VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Jnana Sangama, Belgaum-590018

A PROJECT PHASE I REPORT (18CSP77) ON

E-waste management using deep learning

Submitted in Partial fulfillment of the Requirements for the VII Semester of the Degree of

Bachelor of Engineering in Computer Science & Engineering

By

D VAMSI KRISHNA (1CR20CS050)

REDDY KUSHWANTH (1CR20CS156)

APARNA MOHAN (1CR20CS028)

Under the Guidance of,

Navaneetha M
Associate Professor, Dept. of CSE

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

CMR INSTITUTE OF TECHNOLOGY

#132, AECS LAYOUT, IT PARK ROAD, KUNDALAHALLI, BANGALORE-560037

i
 CMR INSTITUTE OF TECHNOLOGY
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
#132, AECS LAYOUT, IT PARK ROAD, KUNDALAHALLI, BANGALORE-560037

CERTIFICATE
Certified that the project work entitled “Intelligent task Scheduling for Datacenters” carried out by Mr.
Vamsi Krishna D, USN-1CR20CS050, Mr. Reddy Kushwanth USN-1CR20CS156, Ms. Aparna Mohan
USN-1CR20CS028, Bonafede students of CMR Institute of Technology, in partial fulfillment for the
award of Bachelor of Engineering in Computer Science and Engineering of the Visveswaraya
Technological University, Belgaum during the year 2023-2024. It is certified that all
corrections/suggestions indicated for Internal Assessment have been incorporated in the Report
deposited in the departmental library.

The project report has been approved as it satisfies the academic requirements in respect of Project work
prescribed for the said Degree.

_____________________________ _____________________________
Signature of Guide Signature of HOD
Navaneetha M Dr. Shreekanth Mooroor Prabhu
Associate Professor, Professor & HoD
Dept. of CSE, CMRIT Dept. of CSE, CMRIT
 DECLARATION

We, the students of 7th semester of Computer Science and Engineering, CMR Institute of
Technology, Bangalore declare that the work entitled "Intelligent task Scheduling for Datacenters"
has been successfully completed under the guidance of Prof. Navaneetha M Computer Science
and Engineering Department, CMR Institute of technology, Bangalore. This dissertation work is
submitted in partial fulfillment of the requirements for the award of Degree of Bachelor of
Engineering in Computer Science and Engineering during the academic year 2023 - 2024. Further
the matter embodied in the project report has not been submitted previously by anybody for the
award of any degree or diploma to any university.

Place: Bangalore

Date:05/01/2024

Team members:

D VAMSI KRISHNA (1CR20CS050) __________________

REDDY KUSHWANTH (1CR20CS156) __________________

APARNA MOHAN (1CR20CS028) __________________

 ABSTRACT

In response to the pressing demand for automated e-waste sorting, our project strategically
employs a sophisticated Convolutional Neural Network (CNN) image-recognition algorithm.
Recognizing the complexities of e-waste, we have undertaken comprehensive experiments to
optimize the performance of our system. These experiments include the exploration of diverse
image sizes, the implementation of data augmentation techniques such as rotation, and the
incorporation of background removal methodologies. Through these meticulous trials, our
objective is to enhance the training and validation accuracy of the CNN algorithm, ensuring its
effectiveness in accurately classifying and sorting various types of electronic waste. By leveraging
cutting-edge technologies like CNN, we aim to streamline and automate the sorting process,
reducing the burden on manual labor while concurrently contributing to environmental protection
and the well-being of workers involved in e-waste management.

As we advance in the project, the focus remains on creating an efficient system that aligns with
the intricacies of e-waste materials. The significance of our work extends beyond mere automation;
it underscores a commitment to sustainable practices and environmental stewardship. By
developing a robust e-waste sorting solution, we anticipate not only increasing the accuracy and
speed of the sorting process but also establishing a foundation for improved recycling strategies.
This endeavor stands as a testament to the transformative potential of technological innovation in
addressing critical environmental challenges and advancing the responsible management of
electronic waste.

iv
 ACKNOWLEDGEMENT

I take this opportunity to express my sincere gratitude and respect to CMR Institute of
Technology, Bengaluru for providing me a platform to pursue my studies and carry out my final
year project
I have a great pleasure in expressing my deep sense of gratitude to Dr. Sanjay Jain,
Principal, CMRIT, Bangalore, for his constant encouragement.
I would like to thank Dr. Shreekanth M. Prabhu, HoD, Department of Computer Science
and Engineering, CMRIT, Bangalore, who has been a constant support and encouragement
throughout the course of this project.
I consider it a privilege and honor to express my sincere gratitude to my guide Prof.
Navaneetha M, Associate Professor, Department of Computer Science and Engineering, for the
valuable guidance throughout the tenure of this review.
I also extend my thanks to all the faculty of Computer Science and Engineering who
directly or indirectly encouraged me.
Finally, I would like to thank my parents and friends for all their moral support they have
given me during the completion of this work.

v
 TABLE OF CONTENTS
Page No.
Certificate ii
Declaration iii
Abstract iv
Acknowledgement v
Table of contents vii
List of Figures vii
List of Tables viii
List of Abbreviations ix
1 INTRODUCTION
1.1 Relevance of the Project 1
1.2 Problem Statement 2
1.3 Objectives 2
1.4 Scope of the project 2
1.5 Software Engineering Methodology 3
1.6 Tools and Technologies 3
1.7 Chapter Wise Summary 3

2 LITERATURE SURVEY
2.1 Overview 4
2.2 Convolutional Neural Network 4
2.3 Comparison/Inferences 5-6
2.4 Research Gap 6

3 PROBLEM FORMULATION 7-8

4 STATUS AND ROADMAP 9
REFERENCES 10

vi
 LIST OF FIGURES

Page No.
Fig 1.1 waste growth Statistics across the globe 1
Fig 2.1 Set up for segregating waste material 3
Fig 4.1 Schedule for Phase-2 5

vii
 LIST OF TABLES

Page No.
Table 1.1: User Stories 2
Table 2.1: Comparison of various techniques 7

viii
 LIST OF ABBREVIATIONS

CNN Convolutional Neural Network

RNN Recurrent Neural networks
ReLU Rectified Linear Unit

ix
Intelligent task scheduler for Datacenters

CHAPTER 1
INTRODUCTION
In this project, a convolutional neural network (CNN) image-recognition algorithm is used
to classify e-waste into different categories with high accuracy. CNN is a deep learning
neural network that has been studied to analyze visual imagery for image classification
Automated processes to sort and separate e-waste will make e-waste recycling less labor-
intensive and will protect the health of workers.

1.1 Relevance of the Project

The project on "E-waste management using deep learning" holds significant relevance in
addressing the escalating global issue of electronic waste (e-waste). As the world becomes
increasingly digitized, the rapid proliferation of electronic devices has resulted in a surge
in discarded electronics, contributing to environmental pollution and health hazards.
Leveraging deep learning techniques in e-waste management allows for the development
of intelligent systems capable of automating the identification, categorization, and sorting
of electronic waste. This not only streamlines the recycling process but also enhances
efficiency by distinguishing between reusable components and hazardous materials. By
employing deep learning algorithms, the project aims to create a robust and scalable
solution for managing e-waste, thereby mitigating environmental impact and fostering a
sustainable approach to handling the electronic products that have reached the end of their
lifecycle.

Fig 1.1 waste growth Statistics across the globe

Dept. of CSE, CMRIT 2023-2024 Page 1

Intelligent task scheduler for Datacenters

1.2 Problem Statement

Developing an efficient and scalable e-waste management system leveraging machine
learning techniques to accurately identify, classify, and manage electronic waste items.
The aim is to create an intelligent system that can automate the sorting, categorization,
and disposal processes for various types of electronic devices, optimizing resource
allocation and minimizing environmental impact.

1.3 Objectives
 Develop advanced machine learning models, utilizing CNNs and RNNs, for precise
identification and classification of diverse electronic waste items.

 Implement feature extraction and robust data preprocessing techniques to optimize model
performance and address dataset imbalances and noise.

 Integrate IoT and sensor technologies to enhance real-time data collection, facilitating
continuous monitoring and tracking of e-waste across different management stages.

 Establish a comprehensive database and analytics platform to store, manage, and analyze
e-waste data, including types, quantities, geographical distribution, and recycling
outcomes.

 Design models and infrastructure with scalability and adaptability to accommodate future
e-waste challenges and advancements in electronic devices.

 Develop a user-friendly interface for the analytics platform, allowing stakeholders easy
access to meaningful insights for informed decision-making in e-waste management.
1.4 Scope of the project
The scope of the project on "E-waste Management Using Deep Learning" encompasses a
multifaceted approach to address the challenges associated with the growing volume of
electronic waste. The project will involve the development and implementation of
advanced machine learning models to accurately identify, classify, and sort various types
of electronic waste items. This includes exploring the application of convolutional neural
networks (CNNs) and recurrent neural networks (RNNs) for image and sensor data
analysis. The scope extends to integrating Internet of Things (IoT) devices and sensor
technologies to enhance real-time data collection and monitoring throughout the e-waste
management lifecycle.

Dept. of CSE, CMRIT 2023-2024 Page 2

Intelligent task scheduler for Datacenters

1.5 Software Engineering Methodology

Developing a software system for landslide identification involves careful consideration of
the unique challenges associated with the domain. A suitable software engineering
methodology should encompass various stages, from requirements analysis to deployment
and maintenance.

Table 1.1: User Stories Task Scheduler

1.6 Tools and Technologies

IDE: Anaconda
Programming Language: Python
Dataset: Kaggle
1.7 Chapter Wise Summary
In Chapter 1 we have Introduction of the project and the Relevance, problem statement,
objective, scope of project and Software Engineering Methodology.
In Chapter 2 we have overview of Literature Survey, sources, Convolutional Neural
Network algorithm, Comparison, Research Gap.
In Chapter 3 we have Problem Formulation.
In Chapter 4 we have Status and Roadmap, References.

Dept. of CSE, CMRIT 2023-2024 Page 3

Intelligent task scheduler for Datacenters

CHAPTER 2
LITERATURE SURVEY
2.1 Overview
The E-waste management system emphasizes on automating the sorting, categorization,
and disposal processes for various types of electronic devices, optimizing resource
allocation and minimizing environmental impact. It concentrates on high efficiency and
accuracy, thereby protecting the environment and the health of workers. With the current
documented collection and recycling rate of 17.4%, a raw material value of $10 billion
USD is recovered in an environmental sound way from e-waste globally, and 4 Mt of raw
materials could be made available for recycling.
Sources

► IEEE Xplore-MACHINE LEARNING FOR THE CLASSIFICATION AND

SEPARATION OF E WASTE.

► Application of deep learning object classifier to improve e-waste collection

planning.

► Convo Waste: An Automatic Waste Segregation Machine Using Deep Learning.

2.2 Convolutional Neural Network

Input layer: The input layer of the CNN model is fed with images of various common e-
waste items

Hidden layer: multiple hidden layers that are fully connected.

Rectified Linear Unit (ReLU) activation function is chosen for each convolutional layer.

Output layer: A dense layer with 128 neurons is employed to receive all the outputs
from the previous neurons

The activation function for the output layer is not explicitly mentioned, but it is common
to use softmax activation for multi-class classification problems

CNN creates a pattern from the input image and matches it with the existing images
already trained and present in the database.

Experimental Setup (if research):

o Dataset: The data set of the images is manually generated.
o Hardware: A pan-and-tilt set (SPT200 Pan & Tilt Kit-25 Tooth Spline with 2000
Series Dual Mode Servo 25-2 by Servocity) was used as the sorting robot

Dept. of CSE, CMRIT 2023-2024 Page 4

Intelligent task scheduler for Datacenters

Fig 2.1 Set up for segregating waste material[2]

Major Results:
o Generation of patterns from images
o Pattern matching and segregating the images

Shortcomings:
o Size of the image
o Rotation of image (20% original images and 80% images rotated)
o Increased efficiency on background removal

Future Work:
o They should expand the dataset and retrain the model
o Conveyer belt can be added for better hardware implementation

2.3 Comparison

The comparison of various research papers is presented in Table 2.1.

Features Advantages Limitations

1. Zhou, Ethan P. - 3-layer Convolutional - Image size Achieved high
"Machine Learning Neural Network (CNN) limitation. - Rotation training and
For The architecture with ReLU of images not validation accuracy
Classification And activation function. adequately addressed. (96.9% and 93.9%). -
Separation Of E- - Efficiency in Successful pattern
Waste." background removal matching and
can be improved. segregation of
images.
2. Nowakowski, Utilized CNN and Faster Limited dataset size. - Demonstrated use of
Piotr, and Teresa R-CNN for waste object Faster R-CNN deep learning for

Dept. of CSE, CMRIT 2023-2024 Page 5

Intelligent task scheduler for Datacenters

Pamuła. "Application classification. - Image showed lower waste collection

of deep learning size: 128 x 128 pixels. accuracy. planning. - Achieved
object classifier to accuracy range of 90-
improve e-waste 97%.
collection planning."
Nafiz, Md Shahariar, Used ConvoWaste, a pre- Variations in lighting Achieved high
et al. "Convowaste: trained CNN, for waste conditions affecting accuracy (98%). -
An Automatic Waste classification. - image classification. - Remote control
Segregation Machine Integration of ultrasonic Variable time for through an android
Using Deep sensor and servo motor- waste detection app. - Integration of
Learning. based system for waste among different waste communication
segregation. - Dual-band types. technology for waste
GSM-based level alerts.
communication for waste
level notification.
Table 2.1 Comparison of various techniques

2.4 Research Gap

Despite advancements in e-waste management, a significant research gap exists in
the application of deep learning techniques for enhancing the automated sorting and
categorization processes. While traditional methods have been explored, the integration of
deep learning algorithms, such as convolutional neural networks (CNNs) and recurrent
neural networks (RNNs), specifically tailored for e-waste recognition, remains
underexplored. Addressing this gap is essential for achieving higher accuracy and
efficiency in identifying diverse electronic components, thereby improving the overall
efficacy of e-waste recycling systems. Furthermore, limited studies have focused on the
scalability and real-world implementation challenges of deep learning models in large-scale
e-waste facilities, representing another critical area for exploration and innovation. Closing
these gaps will contribute significantly to the advancement of sustainable and
technologically advanced e-waste management practices.

Dept. of CSE, CMRIT 2023-2024 Page 6

Intelligent task scheduler for Datacenters

CHAPTER 3
PROBLEM FORMULATION

The pervasive issue of electronic waste (e-waste) poses a multifaceted challenge, impacting
human health, environmental sustainability, and effective resource management. As our
dependence on electronic devices continues to grow, the need for sophisticated tools to
precisely detect and monitor e-waste becomes increasingly urgent. The core challenge lies
in the rapid identification of new instances of e-waste accumulation and the improvement
of existing inventories. These tasks are essential for the implementation of effective
recycling strategies and timely responses to emerging environmental risks.

The critical aspect of this problem is the dynamic nature of e-waste generation, driven by
rapid technological advancements and a shortening product lifecycle. As consumers
frequently upgrade their electronic devices, the rate of e-waste accumulation accelerates,
creating a pressing need for advanced detection methods. Traditional monitoring systems
often struggle to keep pace with the evolving landscape of electronic consumption, leading
to gaps in identifying and managing e-waste hotspots.

Moreover, the complexity of e-waste, which includes a wide range of materials and
components, further complicates the detection and monitoring process. Differentiating
between various types of electronic devices and understanding their composition is a
formidable task that requires sophisticated tools and expertise. Both experts and non-
experts in e-waste management face significant difficulties in addressing these challenges,
hampering the development and implementation of effective solutions.

The consequences of inadequate e-waste management are far-reaching. Human health risks
arise from the improper disposal and dismantling of electronic devices, as hazardous
materials can leach into the environment, contaminating soil and water sources.
Additionally, the environmental impact extends to the depletion of valuable resources and
the exacerbation of climate change through the release of greenhouse gases during the
incineration of e-waste.

Dept. of CSE, CMRIT 2023-2024 Page 7

Intelligent task scheduler for Datacenters

To address these issues comprehensively, there is a critical need for advancements in

technology and methodologies for e-waste detection and monitoring. The development of
innovative sensors, artificial intelligence algorithms, and remote sensing technologies
could enable more precise and real-time identification of e-waste accumulation. Integration
of these tools into existing waste management systems can enhance their efficiency and
responsiveness, allowing for proactive measures in recycling and minimizing
environmental risks.

In conclusion, the e-waste challenge requires a concerted effort to develop and implement
sophisticated tools for the precise detection and monitoring of electronic waste. This
involves overcoming the dynamic nature of e-waste accumulation, the diverse composition
of electronic devices, and the limitations of current monitoring systems. By addressing
these issues, we can pave the way for effective recycling strategies, mitigate environmental
risks, and promote sustainable resource management in the face of the ever-growing
electronic consumption landscape.

Dept. of CSE, CMRIT 2023-2024 Page 8

Intelligent task scheduler for Datacenters

CHAPTER 4
STATUS AND ROADMAP
Having concluded Phase 1 of our E-waste Management project, we have successfully
conducted an extensive Literature Survey, delving into numerous research papers focused
on sorting and categorization methods. The comprehensive review has equipped us with
valuable insights into the latest advancements and methodologies in the domain of
electronic waste management. Specifically, we have identified and studied various
algorithms pivotal for the efficient sorting and categorization of electronic waste. This
knowledge serves as a solid foundation for the subsequent phases of our project, providing
a roadmap for the implementation of cutting-edge technologies to enhance the recycling
process.

Moving forward, the project transitions into Phase 2, where our focus shifts from theoretical
understanding to practical implementation. Python will be the primary programming
language utilized in this phase, and we plan to incorporate advanced algorithms such as
Convolutional Neural Networks (CNN). These algorithms will play a crucial role in
automating the sorting and categorization processes, contributing to the overall efficiency
of e-waste management. As we embark on the implementation stage, the insights gained
from the Literature Survey will guide our development efforts, ensuring a robust and
effective system for the sustainable management and recycling of electronic waste.

Fig 4.1 Schedule for Phase-2

Dept. of CSE, CMRIT 2023-2024 Page 9

Intelligent task scheduler for Datacenters

REFERENCES
[1] Forti V, Balde CP, Kuehr R, Bel G. “The global e-waste monitor 2020: Quantities, flows
and the circular economy potential”. Available from: https://ewastemonitor.info/gem-
2020/
[2] Earth911. “20 staggering e-waste facts in 2021” [Internet].Available from:
https://earth911.com/eco-tech/20-e-wastefacts/
[3] Dias P, Bernardes AM, Huda N. “Ensuring best E-waste recycling practices in developed
countries: An Australian example”. J Clean Prod. 2019 Feb 1; 209: 846-54. http://doi.org/
10.1016/j.jclepro.2018.10.306.
[4] Kang HY, Schoenung JM. “Economic analysis of electronic waste recycling: modeling the
cost and revenue of a materials recovery facility in California”. Environ Sci Technol. 2006
Mar 1;40(5):1672-80. doi: 10.1021/es0503783. PMID: 16568786.

[5] Agrawal, S.R., Mittal, D., 2017. Need of an online e-waste market in India. Int.
J.Environ. Waste Manag. 19, 21–41. https://doi.org/10.1504/IJEWM.2017.083560.
[6] Asekol, 2019. Collection points < ASEKOL | Asekol [WWW Document]. URL
https://www.asekol.cz/en/asekol/collection-points/ (accessed 8.5.19).
[7] Aziz, F., Arof, H., Mokhtar, N., Mubin, M., Abu Talip, M.S., 2015. Rotation
invariant
[8] bin detection and solid waste level classification. Measurement 65, 19–
28.https://doi.org/10.1016/j.measurement.2014.12.027.Bin Aziz, F., Arof, H.,
Mokhtar, N., M. Shah, N., Khairuddin, A., Hanafi, E., Sofian Abu Talip, M., 2018.
Waste level detection and HMM based collection scheduling of multiple bins.
PLOS ONE 13, e0202092. https://doi.org/10.1371/journal.

Dept. of CSE, CMRIT 2023-2024 Page 10