionic polymer artificial muscles. In:
Intelligent Robots and Systems, IEEE/RSJ International Conference, pp 1861–1866
Triantafyllou GS, Triantafyllou MS, Grosenbaugh MA (1993) Optimal thrust development in
oscillating foils with application to fish propulsion. J Fluids Struct 7:205–224
Triantafyllou GS, Triantafyllou MS (1995) An efficient swimming machine. Sci Am 272:64–70
Triantafyllou MS, Triantafyllou GS, Yue DKP (2000) Hydrodynamics of fishlike swimming. Ann
Rev Fluid Mech 32:33–53
Ummat A, Dubey A, Mavroidis C (2006) Bio-nanorobotics: a field inspired by nature. In: Bar-
Cohen Y (ed) Biomimetics: biologically inspired technologies. CRC Press, Boca Raton,
pp 201–227
Van Valen L (1973) A new evolutionary law. Evol Theory 1:1–30
Vincent J (1990) Structural biomaterials. Princeton University Press, Princeton
Vogel S (1988) Life’s devices. Princeton University Press, Princeton
Vogel S (1994) Life in moving fluids. Princeton University Press, Princeton
Vogel S (1998) Cat’s paws and catapults. W. W. Norton, New York
Wainwright SA (1988) Axis and circumference: the cylindrical shape of plants and animals.
Harvard University Press, Cambridge
Wainwright SA, Biggs WD, Currey JD, Gosline JM (1976) Mechanical design in organisms.
Princeton University Press, Princeton
Webb PW (1975) Hydrodynamics and energetics of fish propulsion. Bull Fish Res Bd Can
190:1–158
Webb PW (1997) Designs for stability and maneuverability in aquatic vertebrates: What can we
learn. In: Tenth international symposium unmanned untethered submersible technology:
proceedings of the species on bio-engineering research related to autonomous underwater
vehicles, Durham, NH, pp. 85-108
Whitley D (1994) A genetic algorithm tutorial. Stat Comp 4:65–85
Yang S, Qiu J, Han X (2009) Kinematics modeling and experiments of pectoral oscillation
propulsion robotic fish. J Bionic Eng 6:174–179
Zhang S, Yokoi H, Zhao X (2006) Molecular design of biological and nano-materials. In: Bar-
Cohen Y (ed) Biomimetics: biologically inspired technologies. CRC Press, Boca Raton,
pp 229–242
Zhou C, Low K (2010) Better endurance and load capacity: an improved design of manta ray
robot (RoMan-II). J Bionic Eng 7:S137–S144
Author Biographies
Wojciech Bejgerowski received the MS degree in mechanical engineering from

Poznan University of Technology, Poznan, Poland in 2003. He received his Ph.D.
degree in mechanical engineering from the University of Maryland, College Park
in 2010. His research interests include In-mold Assembly Processes, Integrated
Product and Process Design and Manufacturing Process Optimization.
Dr. John T. Beneski is a Professor of Biology at West Chester University of

Pennsylvania, where he has been on the faculty since 1986. He earned a BA in
Zoology (1978) and a MA in Biology (1982) from Humboldt State University. He
received his Ph.D. in Evolutionary Morphology from the Department of Zoology,
Washington State University in 1986.
Dr. Beneski’s research emphasizes the evolution and adaptations of primitive
terrestrial vertebrates. Recent projects include the ontogeny, kinematics, evolution
and functional morphology of prey capture and intra-oral transport, the
biomechanics of tail autotomy, the evolution and functional significance of
tooth shape, the quantitative and qualitative analysis of tissue samples, and the 3-
dimensional modeling of biomechanical systems. Primary analytical techniques
include scanning electron microscopy, image analysis, and computer modeling. He
has been funded by the National Science Foundation, the PA Department of
Education, Endo Pharmaceuticals and the PA State System of Higher Education.
Janine Benyus is a biologist, innovation consultant, and author of six books,

including Biomimicry: Innovation Inspired by Nature. In Biomimicry, she names
an emerging discipline that seeks sustainable solutions by emulating nature’s
designs and processes (for instance, solar cells that mimic leaves). Since the
book’s 1997 release, Janine has evolved the practice of biomimicry, consulting
with sustainable businesses and conducting seminars about what we can learn from
the genius that surrounds us. Over the past 14 years, the Biomimicry Guild and the
Biomimicry Institute have developed the practice of nature-inspired innovation
and introduced it to the world. We have reached hundreds of thousands through
TED talks, feature films and documentaries, and media such as Time, Newsweek,
Forbes, Fortune and Wired. We started the world’s first bio-inspired consultancy,
bringing nature’s ideas to the design table and boardroom of 250+ companies
DOI: 10.1007/978-1-4471-5248-4, Ó Springer-Verlag London 2014
314 Author Biographies
including Nike, HOK, GE Procter and Gamble, Levi’s and General Mills. In 2011,
we combined our staffs and administration services into a single entity, to create
Biomimicry 3.8.
Dr. Amaresh Chakrabarti is a Professor at Centre for Product Design and

Manufacturing, Indian Institute of Science (IISc), Bangalore. He has a BE
(Mechanical Engineering) from University of Calcutta, India, ME (Mechanical
Design) from IISc, and a Ph.D. (Engineering Design) from University of
Cambridge, UK. After completing his Ph.D., he led for ten years the Design
Synthesis team at the EPSRC Centre for Excellence Engineering Design Centre at
University of Cambridge, before joining IISc. He works in design synthesis,
creativity, eco-design, sustainability, and knowledge engineering/management. He
has authored or edited 9 books, authored over 250 peer-reviewed articles, and has 8
patents granted or pending. He co-authored DRM, a methodology used widely
around the world as a framework for carrying out engineering design research. He is
an Associate Editor of AI EDAM Journal (CUP), Area Editor for Research in
Engineering Design (Springer), Regional Editor for Journal for Remanufacturing
(Springer), and Advisory Editor for seven International Journals. He is on the
Advisory Board of Design Society, UK, and a member of the CII National
Committee on Design, India. He founded IDeASLab – the first laboratory in India
for research into design, and ICoRD – the first conference series in design research.
He is an Honorary Fellow of the Institution of Engineering Designers, UK.
Hyunmin Cheong is currently completing his Ph.D. at the University of Toronto.

He obtained his MASc and BASc in Mechanical and Industrial engineering at the
University of Toronto. His research interests include applying natural-language
processing techniques to support biomimetic design and understanding designers,
analogical reasoning, and creativity.
Jon-Michael Deldin is a software developer for Biomimicry 3.8. He maintains the

current version of AskNature and is developing its next evolutionary leap.
Additionally, he is the lead programmer of the Biomimicry Student Design
Challenge and Biomimicry Education Network websites. His professional interests
include designing developer-friendly APIs, building highly-usable products, and
applying machine learning techniques for insight into data. He holds a master’s
degree in computer science from the University of Montana-Missoula, and
undergraduate degrees in biology and chemistry from the University of Great
Falls. In his spare time, he can be found riding mountain bike trials, hiking, and
pursuing assorted side-projects in programming and data visualization.
Dr. Frank Fish is a Professor of Biology at West Chester University of

Pennsylvania, where he has been on the faculty since 1980. He received a BA in
Biology from SUNY Oswego in 1975. He completed a MS in 1977 and a Ph.D. in
1980 from the Zoology Department of Michigan State University.
Author Biographies 315
Dr. Fish has served as chairperson of his department at West Chester

University. He was the chairperson of the Division of Vertebrate Morphology of
the Society for Integrative and Comparative Biology. He is currently an associate
editor for the journal Marine Mammal Science, and he served on the editorial
board of the journal Bioinspiration and Biomimetics.
Dr. Fish’s research is focused on the energetics and hydrodynamics of aquatic
locomotion by vertebrates and on biomimetic applications. He has been funded by
the Defense Advanced Research Projects Agency, the National Science
Foundation, and the Office of Naval Research. Recent projects have included
examinations of the evolution of swimming modes in aquatic mammals, energetics
and maneuverability of aquatic animals, hydrodynamic design of biological
control surfaces, and analysis of manta swimming for biomimetic AUV design.
John Gerdes is currently working as a Mechanical Engineer in the Vehicle

Technology Directorate of the US Army Research Laboratory. He received his
M.S. in mechanical engineering in 2009 and is currently pursuing his Ph.D. in
mechanical engineering, both from the University of Maryland, College Park. His
research interests include biological robotics and systems engineering.
Dr. Ashok K. Goel is a Professor of Computer and Cognitive Science in the School
of Interactive Computing at Georgia Institute of Technology in Atlanta, USA. He is
Director of the School’s Design and Intelligence Laboratory, and a Co-Director of
Georgia Tech’s Center for Biologically Inspired Design. He serves on the Boards of
Biomimicry 3.8 Institute and BioInspired! digital magazine. Born and brought up in
India, Ashok immigrated to the U.S. for graduate studies. He lives in mid-town
Atlanta with his two sons. Ashok conducts research into artificial intelligence and
cognitive science, with a focus on computational design, modeling and creativity.
His recent TEDx talk http://www.youtube.com/watch?v=wiRDQ4hr9i8 summa-
rizes some of his research on AI, biologically inspired design, and environmental
sustainability. An interactive tool for supporting some aspects of biologically
inspired design developed by his laboratory is available at http://dilab.cc.
gatech.edu/dane/.
Dr. Satyandra K. Gupta is a Professor in the Department of Mechanical

Engineering and the Institute for Systems Research at the University of Maryland.
He was the founding director of the Maryland Robotics Center. His interest is in
automation problems arising in Computer Aided Design, Manufacturing
Automation, and Robotics areas. The current focus of his research is on
automated planning for human robot collaborations in assembly tasks, autonomous
unmanned sea surface vehicles, autonomous unmanned ground vehicles, and
automated optical micro-manipulation. He has authored or co-authored more than
two hundred fifty articles in journals, conference proceedings, and book chapters.
He has received many awards for his research contributions including a
Presidential Early Career Award for Scientists and Engineers (PECASE) and five
best paper awards in conferences. For description of his latest projects please visit
http://terpconnect.umd.edu/*skgupta/.
Michael Helms holds a BA in Finance from the University of Illinois, Urbana-

Champaign, an MS in Computer Science from DePaul University in Chicago, and
is currently a Ph.D. candidate in Computer Science at the Georgia Institute of
Technology working in the Design and Intelligence Laboratory in the School of
Interactive Computing. He is an instructor in Biologically Inspired Design (BID)
at Georgia Tech, and has taught BID courses at NASA and Tohoku University,
Japan, through his affiliation with the Center for Biologically Inspired Design at
Georgia Tech. His forthcoming Ph.D. thesis investigates problem formulation and
problem-solution coevolution in the context of Biologically Inspired Design. Prior
to his work in artificial intelligence and design theory, Michael provided business
information strategy consulting to Fortune 500 financial services companies.
James K. Hopkins currently works in the Propulsion and Power Engineering

Department at the U.S. Naval Air Systems Command. He received his BS degree
in mechanical engineering from Ohio University in 2001. He received his MS
degree in engineering management from Florida Institute of Technology in 2004.
He is pursuing his Ph.D. in mechanical engineering from University of Maryland
at College Park. His research interests include design and mechanics of snake-
inspired robots.
Dr. Venkat Krovi is an Associate Professor of Mechanical and Aerospace

Engineering and the Director of Automation, Robotics and Mechatronics (ARM)
Laboratory at the State University of New York at Buffalo. His research interests
are in the lifecycle treatment (design, modeling, analysis, control, implementation
and verification) of novel biorobotic and biomechatronic systems. His work has
been funded by NSF, DARPA, ARO, state and industrial grants and has received
multiple awards, including the National Science Foundation (NSF) CAREER
Award, Petro-Canada Young Innovator Award, several best poster and journal
paper awards. See http://mechatronics.eng.buffalo.edu for details.
Dr. Lengfeng Lee is currently working as a post-doctoral researcher in

Department of Kinesiology at Univ. of Massachusetts, Amherst since 2011. He
completed his Ph.D. (2011), MS (2005) and BS (2003) degrees in Mechanical
Engineering from State University of New York at Buffalo. His research interests
include design and control of articulated multi-body systems, motion planning and
musculoskeletal analysis.
Dr. Julie S. Linsey is an assistant professor at the Georgia Institute of Technology

in the Woodruff School of Mechanical Engineering. She earned a Ph.D. and MS in
mechanical engineering from The University of Texas at Austin and a BS in
mechanical engineering from the University of Michigan. Her research focus is on
systematic methods and tools for innovative design with a particular focus on
concept generation and design-by-analogy. Her research seeks to understand
designers’ cognitive processes with the goal of creating better tools and
approaches to enhance innovation. She has co-authored over forty technical
publications including four book chapters and she holds two patents.
Dr. Daniel A. McAdams is currently an associate professor Department of

Mechanical Engineering at Texas A&M University. His research interests are in
the area of design theory and methodology with specific focus on functional
modeling; innovation in concept synthesis; biologically inspired design methods;
inclusive design; and technology evolution as applied to product design. Dr.
McAdams leads the Product Synthesis Engineering Lab (http://www.
prosedesign.org/). He has served in various scholarly leadership roles including
Technical Committee Chair for ASME Design Theory and Methodology Com-
mittee and serving on the editorial board of the Journal of Engineering Design. Dr.
McAdams served as an associate and assistant professor at the Missouri University
of Science and Technology. He received his Ph.D. from the University of Texas at
Austin.
Dr. Frank Mendel is the Vice Chairman and Associate Professor in the
Department of Pathology and Anatomical Sciences at the State University of New
York at Buffalo, NY, USA. He received his MA and Ph.D. degrees in
anthropology from the University of California, Davis in 1971 and 1975,
respectively. His research interests include forensic pathology of extant and extinct
big cats and cat-like carnivores (saber-toothed forms), developing virtual clinical
training platforms and determining best practices for managing sprains and other
soft tissue injuries.
Dr. Jacquelyn K. Nagel is an Assistant Professor in the Department of

Engineering at James Madison University. Dr. Nagel has seven years of
diversified engineering design experience, both in academia and industry, and
has experienced engineering design in a range of contexts, including product
design, bioinspired design, electrical and control system design, manufacturing
system design and design for the factory floor. In 2012, Dr. Nagel was recognized
by the National eWeek Foundation and IEEE-USA as one of the New Faces of
Engineering for her pioneering work in using biological systems as models for
sensors, instrumentation and processes. She earned her Ph.D. in Mechanical
Engineering from Oregon State University, and her MS and BS in Manufacturing
Engineering and Electrical Engineering, respectively, from Missouri University of
Science and Technology. Her research interests include biomimicry, sensors,
engineering design theory, sustainable energy, and manufacturing.
Madusudanan Sathia Narayanan received the BEng degree in mechanical

engineering from the College of Engineering Guindy (Anna University-Main
Campus), India in 2006. He received the MS degree in mechanical engineering in
2008 from the State University of New York at Buffalo, Buffalo, NY, USA. He is
currently pursuing his Ph.D. degree in mechanical engineering since 2009 from the
same institution. His research interests include kinematics, haptics and
medical robotics.
Megan Schuknecht is the Director of University Education at the Biomimicry 3.8

Institute. She works with faculty and administrators to incorporate biomimicry
tools and concepts into university and college classrooms, curricula, and degree
programs. She promotes and fosters formal relationships between the Biomimicry
3.8 Institute and universities to establish biomimicry degree programs through the
Institute’s Biomimicry Affiliate Program. She also develops and evaluates
biomimicry teaching and learning materials, and shares those materials with
educators and students through workshops, in-person and online courses, and
curricula development. Megan orchestrates both the Institute’s annual student
design challenge and its annual Biomimicry Education Summit. She played a lead
role in developing the AskNature website and continues to contribute to the site’s
growth, with an eye on how best the site can evolve to meet student and faculty
needs.
Dr. L.H. Shu, Associate Professor in the Department of Mechanical and Industrial
Engineering at the University of Toronto, received her SM and Ph.D. in
Mechanical Engineering from the Massachusetts Institute of Technology. She has
held research fellowships at the Technical University of Denmark, AlliedSignal
Aerospace Canada Inc., Eastman Kodak, Naval Ocean Systems Center, and Naval
Training Systems Center. She received the CIRP (International Academy for
Production Engineering Research) F. W. Taylor Medal and numerous best paper
awards for her work on biologically inspired design, and held leadership positions
in both the Design Society and the ASME Design Theory and Methodology
community. Dr. Shu’s research interests are in creativity in conceptual design,
systematic identification and application of biological analogies in biomimetic
design, affordances, and identifying and overcoming obstacles to personal
environmentally significant behavior.
Dr. Robert B. Stone is a Professor and the Interim Head of the School of
Mechanical, Industrial and Manufacturing Engineering at Oregon State University.
Dr. Stone s research interests include design theories and
methodologies, specifically automated conceptual design techniques and concept
generation techniques such as biologically inspired design. He has
authored chapters on product architecture and reverse engineering techniques in
product design texts. He leads the Design Engineering Lab (http://www.
designengineeringlab.org). Previously, Stone utilized his design background to
assist in creating the design-focused Interdisciplinary Engineering degree program
at the University of Missouri-Rolla. At Rolla he served as Director of the inter-
disciplinary Student Design and Experiential Learning Center from January 2004
through June 2008 where he oversaw the design competition activities of ten teams
and guided the Center’s new engineering design and experiential learning initia-
tive. He was a Distinguished Visiting Professor at the US Air Force Academy
during the 2006–2007 academic year. Prior to initiating his graduate work, Dr.
Stone worked in the Missions Operation Directorate of NASA-Johnson Space
Center as a Space Shuttle Flight Controller for the Guidance, Navigation and
Control Section. Stone completed his Ph.D. in Mechanical Engineering from The
University of Texas at Austin.
Dr. Craig Tovey is the David McKenney Family Professor in the School of
Industrial and Systems Engineering at Georgia Tech. He received his AB in
applied mathematics from Harvard College in 1977 and both an MS in computer
science and a Ph.D. in operations research from Stanford University in 1981.
Tovey’s principal research and teaching activities are in operations research and its
interdisciplinary applications to social and natural systems, with emphasis on
sustainability, the environment, and energy. He co-founded and co-directs Georgia
Tech’s Center for Biologically Inspired Design. His current research concerns
inverse optimization for electric grid management, classical and biomimetic
algorithms for robots and webhosting, the behavior of animal groups, political
economy, and sustainability measurement.
Dr. Tovey received a Presidential Young Investigator Award in 1985 and the
1989 Jacob Wolfowitz Prize for research in heuristics. He was granted a Senior
Research Associateship from the National Research Council in 1990, was named
an Institute Fellow at Georgia Tech in 1994, and received the Class of 1934
Outstanding Interdisciplinary Activity Award in 2011.
Dr. Swaroop S. Vattam is a Research Scientist at the School of Interactive

Computing at Georgia Institute of Technology. He has degrees in computer
science and artificial intelligence and recently completed his Ph.D. degree in
Computer Science from the Georgia Institute of Technology. His research interests
lie in artificial intelligence, cognitive science, and human-centered computing. He
is primarily interested in the development of novel artificial intelligence systems
that espouse the potential of emerging technologies, but that are also grounded in
theories of human cognition and studies of communities of practice. He is a firm
believer that computational thinking enables us to conceptualize, modularize, and
approach discovery and innovation in fundamentally different ways than we might
consider otherwise. He is also a reviewer of several international journals and
conferences in the areas of design computing and cognition.
Dr. Julian Vincent began his career as an entomologist with a Ph.D. studying
insect hormones acting on cuticle, then a DSc on the mechanical design and
properties of the extensible cuticle of the female locust. This got him interested in
materials and mechanical properties, a topic which he developed to cover aspects
of plants and animals and food, concentrating specifically on fracture mechanics.
All this time he was at the University of Reading in the department of zoology,
collaborating with George Jeronimidis and Jim Gordon in the department of
Engineering. Fascinated by the idea of turning biological phenomena into

engineering, he became interested in biomimetics, eventually being appointed
Professor in the department of Mechanical Engineering at the University of Bath.
He has retired, but continues writing and lecturing around the world, and works
part time for Swedish Biomimetics 3000, the only company in the world
specialising in the development of biomimetic products.
Dr. Vimal Viswanathan is a post-doc at at the Georgia Institute of Technology in

the Woodruff School of Mechanical Engineering. He earned a Ph.D. and MS in
mechanical engineering from Texas A&M University. He had co-authored over
twenty technical publications.
Dr. Marc Weissburg is Professor of Biology and Co-Director of the Center for
Biologically Inspired Design. He is an ecologist/sensory biologist who examines
the mechanisms and consequences of information transfer via aquatic chemical
signals. He uses multidisciplinary approaches and field ecological investigations to
study the structure of aquatic plumes and the dynamics of fluids in the marine
environment and to behaviorally analyse the sensory strategies of aquatic
organisms and their capability to rely on turbulent chemical plumes for
guidance and navigation. He has used biological principles to devise artificial
sensory processing strategies for autonomous navigation in chemical plumes. He
has applied principles of ecological organization to human infrastructure in his
search for more sustainable practices. He has co-taught Biologically Inspired
Design for seven years to a variety of audiences ranging from undergraduates to
professional engineers and scientists.
Bryan Wiltgen is a Ph.D. student in Computer Science at the Georgia Institute of

Technology, where he is advised by Dr. Ashok K. Goel. Bryan conducts research
in the Design & Intelligence Laboratory at Georgia Tech. Bryan holds a BS with
Honors in Computer Science from Kennesaw State University and a MS in
Computer Science from the Georgia Institute of Technology. Bryan has research
interests in the areas of analogical reasoning, knowledge representation,
conceptual change, and distributed cognition. Please go to http://www.cc.
gatech.edu/*bwiltgen/ to find more information about Bryan and his research.
Dr. Jeannette Yen is the Director of the Georgia Institute of Technology’s Center
for Biologically Inspired Design (CBID). The goals of CBID are to bring together
a group of interdisciplinary faculty who seek to facilitate interdisciplinary research
and education for innovative products and techniques based on biologically-
inspired design solutions. The participants of Georgia Tech’s CBID believe that
science and technology are increasingly hitting the limits of approaches based on
traditional disciplines, and Biology may serve as an untapped resource for design
methodology, with concept-testing having occurred over millions of years of
evolution. Experiencing the benefits of Nature as a source of innovative and
inspiring principles encourages us to preserve and protect the natural world rather
than simply to harvest its products. Jeannette team-teaches the interdisciplinary

course in biologically inspired design [http://www.youtube.com/watch?v=
PMlvUJ9_GSk].
Jeannette Yen’s Ph.D. is in interdisciplinary environmental science of
biological oceanography where she studies how fluid mechanical and chemical
cues transported at low Re flow serve as communication channels for micro-
aquatic organisms, primarily zooplankton: key link in aquatic food webs. She has
been to all 7 continents, including Antarctica for her research and education. She is
a Professor of Biology and has been at Georgia Tech since 2000.
Index
A Biomimicry taxonomy, 19, 21, 24, 25

Adaptation, 6, 270, 272, 279, 288, 291, 293, Bionics, 1
306, 307 Brainstorming, 77, 81, 85, 231, 237
Analogical reasoning, 2, 43, 66, 139, 155, 163,
177, 194, 195, 223
Analogical design, 32, 55, 135, 138–140, 160 C
Analogical mapping tools, 42 Cluster analysis, 278, 282
Analogical reasoning, 2, 139 Cognitive biases, 222, 241
Analogical transfer, 8, 44, 135, 181, 187, 202, Compound analogy, 99, 140, 143, 148, 149
217, 230, 231 Computational linguistics, 34, 40, 44, 56
Analogies, 8, 29, 30, 31, 33, 35, 38, 48, 54, 81, Computer-aided design, 259
86, 99, 107, 127, 134–136, 140, 143, Concept development, 78
149, 157, 181, 195, 222, 225, 289 Concept generation, 8, 64, 68, 77, 79, 81, 82,
Asknature, 9, 17–19, 21–24, 33, 55, 269, 26, 87, 89, 97, 107, 122
110 Connections, 64, 65, 68, 77, 81, 89, 97, 99,
Auv (autonomous underwater vehicle), 298, 107, 122, 168, 194, 296
301 Consolidation or merging, 269
Creativity, 7, 99, 155, 178, 189, 202, 206, 213,
218
B Cross-domain analogy, 134, 143, 179
Biological stimuli, 201, 215, 218
Biologically inspired design, 10, 13, 30, 128,
129, 134, 138, 140, 147, 149, 154, 157, D
175, 202, 217, 283 Dane, 172
Biologue, 172 Darwin, 288, 289, 291
Bio-mimetic systems, 30, 31, 43, 55, 56, 109, Database, 9, 17, 18, 21, 22, 24, 26, 30, 35, 33,
205 55, 66, 74, 99, 143, 156, 203, 269, 278
Biomimetic design applications, 57 Design, 10, 11, 12, 17, 18, 19, 22, 24, 26, 43,
Biomimetic design methods, 31, 57, 98 55, 56, 77, 128, 143, 157, 159, 160,
Biomimetics, 30, 109, 203, 216, 269, 287, 289, 173, 189, 249, 252, 255, 259, 263
293, 298, 308 Design by analogy, 222, 238
Biomimicry, 10, 11, 17–19, 21, 24, 26, 99, Design creativity, 7, 207
148, 172, 269, 307, 308 Design evaluation, 160, 178, 194

DOI: 10.1007/978-1-4471-5248-4, Ó Springer-Verlag London 2014
324 Index
Design innovation, 96, 128, 190 K

Design methodology, 8, 96, 97, 129 Kinematics, 11, 299, 302
Design tools, 9, 66, 76, 87, 98, 172
Dialogue facilitation, 68, 77
Dynamics, 11, 190, 224, 246, 251, 269, 282 L
Learning biologically inspired design, 130,
138
E Local quality, 269, 281, 284
Engineering creativity, 155, 213
Engineering design, 3, 4, 7, 63, 66, 68, 77, 87,
96, 99, 204 M
Evolution, 6, 138, 165, 194, 289, 293 Mantabot, 298, 302
Multi-disciplinarity, 12, 13
Muscle, 86, 249, 260, 275, 296, 304, 305
F
Fitness, 289, 291, 293, 304, 305
Fixation, 7, 42, 89, 177, 194, 231, 235, 236, N
237 Natural-language processing, 43, 46
Framework, 7, 21, 56, 67, 90, 97, 99, 170, 180, Natural selection, 288, 290, 291, 294, 304, 306
181, 202, 214, 246, 272, 289, 295, 305 Novelty, 7, 159, 189, 193, 202, 206, 209, 217,
Function-based design, 74, 79, 98 279
Functional basis, 8, 38, 56, 66, 71, 73, 76, 81,
85, 87, 88, 90, 106, 233
Functional biology, 8, 18, 33, 73, 96, 157 O
Functional decomposition, 130, 134, 135, 148, Ontology, 2, 34, 143, 273, 278
173, 194
Functional modeling, 7, 34, 77, 81, 82, 169
P
Parallel manipulators, 248, 263
G Phenotype, 290, 292
GEMS of SAPPhIRE, 209, 214 Prior cushioning, 269, 282
Genetic variation, 291 Problem decomposition, 8, 32, 134, 135, 149,
Guidelines, 7, 101, 163, 203, 205, 216 188, 203
Problem-driven analogy, 139, 140, 148, 149
Problem-driven approach, 31, 109, 175
H Problem-driven design, 148, 157, 175
Hierarchy, 19, 91, 107, 168–170, 273, 279, Problem-solution co-evolution, 129
281 Protocol analysis, 43, 222
I R
Idea-inspire, 20, 67, 74, 161, 204, 205, 211, Reproduction and replication, 57, 109, 216,
215 256, 296, 308
Identification, 74, 85, 130, 247 Robotic design, 260, 303
Information, 3, 7, 17, 18, 21, 35, 47, 55, 80, Robotics, 10, 247, 260, 266, 303
129, 147, 212, 238, 281
Information extraction, 55–57
Information retrieval, 68, 73, 74, 79 S
Inspiration search, 6, 67, 85 Search categorization, 39
Interdisciplinary design, 129, 181 Search keywords, 33, 35, 51
Interdisciplinary education, 10 Sensor design, 116, 120
Inventive principle, 10, 146, 275, 276, 279, Solution-based analogy, 138–140, 143, 148,
283 149
Index 325
Solution-based design, 129, 130, 140, 157, Team-based learning, 194

175, 188 Team design, 177, 190
Solution-driven approach, 31 Technical product development, 206
Structure-behavior-function,, 9, 164, 168, 172 Thesaurus, 8, 66, 68, 71, 74, 76, 79, 86, 90
Systematic design, 96, 99, 106, 108, 109, 129, Translation, 38, 56, 68, 79, 82, 189, 261, 298
205, 214 TRIZ, 11, 30, 129, 145, 269, 274, 278, 283
T U
Task analysis, 129–131, 133, 148, 149 Understanding biological systems, 156, 177,
Task model, 129–131, 139, 140, 143–145, 179
147, 149 Usefulness, 7, 206, 211, 213
Teaching biologically inspired design, 2, 157

2014 Book BiologicallyInspiredDesign

Uploaded by

Copyright:

Available Formats

2014 Book BiologicallyInspiredDesign

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2014 Book BiologicallyInspiredDesign

Uploaded by

Copyright:

Available Formats

What are the main topics covered in the document?

What are the main topics covered in the document?

What are some of the design methodologies discussed?

What are some of the design methodologies discussed?

Ashok K.

ISBN 978-1-4471-5247-7 ISBN 978-1-4471-5248-4 (eBook)

Library of Congress Control Number: 2013943573

Ó Springer-Verlag London 2014

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)

I would like to thank the students I have had

The Moment of Truth

Society of Manufacturing Engineers’ annual list of ‘‘innovations that could change

Transferable Ideas and Downloadable Beaks

Building a biological intelligence tool for inventors is the quintessential exercise in

Helping Innovators Meet Their Mentor

Biologically inspired design is a promising paradigm for design innovation as well

In Chap. 4, Jacquelyn Nagel presents an engineering-to-biology thesaurus,

1 Charting a Course for Computer-Aided Bio-Inspired Design . . . . 1

2 The AskNature Database: Enabling Solutions

3 A Natural Language Approach to Biomimetic Design . . . . . . . . . 29

4 A Thesaurus for Bioinspired Engineering Design. . . . . . . . . . . . . 63

5 Function-Based Biologically Inspired Design . . . . . . . . . . . . . . . . 95

6 Information-Processing Theories of Biologically

7 Adaptive Evolution of Teaching Practices in Biologically

8 Supporting Analogical Transfer in Biologically

9 Overcoming Cognitive Challenges in Bioinspired Design

10 An Engineering Approach to Utilizing Bio-Inspiration

11 An Ontology of Biomimetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

12 Evolution and Bio-Inspired Design: Natural Limitations . . . . . . . 287

Author Biographies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Robert B. Stone, Ashok K. Goel and Daniel A. McAdams

Abstract Bio-inspired design (BID) is an emerging research area in design,

Bio-inspired design (BID or biomimicry or bionics) is an emerging research area

A. K. Goel et al. (eds.), Biologically Inspired Design, 1

of BID researchers at present is fragmented with no professional society, unifying

1.1.1 NSF Workshops

Table 1.1 Presenters and presentation titles at workshop 1

Table 1.2 Workshop 1 Name Affiliation

Following the presentations, three breakout groups were formed to generate

Table 1.3 Presenters and presentation titles at workshop 2

Conference to evaluate the first workshop’s findings and generate a draft

1.2 Current Research from the Disciplines

Table 1.4 Workshop 2 Name Affiliation

1.2.2 Bridging Biology and Engineering

1.2.3 Techniques and Tools for Bio-Inspired Conceptual

1.2.4 Education in Bio-Inspired Design

fundamentals of robotics, including kinematics, inverse kinematics, dynamics,

Biologically inspired design has been practiced on an ad hoc basis throughout

1.3.1 A Design Theory for Bio-Inspired Design

1.3.2 A Funding Program for Bio-Inspired Design Research

• BID system usability, interface design, visualization, and search;

Arciszewski T, Cornell J (2006) Bioinspiration: learning creative design principles. In intelligent

Bridging Biology and Engineering

Techniques and Tools for Bio-Inspired Conceptual Design

Chakrabarti A, Sarkar P, Leelavathamma B, Nataraju B (2005) A functional representation for

Lenau T (2009) Biomimetics as a design methodology—possibilities and challenges, interna-

Education in Bio-Inspired Design

Bruck H, Gershon A, Golden I, Gupta S, Gyger L, Magrab E, Spranklin B (2007) Training

Design Theory for Bio-Inspired Design