AUTONOMOUS SOLAR-POWERED BOAT FOR

PLASTIC WASTE COLLECTION IN OCEANS

A PROJECT REPORT

Submitted by

R.KAVIVANAN 420721106301

N.NAVEEN 420721106035

In partial fulfilment for the award of the degree

of

BACHELOR OF ENGINEERING

IN

ELECTRONICS AND COMMUNICATION ENGINEERING

C.K. COLLEGE OF ENGINEERING & TECHNOLOGY

CUDDALORE

ANNA UNIVERSITY:: CHENNAI 600 025

MAY 2025
 ANNA UNIVERSITY: CHENNAI 600 025

BONAFIDE CERTIFICATE

Certified that this project report “ AUTONOMOUS SOLAR-POWERED

BOAT FOR PLASTIC WASTE COLLECTION IN OCEANS “ is the bonafide
work of “R.KAVIVANAN (420721106301), N.NAVEEN (420721106035)” who
carried out the project work under my supervision.

SIGNATURE SIGNATURE

Dr.M.Arulaalan, M.E., Ph.D., Mr.G.Manikannan, M.Tech.,(Ph.D.)

HEAD OF THE DEPARTMENT SUPERVISOR

Department of Electronics & Department of Electronics &

Communication Engineering Communication Engineering

C.K. College of Engineering & C.K. College of Engineering &

Technology, Cuddalore. Technology, Cuddalore.

Submitted for the project Viva-Voce Examinations held on .............................

INTERNAL EXAMINER EXTERNAL EXAMINER

ii
 ACKNOWLEDGEMENT

We wish to extend our heartfelt thanks to Mr. C.K. RANGANATHAN,

Chairman and Ms. AMUDHAVALLLI RANGANATHAN, Executive Director,
CK Group of Institutions, Cuddalore, for providing the facilities for successful
completion of the project work.

Words and inadequate in offering our sincere thanks and gratitude to our
respect Principal Dr. S. SARAVANAN, M.E., Ph.D., for his amazing support and
continuous encouragement for the successful completion of our project work.

We also thank to Dr. M. ARULAALAN, M.E., Ph.D., Professor, Head of the

Department, Electronics and Communication Engineering, for his encouragement
and permitting us to utilize all the facilities available in the department to complete
our project work.

We wish to express the sense of gratitude and we would like to thank our
Guide Mr. G. MANIKANNAN, M.TECH. (Ph.D.), Assistant Professor,
Department of Electronics and Communication Engineering, for his encouragement
and valuable support, constructive remarks rendered in this project.

We are thankful to our project coordinator Mr. THEEBAN

CHAKARAVARTHY, M.TECH., (Ph.D.), Assistant Professor, Department of
Electronics and Communication Engineering, for his kind help and support
throughout this project.

We owe thanks to all my faculty members and staff members, friends and our
parents who rendered in our endeavor.

iii
 ABSTRACT

Plastic pollution in water bodies, especially in oceans and lakes, poses severe
threats to marine ecosystems and public health. Manual cleanup is often
inefficient, labor-intensive, and lacks real-time adaptability. This project
introduces an autonomous, solar-powered robotic boat designed for intelligent
plastic waste detection and collection using AI-based image processing on a
local system. The boat features a floatable platform driven by DC motors
controlled via an L298N motor driver and navigated using GPS. An ESP32-
CAM module captures real-time images of the water surface and transmits them
to a local Python Flask server. There, a CNN model built with TensorFlow
analyzes the images to detect the presence of plastic waste. Upon detection, the
system triggers a servo-operated collection gate via the ESP8266
microcontroller to collect the waste. Additionally, an ultrasonic sensor is
integrated to detect nearby obstacles such as walls or floating objects and
prevent collisions through basic avoidance logic. The boat operates
independently, using solar-charged batteries, and returns to base automatically
based on GPS coordinates. The entire process is handled locally without
reliance on cloud platforms, enabling offline operation in remote areas. The
proposed system provides a low-cost, energy-efficient, and scalable solution for
mitigating plastic pollution in aquatic environments, reducing the need for
manual intervention and enhancing sustainability in waste management.

iv
 TABLE OF CONTENTS
CHAPTER TITLE PAGE
NO. NO.
ABSTRACT iv
LIST OF FIGURES viii
LIST OF ABBREVIATION ix

1 INTRODUCTION 1
1.1 INTRODUCTION TO IOT 1
1.2 IOT ARCHITECTURE 2
1.2.1 PERCEPTION LAYER 2
1.2.2 NETWORK LAYER 2
1.2.3 MIDDLEWARE LAYER 3
1.2.4 APPLICATION LAYER 4
1.2.5 SECURITY LAYER 4
1.2.3 CLOUD/EDGE COMPUTING LAYER 4
1.3 IOT APPLICATIONS 5
1.3.1 SMART HOME AUTOMATION 5
1.3.2 HEALTHCARE 5
1.3.3 INDUSTRIAL IOT 5
1.3.4 SMART AGRICULTURE 6
1.3.5 SMART CITIES 6
1.3.6 RETAIL AND SUPPLY CHAIN 6
1.4 IOT IN AUTONOMOUS SOLAR-POWERED 7
BOAT FOR PLASTIC WASTE COLLECTION
1.5 AI IN SMART AUTONOMOUS WASTE 8
COLLECTION ROBOT
1.6 ORGANIZATION OF THESIS 10
2 LITERATURE REVIEW 12
3 EXISTING SYSTEM 19

vii
 3.1 INTRODUCTION 19
3.2 EXISTING METHODOLOGY 20
4 PROPOSED SYSTEM 22
4.1 INTRODUCTION 22
4.2 PROPOSED METHODOLOGY 22
4.2.1 HARDWARE SETUP 24
4.2.2 CONNECTIVITY 25
4.2.3 CENTRALIZED CONTROL SYSTEM 25
4.2.4 REAL-TIME BLOCKAGE DETECTION 25
AND CLEANING

4.3 HARDWARE COMPONENTS 26

4.3.1 NODEMCU 26
4.3.2 ULTRASONIC SENSOR 27
4.3.3 ESP32-CAM 28
4.3.4 GPS MODULE 29
4.3.5 MOTOR DRIVER 29
4.3.6 MOTOR DRIVER 27
4.3.7 SOLAR PANEL 31
4.3.8 DC MOTOR 29
4.3.9 SERVO MOTOR 32
4.4 SOFTWARE USED 33
5 RESULT & DISCUSSION 36
5.1 INTRODUCTION 36
5.2 PROJECT PROTOTYPE 36
5.3 PROTOTYPE DESCRIPTION 37
5.4 WORKING PRINCIPLE 37
5.5 PROJECT OUTPUT 38
5.6 HARDWARE ANALYSIS RESULT 37
5.6.1 NODEMCU FUNCTIONALITY 37
5.6.2 POWER SUPPLY STABILITY 37
5.7 OUTPUT DISCUSSION 37

vii
6 CONCLUSION & FUTURE SCOPE 39
6.1 CONCLUSION 39
6.2 FUTURE SCOPE 40
7 REFERENCE 41

vii
 LIST OF FIGURES
FIGURE TITLE PAGE
NO. NO.
1.1 INTERNET OF THINGS 1
1.2 IOT ARCHITECTURE 3
1.3 IOT IN STREET LIGHT MONITORING 8
1.4 ENERGY MANAGEMENT IN IOT 9
1.5 SMART ENVIRONMENT MONITORING SYSTEM 10
3.1 EXISTING BLOCK DIAGRAM 22
4.1 BLOCK DIAGRAM OF PROPOSED SYSTEM 25
4.2 NODEMCU 28
4.3 LDR SENSOR 29
4.4 LIGHT EMITTING DIODE 30
4.5 IR SENSOR 30
4.6 VOLTAGE SENSOR 31
4.7 POWER SUPPLY BOARD 32
5.1 HARDWARE IMPLEMENTATION 35
5.2 WORKING MODEL 36
5.2 OUTPUT OF THE PROJECT 37

viii
 LIST OF ABBREVIATION

IEEE Institute of Electrical and Electronics Engineers

IoT Internet of Things
WAN Light Emitting Diode
MCU Micro Controller Unit

GPS Personal Identification Number

SMS Short Message Service

LDR Light dependent resistor

ix
 CHAPTER - 1

INTRODUCTION

1.1 Introduction to IoT:

The Internet of Things (IoT) represents a transformative force in the realm

of technology, connecting billions of devices worldwide to create a network of
interconnected systems. These devices, ranging from smart home appliances and
wearable gadgets to industrial machinery and urban infrastructure, collect and
exchange data seamlessly over the internet. This interconnectedness enables
enhanced automation, real-time monitoring, and data-driven decision-making
across various sectors. By harnessing the power of IoT, organizations can
optimize operations, improve efficiency, and deliver personalized experiences to
users. However, this expansive connectivity also brings forth challenges, such as
data privacy concerns and security vulnerabilities, requiring robust solutions to
ensure the integrity and security of IoT ecosystems. Despite these challenges, the
potential of IoT to revolutionize industries, drive innovation, and improve the
quality of life for individuals is undeniable, making it a pivotal technology for the
future.

Fig 1.1 Internet of Things

1
 However, as IoT continues to expand, it also presents complex challenges
that need to be addressed to ensure its sustainable growth. Issues such as data
privacy, cybersecurity threats, and interoperability remain significant concerns
that require comprehensive strategies and robust frameworks to mitigate risks and
safeguard IoT ecosystems. Despite these challenges, the transformative potential
of IoT in driving digital transformation, fostering innovation, and creating
smarter, more connected environments is immense, making it a key driver of the
fourth industrial revolution.

1.2 IoT Architecture:

The architecture of the Internet of Things (IoT) typically consists of several

layers that work together to enable seamless connectivity, data collection,
processing, and application deployment. Here's a breakdown of the typical IoT
architecture:

1.2.1 Perception Layer:

Sensors: These are physical devices that collect data from the
environment. They can measure various parameters such as temperature,
humidity, light, motion, and more.
Actuators: These devices perform actions based on the data received from
sensors, such as turning on/off a switch or adjusting a valve.
1.2.2 Network Layer:
Connectivity: This layer facilitates the communication between sensors,
actuators, and other devices. It includes various communication protocols
like Wi-Fi, Bluetooth, Zigbee, Lora WAN, and cellular networks (4G/5G).

Gateways: These devices act as intermediaries between the IoT devices

and the cloud or central server, aggregating data, performing edge
computing, and ensuring secure communication.

2
 Fig 1.2 IoT Architecture

1.2.3 Middleware Layer:

Data Processing and Storage: This layer handles data filtering,
aggregation, and storage. It preprocesses the raw data collected from
sensors before transmitting it to the cloud for further analysis.
Device Management: It includes functionalities for device provisioning,
configuration, monitoring, and firmware updates, ensuring the smooth
operation of IoT devices.

3
1.2.4 Application Layer:
IoT Applications: These are the end-user applications or services that
utilize the processed data to provide insights, trigger actions, or offer
valuable services. Examples include smart home systems, industrial
automation, healthcare monitoring, and environmental monitoring
solutions.
User Interface: This component provides a graphical interface for users to
interact with the IoT system, visualize data, set preferences, and control
connected devices.

1.2.5 Security Layer:

Authentication and Authorization: Ensures only authorized devices and
users can access the IoT network and data.
Data Encryption: Protects the data transmitted between devices,
gateways, and the cloud from unauthorized access.
Security Monitoring and Incident Response: Monitors the IoT ecosystem
for any security threats or anomalies and takes appropriate actions to
mitigate risks.

1.2.6 Cloud/Edge Computing Layer:

Cloud Infrastructure: Provides scalable computing resources for data
storage, processing, and analytics. It enables the storage of large volumes
of data collected from IoT devices and facilitates advanced analytics and
machine learning algorithms to derive actionable insights.
Edge Computing: This approach involves processing data closer to its
source (at the edge of the network), reducing latency, bandwidth usage, and
ensuring real-time data processing and response. Edge devices can perform
initial data analysis, filtering, and actuation locally before sending relevant
data to the cloud.

4
1.3 IoT Applications:

1.3.1 Smart Home Automation:

Home Security: IoT-enabled cameras, motion sensors, and smart locks

provide enhanced security monitoring and control.
Energy Management: Smart thermostats and lighting systems optimize
energy consumption based on user behavior and environmental conditions.
Appliance Control: Remote control and monitoring of home appliances for
improved efficiency and convenience.

1.3.2 Healthcare:

Remote Patient Monitoring: Wearable devices and sensors track vital

signs and health parameters, enabling continuous remote monitoring of
patients.
Telemedicine: IoT facilitates virtual consultations and remote medical
services, improving accessibility to healthcare.
Medication Management: Smart pill dispensers and medication adherence
systems help patients manage their medication schedules effectively.

1.3.3 Industrial Internet of Things (IIoT):

Predictive Maintenance: IoT sensors monitor equipment health and

performance, predicting potential failures and optimizing maintenance
schedules.
Asset Tracking: Real-time tracking and monitoring of assets, inventory,
and logistics operations to improve efficiency and reduce costs.
Smart Manufacturing: Automation and optimization of manufacturing
processes through real-time data analytics and control systems.

5
1.3.4 Smart Agriculture:

Precision Farming: IoT sensors and drones collect data on soil moisture,
temperature, and crop health, enabling farmers to make informed decisions
and optimize crop yields.
Livestock Monitoring: Wearable devices monitor the health and
behaviour of livestock, facilitating early disease detection and
management.
Automated Irrigation Systems: IoT-enabled irrigation systems adjust
water usage based on environmental conditions and plant needs,
conserving water and enhancing crop growth.

1.3.5 Smart Cities:

Traffic Management: IoT sensors and cameras monitor traffic flow and
congestion, enabling smart traffic management and optimization.
Waste Management: Smart waste bins and monitoring systems optimize
waste collection routes and schedules based on fill-level sensors,
improving efficiency and reducing costs.
Environmental Monitoring: Sensors measure air quality, noise levels,
and other environmental parameters to monitor and manage urban
environmental conditions.

1.3.6 Retail and Supply Chain:

Inventory Management: Real-time tracking and monitoring of inventory

levels, reducing stockouts and overstock situations.
Customer Experience: IoT-enabled devices, such as smart mirrors and
interactive kiosks, enhance the in-store shopping experience by providing
personalized recommendations and interactive services.

6
1.4 IoT in Autonomous Solar-Powered Boat for Plastic Waste Collection:

The IoT framework in the autonomous solar-powered boat project is

fundamental for real-time monitoring, communication, and control of the
boat’s operations during plastic waste collection in oceans. IoT enables
seamless integration of sensors, controllers, and cloud services to optimize
efficiency and autonomy.

Sensor Integration: The boat is equipped with IoT-enabled devices

including GPS modules for location tracking, ultrasonic sensors for
obstacle detection, water quality sensors, and ESP32-CAM for live video
and image capture. These sensors continuously gather data on the boat’s
position, surroundings, plastic waste presence, and environmental
conditions.

Data Transmission: Sensor data is processed locally by a microcontroller

such as ESP32 or Raspberry Pi and transmitted wirelessly to a cloud-based
IoT platform using Wi-Fi, 4G/LTE, or LoRaWAN networks. This ensures
reliable and near real-time data communication even in remote ocean areas

Remote Monitoring and Control: The cloud platform provides a user

interface accessible via web or mobile apps, allowing operators to remotely
monitor the boat’s status, battery levels, sensor readings, and mission
progress. Alerts are generated automatically when obstacles, plastic waste,
or operational anomalies are detected.

Autonomous Navigation and Coordination: IoT enables the boat to

receive updated navigation commands based on live data analytics and
environmental inputs. Multiple boats can communicate within an IoT mesh
network to coordinate coverage areas and avoid collisions, improving
overall collection efficiency.

7
 Energy Management: Solar panel performance and battery health are
continuously monitored via IoT sensors. This data helps optimize power
usage and ensures uninterrupted operation, triggering alerts when energy
levels are low to prompt docking or recharging.

Fig 1.3 Smart Autonomous Waste collection Robot

1.5 AI in Smart Autonomous Waste collection Robot:

The Artificial Intelligence (AI) component is central to the

autonomous operation of the solar-powered boat for plastic waste
collection. AI-powered computer vision models, typically convolutional
neural networks (CNNs), are deployed to analyze real-time images
captured by the onboard camera to detect and classify plastic waste
floating on the ocean surface. These models are trained on diverse datasets
containing images of different types, shapes, and sizes of plastic debris,
enabling accurate identification even under varying lighting and weather
conditions. Once plastic waste is detected, AI algorithms process location
and spatial data to guide the boat’s navigation system towards the target
8
waste. Path planning is enhanced by AI techniques that fuse data from
multiple sensors like GPS, ultrasonic sensors, and inertial measurement
units, allowing the boat to avoid obstacles and navigate complex marine
environments safely. AI also manages energy consumption by optimizing
solar panel usage and battery charging cycles for continuous operation.
Additionally, AI facilitates communication between multiple boats in a
swarm robotics setup, enabling coordinated coverage of large ocean areas
for efficient waste collection. Real-time data analytics powered by AI help
monitor environmental conditions and improve the system’s adaptability.
Through continuous learning and updates, the AI system enhances
detection accuracy and navigation efficiency over time, making the boat
smarter and more effective in combating ocean plastic pollution.

9
1.6 ORGANIZATION OF THESIS:

Chapter 1: Introduction

 Introduction to IoT and its significance in urban infrastructure management

 Overview of street lighting systems and their importance in urban
environments
 Objectives of the project and scope of the thesis

Chapter 2: Literature Review

 Review of existing literature and research studies related to IoT-based

street light monitoring and fault detection systems
 Examination of various IoT technologies, sensors, and communication
protocols used in similar projects
 Analysis of challenges and opportunities in implementing IoT solutions for
street light management

Chapter 3: Existing System

 Overview of the current methods and technologies used for monitoring and
maintaining street lights
 Evaluation of the limitations and shortcomings of traditional street light
management systems
 Case studies or examples of existing IoT-based street light monitoring
systems
Chapter 4: Proposed System

 Description of the proposed IoT-based street light fault detection system

 Explanation of the architecture, components, and functionality of the
system
 Details of the sensors, communication protocols, and data processing
techniques employed

10
  Discussion on how the proposed system addresses the limitations of
existing systems

Chapter 5: Results and Discussion

 Presentation of experimental results and findings from the implementation

of the proposed system
 Analysis of the system's performance, including accuracy, efficiency, and
reliability
 Discussion on the implications of the results and their significance for
street light management practices

Chapter 6: Conclusion and Future Scope

 Summary of the key findings and contributions of the thesis

 Reflection on the effectiveness of the proposed system in addressing the
research objectives
 Concluding remarks on the importance of leveraging IoT technology for
improving urban infrastructure sustainability and efficiency

11
 CHAPTER – 2
LITERATURE REVIEW
Title: AI-Based Autonomous Boat for Plastic Waste Collection in Water
Bodies
Author & Year: S. K. Singh, A. Kumar, and P. K. Sahu, 2023
Description: This paper discusses the design and implementation of an AI-
enabled autonomous boat aimed at detecting and collecting plastic waste from
water bodies. The system employs machine learning algorithms to classify
waste using onboard cameras and navigates using GPS and ultrasonic sensors.
The boat is solar-powered to enhance operational efficiency and sustainability.
Real-time data on waste locations is transmitted to a centralized monitoring
platform for further analysis.
The research highlights challenges in marine waste collection and proposes
adaptive strategies for different water conditions. The robot’s hardware includes
a microcontroller interface for motor control and sensor data acquisition. It uses
convolutional neural networks (CNN) to differentiate plastic waste from organic
debris. The study also presents field trials in controlled water bodies,
showcasing over 85% accuracy in detection and successful autonomous
collection missions. The authors address the energy constraints by integrating
solar panels and battery management systems.
Additionally, the system incorporates fail-safe mechanisms to avoid collisions
with aquatic animals. The paper concludes with a discussion on scalability for
urban waterway cleanup programs.

12
Title: Deep Learning-Based Plastic Waste Detection Using ESP32-CAM
and TensorFlow
Author & Year: J. Park, M. Lee, and H. Kim, 2022
Description: This study presents a low-cost embedded system integrating
ESP32-CAM with TensorFlow-based convolutional neural networks for plastic
waste detection in aquatic environments. The system captures images and
processes them locally to detect plastics with high accuracy, enabling real-time
decisions for robotic waste collection. The work demonstrates the feasibility of
deploying lightweight AI models on edge devices and discusses trade-offs
between accuracy and computational resources. The authors detail the training
dataset comprising over 10,000 annotated images of various plastic types under
different lighting and water conditions.
Techniques such as data augmentation and transfer learning were used to
improve model robustness. The system’s latency for image capture to
classification is under 300 ms, suitable for dynamic environments. The paper
discusses energy-efficient coding practices to extend battery life in field
deployments. Experimental results show precision and recall rates exceeding
90%, outperforming previous works using heavier processing units. Challenges
such as occlusions, reflections, and varying water turbidity are addressed
through image preprocessing steps. The authors propose future integration with
multi-sensor fusion for enhanced reliability.

Title: Solar-Powered Autonomous Robots for Water Pollution Cleanup

Author & Year: R. Sharma and V. Gupta, 2024
Description: This paper explores the development of solar-powered
autonomous robots designed for water pollution cleanup, focusing on renewable
energy integration. The robots utilize sensors and AI algorithms to detect and
collect pollutants, including plastics and chemical contaminants. The study

13
evaluates energy management strategies to maximize operational time and
discusses the environmental impact reduction achieved by such autonomous
systems. The hardware design includes photovoltaic panels optimized for
aquatic use and rechargeable lithium-ion batteries. The control system adapts
cleaning operations based on real-time pollution density maps generated via
onboard sensors.
Autonomous navigation is supported by GPS and obstacle avoidance sensors to
ensure safe operation in complex water environments. The robot communicates
with a cloud-based monitoring system for data logging and remote control.
Field tests demonstrate continuous operation for up to 10 hours under partial
sunlight. The authors analyze lifecycle costs and potential environmental
benefits, including reduced human intervention and improved water quality.
Limitations related to weather dependency and maintenance are discussed,
along with proposed hybrid power solutions.

Title: Design and Implementation of an IoT-Based Plastic Waste Collection

Boat
Author & Year: M. N. Islam, S. S. Chowdhury, and M. A. Rahman, 2023
Description: The authors propose an IoT-based system that allows remote
monitoring and control of plastic waste collection boats. The system integrates
sensors for water quality, GPS for navigation, and communication modules for
data transmission. The study highlights the benefits of IoT in improving waste
management in aquatic environments through enhanced situational awareness
and automation.
The boat uses multiple sensor arrays to detect waste density, water pH, and
temperature to optimize collection routes. The communication infrastructure
employs low-power wide-area network (LPWAN) protocols for long-range data
transmission. The control unit processes sensor inputs and sends motor
commands via a microcontroller. Real-time visualization dashboards provide

14
operators with status updates and environmental metrics. The system supports
geofencing to prevent navigation into restricted or hazardous zones.
Experimental deployments in urban rivers showed a 30% improvement in waste
collection efficiency compared to manual methods. The paper discusses security
challenges and proposes encryption methods for IoT data streams. Future work
includes integrating machine learning for predictive route optimization.

Title: Obstacle Avoidance and Navigation in Autonomous Boats Using

Ultrasonic Sensors and GPS
Author & Year: L. Zhang, Y. Wang, and Q. Zhou, 2021
Description: This research investigates autonomous navigation strategies for
boats using ultrasonic sensors for obstacle detection and GPS for path planning.
The paper describes sensor fusion techniques and real-time control algorithms
enabling safe and efficient movement in cluttered aquatic environments.
Experimental results validate the robustness of the system under varying
conditions. The ultrasonic sensors detect obstacles within a 5-meter radius,
providing timely data for evasive maneuvers. The navigation module uses GPS
waypoints and dynamically adjusts paths based on obstacle proximity and water
current data.
The control system implements PID controllers for smooth velocity and
direction adjustments. The authors simulate various scenarios including static
and moving obstacles, with collision avoidance success rates over 95%. The
study also addresses sensor noise filtering and calibration methods. The work
contributes to safer deployment of autonomous boats in populated waterways
and environmental monitoring. Recommendations for integrating additional
sensor types, such as LiDAR, are included for future enhancements.

15
Title: IoT Based Smart Monitoring System for Sewage Workers with Two-
Way Communication
Author & Year: A. Vellingiri, K. Dharni, M. Arunadevi, R. L. Aravind Lal,
2020
Description: This study presents an IoT-enabled smart monitoring system
designed to ensure the safety of sewage workers. The system incorporates gas
sensors, wearable devices, and two-way communication modules to monitor
hazardous gases and worker health remotely. The framework emphasizes real-
time alerts and remote intervention to reduce occupational risks in sewage
management, which could be adapted for autonomous drainage cleaning robots.
Gas sensors detect toxic gases such as methane, hydrogen sulfide, and carbon
monoxide to prevent worker exposure.
Wearable health monitors track vital signs, including heart rate and body
temperature, transmitting data to a central control unit. Two-way
communication allows emergency instructions and status updates. The system
includes GPS tracking for location monitoring. Field trials in urban sewage
systems demonstrated reduced accident rates and improved emergency response
times. The authors suggest integrating robotic automation with this safety
framework to replace manual cleaning. Challenges such as network reliability
and sensor calibration in harsh environments are addressed.

Title: Low-Cost Embedded Systems for Environmental Monitoring: A

Review
Author & Year: P. Patel and S. Mehta, 2022
Description: This review paper surveys various low-cost embedded platforms
used for environmental monitoring, focusing on air and water quality sensors,
wireless communication, and data processing techniques. It discusses the
challenges of deploying affordable solutions in remote and harsh environments

16
and highlights innovations in sensor technology and power management
suitable for robotic environmental applications.
Platforms such as Arduino, Raspberry Pi, and ESP8266/ESP32 are compared in
terms of computational power, energy consumption, and ease of integration.
The authors emphasize modular designs enabling customization for specific
monitoring needs. Wireless protocols including Zigbee, LoRa, and Wi-Fi are
analyzed for their applicability in different geographic and infrastructural
conditions.
Power-saving techniques such as duty cycling and solar charging are explored.
The paper also reviews data analytics methods, from basic threshold alarms to
advanced machine learning-based anomaly detection. Case studies on river
pollution monitoring and air quality assessment demonstrate real-world
applicability. The review concludes by outlining gaps in existing technologies,
including sensor accuracy and long-term durability.

Title: Autonomous Surface Vehicles for Environmental Cleanup: A Survey

Author & Year: H. K. Verma and R. Singh, 2023
Description: This comprehensive survey covers the state-of-the-art autonomous
surface vehicles (ASVs) used in environmental cleanup operations, including
waste collection and pollution detection. The paper categorizes different
sensing, navigation, and power systems, providing insights into design trade-
offs and future research directions.
The authors examine sensor suites such as optical cameras, sonar, LiDAR, and
chemical sensors for detecting diverse pollutants. Navigation techniques
including waypoint-based routing, dynamic obstacle avoidance, and swarm
coordination are discussed. The survey highlights power sources from batteries
to hybrid solar-fuel cells, emphasizing the need for sustainability.
Control architectures are reviewed, ranging from reactive to model-predictive
control schemes. The survey also analyzes communication protocols enabling

17
real-time remote monitoring and data sharing. Challenges in harsh
environmental conditions like waves, weather, and biofouling are discussed.
The paper identifies emerging trends such as AI-driven decision making and
autonomous docking for maintenance. Practical deployment cases and testing
methodologies are summarized to guide future innovations.

18
 CHAPTER – 3

EXISTING SYSTEM

3.1 INTRODUCTION

In most cities and towns, the current drainage maintenance system is

heavily dependent on manual labor, where workers are deployed to clean
clogged drains and sewers using basic tools such as rods, hooks, and buckets.
These methods are not only inefficient and time-consuming, but they also
expose workers to extreme health hazards, including direct contact with
harmful sewage, toxic gases like methane and hydrogen sulfide, and the risk of
physical injury in confined spaces.

 Some semi-automated mechanical solutions like suction pumps and

jetting machines have been introduced in urban areas, but they still
require significant human supervision and operation, and are limited
in detecting blockages deep within drainage pipelines.

 Furthermore, real-time monitoring systems are rarely integrated in

these traditional setups, leading to delayed responses during emergencies
like flooding or gas leaks.

 Moreover, the lack of intelligent sensing or alert systems means that

potential hazards are identified only after damage has occurred. The
absence of live surveillance, predictive maintenance, or autonomous
action further limits the effectiveness of existing methods.

 Thus, there is a critical need for an automated, sensor-driven, and IoT-

enabled solution that can operate independently, detect hazards early,
and reduce the need for human intervention in hazardous environments.

19
  Plastic pollution in water bodies such as lakes, rivers, and oceans is a
growing global concern, contributing to severe environmental
degradation, marine ecosystem damage, and threats to aquatic life.
Despite several government initiatives and awareness campaigns, the
accumulation of plastic waste continues to rise due to poor waste disposal
practices and lack of continuous monitoring systems.

 Traditional methods for cleaning water bodies involve manual labor

using nets or boats, which are inefficient, time-consuming, and incapable
of large-scale deployment.

 These manual approaches also expose workers to health risks, especially

when operating in polluted or hazardous environments.

 Additionally, these methods are reactive rather than proactive, often

addressing pollution only after it has already accumulated significantly.

 Hence, there is a growing need for an intelligent, automated solution that

can detect and collect plastic waste in real-time, operate with minimal
human intervention, and use renewable energy sources for sustainable
operation.

3.2 EXISTING METHODOLOGY

 Current water body cleaning methods primarily rely on manual

operations using fishing nets or basic motorboats to collect floating
waste.
 In some cities, government or private bodies deploy manual or semi-
mechanized boats for periodic cleaning drives, but these are labor-
intensive and limited by the coverage area.
20
 These systems do not include real-time image recognition or plastic
detection capabilities, and hence, waste collection is often random or
inefficient.
 Image capturing and plastic detection technologies are rarely integrated,
leading to missed debris or partial cleanup operations.
 Navigation is usually manual or GPS-assisted but not autonomous,
requiring human drivers for movement and control.
 There is also a lack of continuous data logging, remote monitoring, or
predictive response, limiting the scalability and responsiveness of
existing methods.

21
 CHAPTER – 4

PROPOSED SYSTEM

4.1 INTRODUCTION

Plastic waste accumulation in water bodies poses a significant threat to

aquatic ecosystems, public health, and environmental sustainability. Traditional
methods of waste collection, such as manual netting or occasional boat patrols,
are inefficient, labor-intensive, and unable to ensure continuous cleaning. To
overcome these limitations, this project introduces a Smart Autonomous Plastic
Waste Collection Boat designed to detect and collect plastic waste using real-
time image processing and autonomous navigation.

The boat is built on a floatable platform powered by a solar-charged 12V

battery, which drives DC motors through an L298N motor driver, enabling the
boat to move independently on water surfaces. An ESP32-CAM module is
mounted on the boat to capture live images, which are transmitted to a local
system running a Python Flask server. A TensorFlow-based CNN model
processes these images to identify plastic waste. When plastic is detected, a
command is sent to an ESP8266 microcontroller, which controls a servo-
operated gate mechanism to collect the waste into a chamber. This solution
enables continuous, automated plastic collection without relying on cloud
services, offering a sustainable and scalable approach to water body
maintenance.

4.2 PROPOSED METHODOLOGY

The proposed smart plastic waste collection system operates autonomously

to maintain cleanliness in lakes, rivers, and coastal water bodies. It is built on a
floatable robotic boat platform powered by two DC motors with propellers,
ensuring smooth navigation across the water surface. The entire system is
energized by a solar-powered, rechargeable 12V battery, allowing for
22
continuous operation in remote or off-grid environments.

The boat is equipped with an ESP32-CAM module, which continuously

captures real-time images of the water surface. These images are transmitted to
a local computer running a Python Flask server, where a TensorFlow-based
Convolutional Neural Network (CNN) model processes them to detect plastic
waste. Upon detection, a signal is sent to an ESP8266 microcontroller, which
then activates a servo motor to open a gate mechanism and collect the waste
into an onboard chamber. Additionally, an ultrasonic sensor is integrated to
detect nearby obstacles such as walls or floating objects and prevent collisions
through basic avoidance logic.

Fig 4.1 Block diagram of Proposed System

23
 For navigation, the boat uses a GPS module (NEO-6M) to maintain its
course and automatically return to base once the task is completed or the battery
is low. The system runs entirely offline, eliminating the need for internet or
cloud services, making it suitable for deployment in remote areas. The
combination of real-time image detection, local processing, autonomous
collection, and GPS-based navigation ensures an efficient, low-cost, and eco-
friendly solution to combat plastic pollution in water bodies.

4.2.1 Hardware Setup:

 Design and deploy a floatable boat capable of moving independently

across water surfaces. Power the boat using a solar panel connected to a
rechargeable 12V battery, ensuring sustainable energy usage.

 Use two DC motors with propellers for forward and directional

movement, controlled via an L298N motor driver. Additionally, an
ultrasonic sensor is integrated to detect nearby obstacles such as walls or
floating objects and prevent collisions through basic avoidance logic.

 Mount an ESP32-CAM on the boat for capturing real-time images of the

water surface.

 Use an ESP8266 (NodeMCU) microcontroller to control the servo motor

and handle communication from the local server.

 Add a servo motor to open and close the collection chamber door when
plastic is detected.

 Integrate a GPS module (NEO-6M) for real-time location tracking and

return-to-base functionality.

24
4.2.2 Connectivity:

 The ESP32-CAM sends images to a Python Flask server hosted on a local

laptop/system using HTTP requests.

 The Flask server processes the images using a TensorFlow CNN model
trained to detect plastic waste.

 Upon detection, the server sends a response command to the ESP8266,

which triggers the servo to collect the waste.

 GPS data is used to guide the boat’s path and ensure it returns to base
when the battery is low or collection is complete.

4.2.3 Centralized Control System:

 A local Python Flask dashboard may be used to view logs of detections

and servo actions.

 All operations are executed offline, ensuring the system works in remote
areas without internet access.

 The server can be expanded to log timestamps, GPS coordinates, and

plastic detection status for analysis and future optimization.

4.2.4 Real-time Blockage Detection and Collection:

 The ESP32-CAM continuously captures images and transmits them to

the local system. The CNN model classifies each frame as "plastic" or
"clean water."

 On detection, the ESP8266 receives a signal to activate the servo

motor, opening the chamber to collect the detected waste.

 The boat navigates autonomously while continuously performing

25
 detection and collection tasks.

 The GPS module ensures the boat can return to its starting point
automatically after completing the task or when battery level drops.

4.3 HARDWARE COMPONENTS

 NodeMCU
 Ultrasonic Sensor
 ESP32 CAM
 GPS Module
 Motor Driver
 Solar Panel
 DC Motor
 Servo Motor
 Power Supply

4.3.1 NodeMCU
The NodeMCU is a widely-used open-source IoT platform based on the
ESP8266 WiFi module. It serves as the core processing unit and communication
gateway for the street light monitoring system. The NodeMCU serves as the core
component responsible for managing the hardware and facilitating
communication with external systems. It functions as a versatile microcontroller
unit based on the ESP8266 WiFi module, enabling wireless connectivity crucial
for transmitting data from sensors like the LDR and voltage sensor to monitor
street light conditions. The NodeMCU collects and processes data regarding
faults and location detection from these sensors, and it controls feedback
circuits and LED indicators to signify issues. Moreover, it interfaces with the
keypad for user input and communicates with the Telegram bot, allowing for
fault notifications to be sent to authorized personnel. Essentially, the
NodeMCU acts as the central control unit, orchestrating the operation of
26
various hardware components and enabling remote monitoring and
management of the street lighting network.

Fig 4.2 NodeMCU

The NodeMCU (Node MicroController Unit) is an open-source software

and hardware development environment built around an inexpensive System-on-
Chip (SoC) called the ESP8266. That makes it an excellent choice for Internet of
Things (IoT) projects of all kinds.
4.3.2 Ultrasonic Sensor

An Ultrasonic Sensor measures distance using ultrasonic sound waves. It

emits a high-frequency sound pulse and measures the time it takes for the echo
to return after bouncing off an object. This helps detect the water level or any
obstacle in front of the sensor. It is commonly used in applications like obstacle
avoidance, liquid level detection, and smart parking systems.

27
 Fig 4.3 ULTRASONIC Sensor

4.3.3 ESP32 CAM

The ESP32-CAM is a compact and cost-effective camera module
integrating an ESP32-S chip with a built-in OV2640 camera and a microSD
card slot. This module offers both Wi-Fi and Bluetooth connectivity, making it
highly versatile for various IoT applications. It operates as a self-contained
system with low power consumption, including a deep sleep mode that draws as
little as 6mA. The ESP32-CAM is powered by a dual-core 32-bit LX6 CPU,
with an adjustable main frequency ranging from 80MHz to 240MHz, providing
substantial processing power. Its integrated features make it ideal for projects
requiring image capture, video streaming, and wireless communication, such as
home security, smart monitoring, and facial recognition systems.

Fig 4.4 ESP32 CAM

28
4.3.4 GPS Module

Fig 4.5 GPS Module

A GPS module is a compact electronic device that receives signals from

orbiting GPS satellites to determine its precise location on Earth. It triangulates
these signals to calculate latitude, longitude, and altitude, often with impressive
accuracy. These modules are fundamental in applications ranging from
navigation systems and asset tracking to geocaching and scientific research.
They typically output location data in a standardized format like NMEA,
making it easy for microcontrollers or other processing units to interpret. Many
GPS modules also provide additional information such as speed, heading, and
time, which are invaluable for dynamic applications.

4.3.5 Motor Driver (L298N)

The L298N is a popular dual H-bridge motor driver IC module. It allows
you to control the speed and direction of two DC motors or one stepper motor
simultaneously. It can handle motor voltages from 5V to 35V and peak currents
up to 2A per channel. The module typically includes a heatsink and an onboard
5V regulator. It's widely used in robotics, automation, and hobby projects due to
its robustness and ease of interfacing with microcontrollers like Arduino.

29
 Fig 4.6 Motor Driver (L298N)

4.3.6 Power Supply

A rechargeable battery acts as a portable power supply by storing

electrical energy through reversible chemical reactions. Unlike single-use
batteries, they can be recharged multiple times, significantly reducing waste
and long-term costs. Common types include Lithium-ion (Li-ion), Nickel-
Metal Hydride (NiMH), and Lead-acid, each suited for different applications
based on energy density and discharge rates. They are fundamental in devices
ranging from smartphones and laptops to electric vehicles and renewable
energy storage systems, offering convenience and promoting sustainability.
Proper charging and maintenance are crucial for maximizing their lifespan and
performance.

Fig 4.7 Power Supply rechargeable battery

30
4.3.7 Solar Panel

A solar panel is a device that converts sunlight directly into electricity using
the photovoltaic effect. It's composed of multiple photovoltaic cells, usually
made from silicon, which absorb photons from the sun and release electrons,
generating an electric current. These panels are a cornerstone of renewable
energy, offering a clean and sustainable power source that reduces reliance on
fossil fuels. They come in various sizes and efficiencies, suitable for
applications from powering small electronic devices to large-scale grid-
connected power plants. The amount of electricity a solar panel generates
depends on factors like sunlight intensity, panel size, and efficiency, as well as
the angle at which it's positioned relative to the sun.

Fig 4.8 Solar Panel

31
4.3.8 DC Motor

A DC motor is an electric motor that converts direct current (DC)

electrical energy into mechanical energy through the interaction of magnetic
fields and electrical conductors. It operates on the principle that a current-
carrying conductor placed in a magnetic field experiences a force. In a DC
motor, stationary magnets (stator) create a magnetic field, and a rotating coil
(rotor or armature) with current flowing through it experiences this force,
causing it to spin. The direction and speed of rotation can be controlled by
varying the polarity and magnitude of the applied voltage, respectively. DC
motors are widely used in a vast array of applications, including electric
vehicles, industrial machinery, robotics, and household appliances, due to their
simplicity, reliability, and ease of control. Their versatility makes them a
fundamental component in many electromechanical systems.

Fig 4.9 DC Motor

4.3.9 Servo Motor

A servo motor is a rotary or linear actuator that allows for precise control
of angular or linear position, velocity, and acceleration. Unlike standard DC
motors, a servo motor typically consists of a DC motor, a gearbox, a position
sensor (like a potentiometer or encoder), and a control circuit. The control

32
circuit receives a command signal, usually a pulse-width modulated (PWM)
signal, and then drives the motor to the desired position. This closed-loop
control is what enables their high accuracy and repeatability, making them
indispensable in applications requiring precise movement, such as robotics,
CNC machinery, remote control airplanes, and industrial automation.

Fig 4.10 Servo Motor

4.4 SOFTWARE USED

Arduino IDE

A program for Arduino hardware may be written in any programming

language with compilers that produce binary machine code for the target
processor. Atmel provides a development environment for their 8-bit AVR and
32-bit ARM Cortex-M based microcontrollers: AVR Studio (older) and Atmel
Studio (newer). Arduino is an open-source hardware and software company,
project and user community that designs and manufactures single-board
microcontrollers and microcontroller kits for building digital devices and
interactive objects that can sense and control both physically and digitally. Its
products are licensed under the GNU Lesser General Public License (LGPL) or
the GNU General Public License(GPL), permitting the manufacture of Arduino
boards and software distribution by anyone. Arduino boards are available
commercially in preassembled form or as do-it-yourself (DIY) kits.

33
 Arduino board designs use a variety of microprocessors and controllers.
The boards are equipped with sets of digital and analog input/output (I/O) pins
that may be interfaced to various expansion boards or breadboards (shields) and
other circuits. The boards feature serial communications interfaces, including
Universal Serial Bus (USB) on some models, which are also used for loading
programs from personal computers. The microcontrollers are typically
programmed using a dialect of features from the programming languages
C and C++. In addition to using traditional compiler tool chains, the Arduino
project provides an integrated development environment (IDE) based on the
Processing language project.
It includes a code editor with features such as text cutting and pasting,
searching and replacing text, automatic indenting, brace matching, and syntax
highlighting, and provides simple one- click mechanisms to compile and upload
programs to an Arduino board. It also contains a message area, a text console, a
toolbar with buttons for common functions and a hierarchy of operation menus.
The source code for the IDE is released under the GNU General Public License,
version 2.

The Arduino IDE supports the languages C and C++ using special rules of
code structuring. The Arduino IDE supplies a software library
from the Wiring project, which provides many common input and output
procedures. User-written code only requires two basic functions, for starting the
sketch and the main program loop, that are compiled and linked with a program
stub main() into an executable cyclic executive program with the GNU tool
chain, also included with the IDE distribution. The Arduino IDE employs the
program avr port to convert the executable code into a text file in hexadecimal
encoding that is loaded into the Arduino board by a loader program in the board's
firmware.

34
 Python Flask & Tensor Flow
Python Flask is a lightweight and flexible web framework widely used to
develop web applications and APIs. When combined with TensorFlow, a
powerful open-source machine learning library, Flask enables developers to
deploy deep learning models as web services accessible through HTTP requests.
In a typical setup, a pre-trained TensorFlow model is loaded within a Flask app,
allowing the server to process incoming data, such as images or text, and return
predictions in real time.
Flask’s simplicity makes it easy to create RESTful endpoints where clients
can send input data, triggering TensorFlow’s inference operations on the
backend. This integration is particularly useful for AI applications like image
classification, object detection, and natural language processing, where low-
latency responses are crucial. The Flask app handles tasks like data
preprocessing before feeding it into the TensorFlow model, then post-processes
the output to generate human-readable results. Developers benefit from Flask’s
modular design, which supports middleware for logging, security, and
scalability enhancements.
Additionally, Flask apps can be containerized using Docker and deployed on
cloud platforms, ensuring reliable and scalable access to AI models. This
architecture democratizes AI by making sophisticated models accessible via
simple web requests, enabling integration with front-end applications, mobile
apps, or IoT devices. Overall, using Flask with TensorFlow creates an efficient,
user-friendly interface to complex machine learning models, accelerating the
development of intelligent applications across various domains.

35
 CHAPTER – 5
RESULT & DISCUSSISON
5.1 INTRODUCTION
This chapter presents the results obtained from the implementation and
testing of the Autonomous Plastic Waste Collection Boat. The performance of
each hardware component and the overall functionality of the system are
analyzed. Real-time plastic detection, obstacle avoidance, autonomous
navigation, and servo-based collection are tested in simulated water
environments. The system’s effectiveness, challenges faced, and improvement
opportunities are also discussed.

5.2 PROJECT PROTOTYPE

Fig. 5.1 Hardware Implementation

36
5.3 PROTOTYPE DESCRIPTION
The prototype consists of a floatable robotic boat equipped with an
ESP32-CAM for real-time image capture, an ESP8266 microcontroller for
motor and servo control, and a GPS module for navigation. A TensorFlow-
based CNN model runs on a local system using a Flask server to detect plastic
waste from incoming images. A servo motor is activated to open a collection
gate when plastic is detected. For obstacle detection, an ultrasonic sensor is
installed at the front to detect nearby objects and prevent collisions. The boat is
powered by a solar-charged 12V battery, enabling eco-friendly and continuous
operation. The GPS module allows the boat to autonomously return to base after
task completion.

5.4 Working Principle

The Autonomous Solar-Powered Plastic Waste Collection Boat functions

by combining renewable energy, AI-based image detection, and IoT
technologies to monitor and clean water bodies effectively. Once powered on,
the solar panel charges the onboard lithium battery, which supplies power to all
boat components, including the ESP32-CAM, microcontroller
(ESP8266/NodeMCU), and motor systems.
The ESP32-CAM continuously captures images of the water surface.
These images are either processed on the cloud or locally, where an AI-based
CNN model detects the presence of plastic waste. If plastic is identified, the
ESP8266 receives a signal and activates the waste collection mechanism by
triggering a servo motor or actuator to open a small collection door.
The boat moves autonomously using DC motors driven by an L298N
motor driver. Navigation is aided by a GPS module to ensure route coverage and
return-to-base capability. Obstacle avoidance is managed using ultrasonic
sensors to prevent collisions with debris or other objects.

37
 All sensor data, GPS location, and detection events are transmitted in
real-time to a Firebase cloud platform. This enables remote monitoring through
a mobile or web dashboard. Alerts, such as full waste bin or low battery, are sent
automatically via email or app notifications to concerned authorities. Once the
waste is collected or the mission is completed, the boat returns to its docking
station for unloading and recharging, ensuring an efficient, self-sustained waste
management system.

5.5 Project Output:

Fig 5.2 Output of the project

5.6 HARDWARE ANALYSIS RESULT:

5.6.1 NodeMCU Functionality

The ESP8266 (NodeMCU) efficiently handles control logic, motor

control, and response to Flask server commands. It processes ultrasonic
readings, GPS data, and triggers the servo motor. During testing, it
demonstrated stable performance and accurate responses to detection signals.

38
 5.6.2 ESP32-CAM Performance

The ESP8266 (NodeMCU) efficiently handles control logic, motor

control, and response to Flask server commands. It processes ultrasonic
readings, GPS data, and triggers the servo motor. During testing, it
demonstrated stable performance and accurate responses to detection signals.

5.6.3 Power Supply Stability

The system is powered by a 12V rechargeable battery, continuously charged

using a solar panel. During test runs, the power system provided stable output
for all components including motors, ESP modules, GPS, and servo. This setup
ensures uninterrupted performance in outdoor environments.

5.7 Output Discussion:

The implementation of this AI-based autonomous boat presents a practical and

scalable solution to plastic pollution in water bodies. The TensorFlow-based CNN
model successfully identified plastic in most test images with good accuracy. The real-
time detection and servo-based waste collection worked reliably in controlled
conditions.
The ESP8266 and ESP32-CAM integration proved successful, ensuring smooth
communication and fast response to detection signals. Ultrasonic obstacle detection
added an extra layer of safety to avoid collisions with debris or walls. GPS navigation
and return-to-base functions operated well during testing.
 Variations in image quality due to lighting or water reflections.
 Minor delays in image transmission and processing.
 Need for stronger power management for extended operation.
Despite these, the boat functioned autonomously with minimal human intervention
and is ready for further field deployment. Future improvements may include adding
object tracking, waste bin fill sensors, and better edge detection for advanced path
planning.

39
 CHAPTER – 6

CONCLUSION & FUTURE SCOPE

6.1 CONCLUSION
The implementation of the Smart Autonomous Plastic Waste Collection
Boat presents an innovative, efficient, and eco-friendly approach to mitigating
plastic pollution in water bodies. By integrating AI-based image processing with
autonomous navigation and sensor-driven obstacle avoidance, the system
effectively detects and collects floating plastic waste with minimal human
intervention. The use of solar-powered energy combined with a rechargeable
battery ensures sustainable operation, making the boat suitable for extended
deployment in remote or off-grid areas.

The collaboration between ESP32-CAM for real-time image capture, a

TensorFlow CNN model for plastic detection, and the ESP8266 microcontroller
for control logic creates a robust and reliable platform.

The ultrasonic sensor and GPS module further enhance autonomy by

enabling collision avoidance and return-to-base navigation. This project reduces
the time and labor required for plastic waste cleanup, offers scalability for larger
water bodies, and promotes environmental conservation.

Overall, the Smart Autonomous Plastic Waste Collection Boat

demonstrates a practical advancement in automated environmental cleanup
technology, providing a scalable, low-cost, and sustainable solution for
combating plastic pollution in aquatic ecosystems.

40
6.2 FUTURE SCOPE:

To enhance the functionality and adaptability of the Smart Autonomous Plastic Waste
Collection Boat, the following future developments can be explored:

1. Onboard AI Edge Computing:

Deploying lightweight deep learning models directly on the ESP32 or similar
edge devices can enable real-time plastic detection and decision-making
without reliance on external servers. This will reduce latency and increase
system robustness in areas without network connectivity.
2. Advanced Navigation Algorithms:
Incorporating Simultaneous Localization and Mapping (SLAM) techniques
combined with multi-sensor fusion (ultrasonic sensors, LIDAR, IMU) could
improve autonomous navigation in complex water environments, enabling
more precise obstacle avoidance and path planning.
3. Hybrid Renewable Energy System:
Integrating solar panels with energy harvesting techniques (such as small wind
turbines or wave energy converters) and advanced battery management can
ensure longer operational duration, especially in challenging environmental
conditions.
4. Swarm Robotics for Large-Scale Cleanup:
Deploying multiple coordinated boats using mesh networking would allow
collaborative cleaning over extensive water bodies, improving efficiency,
coverage, and fault tolerance.
5. Expanded Environmental Monitoring:
Adding sensors to monitor water quality parameters (e.g., pH, turbidity,
chemical pollutants) alongside plastic waste detection would transform the boat
into a comprehensive environmental monitoring platform.

41
 6. 5G and IoT Integration:
Leveraging 5G connectivity can enable ultra-fast communication for remote
monitoring and control, real-time data analytics, and cloud integration for
enhanced decision-making and maintenance.

REFERENCES

[1] A. Vellingiri, K. Dharni, M. Arunadevi et al, “Iot Based Smart Monitoring

System for Sewage Workers with Two way Communication “, International
Research Journal of Engineering and Technology, vol 7, issues:6 june,2020

[2] S. K. Singh, A. Kumar, and P. K. Sahu, "AI-Based Autonomous Boat for

Plastic Waste Collection in Water Bodies," International Journal of
Environmental Science and Technology, vol. 18, no. 4, pp. 987–996, 2023.

[3] J. Park, M. Lee, and H. Kim, "Deep Learning-Based Plastic Waste Detection
Using ESP32-CAM and TensorFlow," Journal of Cleaner Production, vol.
350, 2022.

[4] R. Sharma and V. Gupta, "Solar-Powered Autonomous Robots for Water

Pollution Cleanup," International Journal of Robotics and Automation, vol.
37, no. 2, pp. 145–153, 2024.

[5] M. N. Islam, S. S. Chowdhury, and M. A. Rahman, "Design and

Implementation of an IoT-Based Plastic Waste Collection Boat," IEEE
Access, vol. 11, pp. 34145–34153, 2023.

[6] L. Zhang, Y. Wang, and Q. Zhou, "Obstacle Avoidance and Navigation in

Autonomous Boats Using Ultrasonic Sensors and GPS," Sensors, vol. 21,
no. 11, 2021.

42
[7] A. Vellingiri, K. Dharni, M. Arunadevi, R. L. Aravind Lal, “IoT Based
Smart Monitoring System for Sewage Workers with Two-Way
Communication,” International Research Journal of Engineering and
Technology, vol. 7, no. 6, June 2020.

[8] P. Patel and S. Mehta, "Low-Cost Embedded Systems for Environmental

Monitoring: A Review," Environmental Monitoring and Assessment, vol.
194, no. 5, 2022.

[9] H. K. Verma and R. Singh, "Autonomous Surface Vehicles for

Environmental Cleanup: A Survey," Journal of Marine Science and
Engineering, vol. 11, no. 3, 2023.

[10] R. T. Mohan and P. Singh, "Implementation of CNN-Based Image

Recognition for Marine Litter Detection," International Journal of Computer
Vision and Pattern Recognition, vol. 8, no. 2, pp. 70–85, 2024.

[11] S. Kim, J. Choi, and M. Park, "Energy-Efficient Solar Power Systems for
Autonomous Marine Robots," Renewable Energy, vol. 183, pp. 1376–1385,
2022.

[12] Y. Lee and S. J. Park, "Ultrasonic Sensor-Based Obstacle Detection and

Avoidance for Autonomous Watercraft," International Journal of Robotics
Research, vol. 40, no. 1, pp. 42–55, 2023.

43