Cooperative Robots Architecture For An Assistive Scenario

Cooperative Robots Architecture for an Assistive
Scenario
L. Ciuccarelli∗ , A. Freddi∗ , S. Longhi∗ , A. Monteriù∗ , D. Ortenzi† and D. Proietti Pagnotta∗
∗ Dipartimento di Ingegneria dell’Informazione, Università Politecnica delle Marche, Ancona, Italy
Email: {a.freddi, sauro.longhi, a.monteriu}@univpm.it, d.proietti@pm.univpm.it
† Dipartimento di Ingegneria dell’Energia Elettrica e dell’Informazione, Università di Bologna, Bologna, Italy
Email: davide.ortenzi@unibo.it
Abstract—This paper proposes an assistive robotic solution for

elderly people and/or people suffering from lower limbs problem.
The proposed architecture consists of a Smart Wheelchair (SWC)
and a Robotic Workstation (RW) which cooperate in order to
perform an autonomous navigation and manipulation task. The
system allows the user to choose a point on the map to move to
and/or an object to manipulate. The SWC is able to autonomously
navigate to the desired RW, while the localization is performed
via an Unscented Kalman Filter (UKF). Moreover, QR codes
have been used as landmarks to compensate the pose estimation
error. The system was tested both in simulation and in a real
scenario. (a) (b)
I. I NTRODUCTION Figure 1: System simulation.

In the last years, increasingly more attention is given to the
Socially Assistive Robotics (SAR) field, which has taken ad- proposed in [6]. However, this type of solutions has a number
vantage from important financial investments [1]. A category of drawbacks, including the working of the arm in an unstruc-
of SAR consists of physically assistive robots, which aim to tured environments and the power supply of the system.
cooperate with humans in order to give physical assistance This paper presents the development and the implementation
and provide help to perform daily living tasks, such as shared of a real prototype of the architecture proposed in authors’ pre-
load transportation or passing objects [2]. vious work [7]. In detail, the architecture proposed consists of
In this context, an application scenario with large potential a Smart Wheelchair (SWC) and a Robotic Workstation (RW)
is using assistive robots for increasing the autonomy and which cooperate to perform an autonomous navigation and
promoting a more independent life style of disabled users manipulation task. The prototype is tested in our laboratory at
and/or elderly people. In the literature, various assistive robot the University of Marche.
solutions, which aim at providing different types of support,
have been proposed. In [3] a system which aims to detect II. S YSTEM OVERVIEW
falls in real-time based on a smartphone and a mobile robot A. Hardware
is developed. Acceleration data are gathered through the In this section, it is listed the main hardware components
accelerometer of the smartphone in order to detect body which constitute the prototype. The manipulator of the RW is
movements and falls, while the mobile robot has the object a Baxter Research Robot, which is equipped with a couple
to establish if the detected fall has really happened and, of seven Degrees of Freedom (DOFs) arms. The RW is
eventually, call the aids. A compliant robotic manipulator for equipped also with a Kinect v2 camera. The SWC is a Quickie
supporting the dressing task is proposed in [4]. In detail, Salsa R2, an electric wheelchair with two rear driving wheels
a Baxter robot dresses a mannequin by tracking the joint and two forward caster ones, equipped with the following
locations and computing the arm trajectory. Authors of [5] sensors: an Arduino ATMega2560 micro-controller, a LORD
realized a framework where haptic robots act as mediators MicroStrain 3DM-GX3-25 Inertial Measurement Unit (IMU),
allowing multiple users to interact to each other and supporting a Sicod F3-1200-824-BZ-K-CV-01 encoder for each driving
a human-human group interaction in the same environment. wheel, a Hokuyo URG-04LX laser scanner and a Logitech
A feasible and effective solution for the navigation and C270 webcam.
manipulation tasks for people suffering from lower limbs
problem consists in the cooperation among a smart wheelchair B. Software
and one or more robotic manipulators. The common approach The user is allowed by the SWC controller to choose a
is mounting a light weight arm on a smart wheelchair, as point on the map to move to and/or an object to manipulate.
978-1-5386-4927-5/18/$31.00 ©2018 IEEE 128

Objects information are stored in a database in order to, at
each time, the SWC controller knows where a desired object
is located and is able to move the SWC towards that point in
an autonomously way.
The controllers of both the RW and the SWC are developed
under the Robot Operating System (ROS) middleware. Both
the controllers consist of a package (one for the RW and one
for the SWC) which collects several nodes that implement
various algorithms, which are explained in the following
section. The Gazebo simulator is exploited to test the system
(and its subsystems) in a simulated scenario. Figure 2: Evaluation of the real system.
III. M ETHODS
using Eq. (1) in [8]. Finally, the desired object is picked by
A. SWC Localization Algorithm
the arm and placed on the user’s hand. This is accomplished
This phase is articulated in two steps. Firstly, an Unscented by means of the external Kinect sensor, which is able to track
Kalman Filter (UKF) is applied to remove the noise and to online the skeleton of the user, and thus the location of his
determine the position of the SWC with respect to an absolute hand.
coordinate reference frame. The filter carries out a sensor
fusion of the data gathered from the encoders, the IMU and IV. E XPERIMENTAL T RIAL
the webcam. In detail, the IMU is exploited to improve the The proposed system is evaluated both in simulation and in
heading estimation, while the webcam is used to compensate a real scenario. Regarding the simulation, the 3D model of the
the drift in the estimation of the pose obtained from the SWC is developed in the Gazebo 7 simulator, complete with
proprioceptive sensors (encoders and IMU). QR codes have the webcam, IMU and laser scanner models. Moreover, 9 QR
been used as landmarks, placed on the ceiling, and encapsulate codes have been positioned in the simulated laboratory, with
the information of the real pose with respect to the absolute a density of about 0.18 codes/m2 (Figure 1(a)). A graphical
reference frame in the form X#Y#Z#room. When a code is user interface has been realized by which the user can choose
focused by the webcam, this information feeds the UKF which the desired object (Figure 1(b)). Once the object is chosen, the
can correct the SWC pose estimation. SWC autonomously navigates to the desired RW, where the
Secondly, the Monte Carlo algorithm employs the data from robotic arm picks the user selected object and places it in the
the laser scanner and the output of the UKF to localize the neighborhood of the wheelchair. The user can then choose a
SWC in the map. new point on the navigation map. Finally, the real system is
tested in our laboratory, as shown in Figure 2.
B. SWC Navigation Algorithm
R EFERENCES
The aim of the navigation algorithm is to compute the
velocity commands to send to the wheels controller. The inputs [1] D. Feil-Seifer and M. J. Mataric, “Defining socially assistive robotics,” in
IEEE 9th International Conference on Rehabilitation Robotics (ICORR),
are the estimated pose of the SWC, from the localization 2005, pp. 465–468.
algorithm, the transformations between the considered frames [2] A. Sciutti and G. Sandini, “Interacting with robots to investigate the
and the laser scanner data. The algorithm is based on the bases of social interaction,” IEEE Transactions on Neural Systems and
Rehabilitation Engineering, vol. 25, no. 12, pp. 2295–2304, 2017.
Dynamic Window Approach (DWA), which associates the scan [3] L. Ciabattoni, F. Ferracuti, G. Foresi, A. Freddi, A. Monteriù, and D. P.
map to a grid, where each cell assumes a number that identifies Pagnotta, “Real-time fall detection system by using mobile robots in smart
the probability of its occupation. The trajectory is generated homes,” in IEEE 7th International Conference on Consumer Electronics-
Berlin (ICCE-Berlin), 2017, pp. 15–16.
considering the original global path, the presence of obstacles [4] G. Chance, A. Camilleri, B. Winstone, P. Caleb-Solly, and S. Dogramadzi,
and the location of the goal. “An assistive robot to support dressing-strategies for planning and error
handling,” in IEEE 6th International Conference on Biomedical Robotics
C. RW Pick and Place Algorithm and Biomechatronics (BioRob), 2016, pp. 774–780.
[5] H. H. Le, M. J. Loomes, and R. C. Loureiro, “Group interaction through
A visual servoing pick and place is carried out by the Baxter a multi-modal haptic framework,” in IEEE 12th International Conference
arm when the SWC completed the navigation. In detail, the on Intelligent Environments (IE), 2016, pp. 62–67.
[6] F. Achic, J. Montero, C. Penaloza, and F. Cuellar, “Hybrid bci system
algorithm is articulated in the following steps. Firstly, the to operate an electric wheelchair and a robotic arm for navigation and
object location is estimated in the camera frame from the manipulation tasks,” in IEEE Workshop on Advanced Robotics and its
image data acquired via the camera mounted at the end of the Social Impacts (ARSO), 2016, pp. 249–254.
[7] G. Foresi, A. Freddi, A. Monteriù, D. Ortenzi, and D. P. Pagnotta,
Baxter arm. In order to find the object center of gravity, the rgb “Improving mobility and autonomy of disabled users via cooperation
image is converted into a black and white one to find the object of assistive robots,” in IEEE International Conference on Consumer
contours. Then, via the moments calculus, the center of gravity Electronics (ICCE), 2018, pp. 1–2.
[8] G. Foresi, A. Freddi, S. Iarlori, A. Monteriù, D. Ortenzi, and D. P. Pag-
is computed. Secondly, the position of the object is expressed notta, “Human-robot cooperation via brain computer interface,” in IEEE
in the Baxter base frame by knowing the arm joint angles, its 7th International Conference on Consumer Electronics-Berlin (ICCE-
height from the object and the camera calibration factor by Berlin), 2017, pp. 1–2.
129

Cooperative Robots Architecture For An Assistive Scenario

Uploaded by

Copyright:

Available Formats

Cooperative Robots Architecture For An Assistive Scenario

Uploaded by

Document Information

Original Title

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

Cooperative Robots Architecture For An Assistive Scenario

Uploaded by

Copyright:

Available Formats

Cooperative Robots Architecture for an Assistive

Abstract—This paper proposes an assistive robotic solution for

I. I NTRODUCTION Figure 1: System simulation.

978-1-5386-4927-5/18/$31.00 ©2018 IEEE 128

You might also like