GESTURE CONTROLLED ROVER WITH
VISUAL PROCESSING USING ESP32-CAM
AND EDGE AI
A Thesis Submitted In Partial Fulfillment of the Requirements for the Award of the Degree of
BACHELOR OF TECHNOLOGY ([Link])
In
ELECTRONICS AND COMMUNICATION ENGINEERING
Submitted by
Student Name Registration No-
SUNIL KUMAR 21052465007
VIJAY MAHALI 21052465008
VISHAL KUMAR 21052465009
Under the Supervision of
Mrs. Anamika Kumari (Asst. Prof. Deputy H.O.D)
Mr. Hare Krishna (Asst. Prof.)
DEPARTMENT OF ELECTRONICS AND COMMUNICATION
ENGINEERING
Academic Year: 2024-2025
CANDIDATE’S DECLARATION
I hereby certify that I completed the work mentioned in this report, titled " GESTURE
CONTROLLED ROVER WITH VISUAL PROCESSING USING ESP32-CAM AND EDGE AI
". I have not submitted the material in this report to any other University or Institute in order to obtain
a degree or diploma. For all words, concepts, figures, graphics, computer programs, tests, and results
that were not my own, I credited the original author or sources. I've put quotation marks around direct
quotes and given credit to the original author or sources.
I assure that none of my work has been plagiarized and that the experiments and results presented in
the report have not been altered. If someone claims I duplicated someone else's work or changed the
experiments and results, I will be entirely responsible and answerable.
Name Reg. No. Signature
SUNI KUMAR 21052465007
VIJAY MAHALI 21052465008
VISHAL KUMAR 21052465009
I
CERTIFICATE OF DECLARATION
This is to declare that
SUNI KUMAR 21052465007
VIJAY MAHALI 21052465008
VISHAL KUMAR 21052465009
Has conducted the research for this thesis, " GESTURE CONTROLLED ROVER WITH VISUAL
PROCESSING USING ESP32-CAM AND EDGE AI " for his Bachelor of Technology work at RTC
Institute of Technology, Anandi (RANCHI), while I was his supervisor. The thesis is based on the
student's own research and studies. Based on what is in the thesis, the candidate or anybody else cannot
obtain another degree from this or any other university or college.
Name of Supervision Signature
Mrs. Anamika Kumari (Asst. Prof. Deputy H.O.D)
Mr. Hare Krishna (Asst. Prof.)
Department of Electronics & Communication Engineering
RTC Institute of Technology, Anandi (RANCHI)
Date:
Signature of HOD
Electronics and Communication Engineering
II
CERTIFICATE OF APPROVAL FROM THE EXAMINATION
The foregoing dissertation entitled “GESTURE CONTROLLED ROVER WITH VISUAL
PROCESSING USING ESP32-CAM AND EDGE AI” is hereby approved as creditable study of the
research topic and has been presented satisfactorily to warrant its acceptance as a perquisite to the
degree for Bachelor of Technology in ELECTRONICS AND COMMUNICATION ENGINEERING at
RTCIT, it is understood that by this approval, the undersigned do not necessarily endorse any conclusion
drawn or approval expressed there in but approve the report for the purpose for which it is submitted.
Signature of Internal Examiner Signature of External Examiner
Signature of HOD
Electronics and Communication Engineering
III
ACKNOWLEDGEMENT
We would like to articulate our deep gratitude towards our project guide Mr. Hare Krishna,
Mrs. Anamika Kumari who has been a source of motivation and firm support for carrying out
the project, through his invaluable suggestions and constant encouragement all through the
dissertation work. We would like to express our sincere thanks to all the faculty member of the
Electronics and Communication Engineering Department for their kind cooperation. We would
like to acknowledge of this nature could never have been attempted without reference to and
inspiration from the works of others whose details are mentioned in the Reference section. We
acknowledge our indebtedness to all of them. Further we would like to express our feelings
towards our family member who directly or indirectly encouraged and motivated us during this
project.
SUNI KUMAR 21052465007
VIJAY MAHALI 21052465008
VISHAL KUMAR 21052465009
Place: RTCIT, Ranchi
Date:
IV
ABSTRACT
This project introduces a compact, intelligent robotic rover that seamlessly combines wireless gesture
control with onboard visual perception. The system empowers users to guide the rover’s movement
through intuitive hand gestures transmitted via a smartphone interface, while simultaneously equipping
the rover with autonomous object awareness using edge-deployed artificial intelligence. At the heart of
the design lies the ESP32-CAM module, which captures visual data and processes it locally using a
lightweight, embedded machine learning model. This self-sufficient approach enables the rover to detect
objects and obstacles in real-time, independent of cloud connectivity or external servers.
The rover’s architecture prioritizes simplicity, affordability, and portability, making it ideal for tasks
such as remote exploration, situational assessment in inaccessible environments, and practical learning
in robotics and embedded AI. Through the harmonious integration of gesture-driven navigation and
edge-based visual recognition, the system demonstrates how cost-effective components can deliver
intelligent behavior in dynamic, real-world settings. This work contributes to the growing landscape of
accessible autonomous robotics, highlighting the potential of intuitive human-machine collaboration in
resource-limited-conditions.
V
TABLE OF CONTENTS
CERTIFICATE II
CANDIDATE’S DECLARATION III
SUPERVISOR DECLARATION IV
APPROVAL FROM EXAMINATION V
ACKNOWLEDGEMENT VI
ABSTRACT VII
TABLE OF CONTENTS VIII
LIST OF FIGURES IX
LIST OF CONTENTS
Chapter 1:
1.1 Introduction 1
1.2 Motivation 1-2
1.3 Objective 2-3
Chapter 2: Literature Review
2.1 Introduction 4-5
2.2 Research References 5-8
Chapter 3: Hardware requirements and Specifications
3.1 Components Required 9
3.2 Block Diagram 9-10
3.3 Circuit Diagram 11-13
3.4 Arduino IDE 13-14
3.5 Arduino Uno 15
3.6 Arduino Uno Features and Pin Configuration 16-18
3.7 HC-05 Bluetooth Module 19
3.8 Bluetooth Communication Between Devices 20
3.9 Motor Driver Module L298N 21-22
3.10 Voltage Drop of L298N 23
3.11 Micromors and Grippy Wheels 24
3.12 Jumper Wires 25
3.13 Batteries 26
3.14 Switch 27
3.15 Smartphone with app 28
VI
3.16 ESP32-CAM module 28-30
Chapter 4: Design and Implementation
4.1 Wiring L298N Motor Driver Module to Arduino UNO 31
Chapter 5: Edge impulse 32-34
Chapter 6: Result and Analysis
6.1 Resultant and Analysis 35-36
6.2 Applications 37
6.3 Advantage 37
Chapter 7: Conclusion and Future Scope
7.1 Conclusion 39
7.2 Future Scope of the Project 39-40
Further Enhancements/Recommendations 41-42
References/Bibliograph
VII
LIST OF FIGURES
Sr. Number Figure Name Page Number
1. Components Block diagram 9
2. Circuit Diagram 11
3. Arduino IDE 14
4. Arduino Uno 15
5. Arduino Uno connections 18
6. Bluetooth Module 19
7. Bluetooth communications 20
8. Motor driver module 22
9. Voltage Drop of L298N 23
10. Motor and Wheels 24
11. Jumper Wires 25
12. Batteries 26
13. Switch 27
15. Smartphone with app 28
16. ESP32-CAM 29
17. Connecting Motor Driver Module to an 31
Arduino
18. Edge impulse 32
19. Working Explanation 36
VIII
CHAPTER 1
1.1 INTRODUCTION
The modern era has witnessed significant advancements in the fields of robotics, embedded systems, and
artificial intelligence (AI). With these advancements, there is a growing demand for intelligent,
autonomous, and cost-effective robotic platforms that can assist in performing tasks in remote,
hazardous, or inaccessible environments. The Gesture Controlled Rover with Visual Processing is a
prime example of how affordable technologies can be integrated to create such smart robotic systems.
This project focuses on developing a rover that combines gesture-based wireless control with real-time
visual processing powered by AI. The system allows users to control the rover's movement through
simple gestures made on a smartphone application. Simultaneously, the rover is equipped with an
ESP32-CAM module, which captures live video feed and processes images using onboard AI to detect
objects and navigate intelligently. By leveraging edge computing, the rover operates without relying on
constant internet connectivity or remote servers, making it ideal for field deployments in remote or risky
areas.
The rover is designed with portability and affordability in mind, ensuring that it can be utilized by
researchers, students, and professionals alike for various applications such as surveillance, search-and-
rescue operations, environmental monitoring, and educational purposes. This project demonstrates how
embedded AI, combined with intuitive human-machine interaction, can lead to impactful, real-world
robotic solutions.
1.2 Motivation
The continuous evolution of robotics, artificial intelligence, and embedded systems has opened new
possibilities for developing intelligent machines that can assist humans in everyday tasks. However,
many advanced robotic platforms remain costly, complex, and dependent on heavy computing
infrastructure, making them inaccessible for common people, students, or small organizations.
The motivation behind this project is to design a simple, portable, and affordable robotic system that
offers intelligent behavior without relying on cloud servers or expensive hardware. By using easily
available components like the ESP32-CAM, Arduino UNO, and a Bluetooth module, combined with
lightweight AI models, it becomes possible to create a functional robotic rover suitable for real-world
applications.
1
Another key motivation is to promote the use of gesture control, an intuitive and human-friendly method
of interacting with machines. Instead of using complex remote controls, users can guide the rover's
movement with simple hand gestures, making the system easy to use for beginners, students, and even
individuals with limited technical knowledge.
Additionally, the integration of onboard AI-based visual processing allows the rover to understand its
surroundings, detect obstacles, and respond intelligently. This enhances safety, autonomy, and practical
usability, especially in fields like surveillance, search and rescue, and educational robotics.
The project also aims to encourage learning and innovation in the field of embedded AI and robotics by
demonstrating that advanced, smart robotic solutions can be achieved with minimal resources and
accessible technology.
1.3 Objective
The primary objective of this project is to design and implement a smart, portable robotic rover that
combines gesture-based wireless control with onboard visual processing using embedded artificial
intelligence.
The specific goals of the project are as follows:
To develop a user-friendly rover that can be controlled wirelessly through simple hand gestures using a
smartphone application.
To integrate an ESP32-CAM module for real-time image capture and object detection directly on the
device without relying on cloud servers.
To utilize lightweight, memory-efficient AI models, trained using Edge Impulse, for detecting objects
such as bottles, humans, laptops, or obstacles.
To demonstrate how affordable, off-the-shelf components like Arduino UNO, HC-05 Bluetooth module,
and ESP32-CAM can be combined to build an intelligent, low-cost robotic platform.
To promote the concept of Edge AI by enabling real-time object detection on resource-constrained
embedded hardware.
2
To create a modular and scalable system design that allows for easy upgrades, additional sensor
integration, or future improvements such as autonomous navigation.
To validate the effectiveness of the system for real-world applications, including remote surveillance,
search operations, environmental monitoring, and educational use.
3
CHAPTER 2
LITERATURE REVIEW
2.1 INTRODUCTION
In recent years, the demand for intelligent and user-friendly robotic systems has increased significantly,
driven by advancements in embedded systems, wireless communication, and artificial intelligence.
Numerous studies have explored different approaches to developing mobile robots that combine easy
control methods with autonomous environmental awareness.
Earlier research primarily focused on remote-controlled vehicles operated using dedicated remotes or
wired interfaces. While these methods provided reliable control, they lacked flexibility, portability, and
often required bulky hardware. With the rise of smartphones, researchers began exploring mobile-based
control systems that offer greater convenience and accessibility. Studies have shown that gesture
control, when combined with mobile applications provides an intuitive and efficient method to operate
robotic platforms, eliminating the need for complex hardware controllers.
On the other hand, the integration of visual perception into mobile robots has been a key area of interest
for enhancing autonomy. Conventional approaches often relied on high-end computing devices or cloud-
based image processing, which increased system cost and required constant internet connectivity. This
made such systems less suitable for resource-limited environments or outdoor field operations.
In response to these limitations, recent works have explored the use of compact microcontrollers, such
as the ESP32-CAM, which offer affordable onboard image capture and processing capabilities. Studies
utilizing platforms like Edge Impulse and TensorFlow Lite have demonstrated that lightweight machine
learning models can be deployed directly on such devices, enabling real-time object detection without
the need for cloud servers. This approach significantly reduces system cost, power consumption, and
dependency on internet infrastructure.
Combining gesture-based control and embedded visual intelligence presents a promising solution for
creating smart, affordable, and portable robotic platforms. While several works have addressed these
features independently, limited research has been done on integrating both technologies into a single,
low-cost rover suitable for real-world applications such as surveillance, rescue, or environmental
monitoring.
The present work builds upon these concepts by introducing a gesture-driven rover that incorporates
4
real-time object detection using the ESP32-CAM. Unlike previous studies that either relied on manual
control or expensive AI hardware, this project emphasizes accessibility, low-power operation, and
independence from external processing units. The use of a smartphone as the control interface,
combined with edge AI for visual understanding, offers a unique, easy-to-use solution for a wide range
of practical scenarios.
Overall, the literature suggests that while various components for gesture control and AI-based vision
are available individually, their effective integration into a compact, cost-efficient rover remains an area
with considerable potential. This project contributes to filling that gap by providing a working,
affordable prototype that showcases the benefits of such integration.
2.2 Research References
Howard, Andrew et al. (2017). MobileNets: Efficient Convolutional Neural Networks for Mobile Vision
Applications.
This work introduces MobileNets, a lightweight and efficient deep learning architecture designed
specifically for mobile and embedded vision applications. The paper proposes methods to reduce
computational complexity while maintaining good accuracy, making it ideal for devices like the ESP32-
CAM used in this project for real-time object detection.
Jia Deng et al. (2009). ImageNet: A large-scale hierarchical image database.
The ImageNet project presents a massive, labeled image dataset that has significantly influenced the
progress of computer vision and deep learning. Datasets like ImageNet form the backbone of training
models that can recognize everyday objects, similar to the object detection system deployed in this
project.
Raheja, J. L. et al. (2010). Real-Time Robotic Hand Control using Hand Gestures.
This paper explores the use of hand gestures for controlling robotic devices in real time. The work
demonstrates the practicality of gesture-based human-machine interaction, which directly relates to the
gesture control system implemented for the rover in this project.
P. B. Nayana and S. Kubakaddi (2014). Implementation of Hand Gesture Recognition Technique for
5
HCI Using OpenCV.
5
The authors present a simple and effective approach for hand gesture recognition using OpenCV. Their
work highlights how computer vision tools can be used for human-computer interaction (HCI), which is
a fundamental part of gesture-controlled robotic systems like the one developed in this project.
Ahuja, M. K., & Singh, A. (2015). Static Vision-Based Hand Gesture Recognition using Principal
Component Analysis.
This paper proposes a vision-based method for recognizing hand gestures using Principal Component
Analysis (PCA). The approach emphasizes simplicity and efficiency, both of which align with the
objectives of this project in developing an affordable, gesture-controlled rover.
Bretzner, L., Laptev, I., & Lindeberg, T. (2002). Hand Gesture Recognition using Multi-Scale Colour
Features, Hierarchical Models and Particle Filtering.
The authors present a robust technique for hand gesture recognition using color features and advanced
filtering methods. Their work contributes to the understanding of how reliable gesture recognition can
be implemented, an essential aspect of the gesture-controlled rover in this project.
Chen, F.-S., Fu, C.-M., & Huang, C.-L. (2003). Hand Gesture Recognition using a Real-Time Tracking
Method and Hidden Markov Models.
This paper proposes a real-time hand gesture recognition system based on tracking methods and Hidden
Markov Models (HMMs). The idea of combining real-time tracking with pattern recognition provides
valuable insight for gesture-based control applications.
DiPietro, L., Sabatini, A. M., & Dario, P. (2008). A Survey of Glove-Based Systems and Their
Applications.
This survey reviews glove-based gesture recognition systems and their practical applications. While this
project does not use gloves, the study helps understand different gesture recognition approaches,
highlighting the advantages of camera-based systems for intuitive control, as implemented here.
Dong, G., Yan, Y., & Xie, M. (1998). Vision-Based Hand Gesture Recognition for Human-Vehicle
Interaction.
This paper presents an early approach to using computer vision for recognizing hand gestures to control
vehicles. The research highlights how vision-based gesture systems can improve human-machine
interaction, which directly relates to this project's use of gesture control for robotic navigation.
6
Garg, P., Aggarwal, N., & Sofat, S. (2009). Vision-Based Hand Gesture Recognition.
This work provides an overview of various techniques for recognizing hand gestures using vision
systems. It discusses challenges and advancements in this field, offering valuable background for
implementing reliable gesture control in robotic platforms like the one developed in this project.
Gupta, S., Jaafar, J., & Ahmad, W. F. W. (2012). Static Hand Gesture Recognition Using Local Gabor
Filter.
The authors propose a method for recognizing static hand gestures by extracting features with Local
Gabor Filters. This technique offers enhanced feature detection, contributing to more accurate gesture
recognition systems, which supports the design goals of this project.
Gupta, S., Jaafar, J., & Ahmad, W. F. W. (2012). Static Hand Gesture Recognition Using Local Gabor
Filter.
The authors propose a method for recognizing static hand gestures by extracting features with Local
Gabor Filters. This technique offers enhanced feature detection, contributing to more accurate gesture
recognition systems, which supports the design goals of this project.
OpenCV Library. [Online] Available: [Link]
OpenCV is a powerful open-source computer vision library widely used for image processing and object
detection tasks. Its tools and functions are fundamental for implementing gesture recognition and visual
processing in embedded systems like the ESP32-CAM.
Arduino. [Online] Available: [Link]
The official Arduino platform provides resources, tutorials, and technical guides for working with
microcontrollers. It serves as the foundation for controlling the rover's hardware, including motors,
sensors, and Bluetooth modules, in this project
S. Soo (2014). Object Detection Using Haar-Cascade Classifier.
This work explains how Haar-cascade classifiers can be used for object detection. The technique, based
on feature-based classification, is efficient for real-time applications and has been applied widely in
7
computer vision tasks, including object detection for robotics.
P. Viola and M. Jones (2001). Rapid Object Detection Using a Boosted Cascade of Simple Features.
This seminal paper introduced the Viola-Jones object detection framework, which remains a cornerstone
for real-time face and object detection. The method's lightweight nature makes it suitable for embedded
systems with limited processing power, similar to the ESP32-CAM used in this project.
8
CHAPTER 3
HARDWARE REQUIREMENTS AND SPECIFICATIONS
3.1 Components Required
1. Arduino UNO
2. HC – 05 Bluetooth Module
3. Motor Driver L298N
4. Jumper Wire
5. Drilling Machine
6. Soldering Gun
7. 4 X 5V Geared Motors
8. Connecting Wires
9. Battery holder
10. Power Supply
11. Android Phone
12. Bluetooth Controller App
13. ESP32-CAM module
3.2 Block diagram and working principle
ESP32-CAM module
Figure 3.2: Block Diagram
Figure 3.2: Block Diagram
The overall working of the rover is based on two main operations happening simultaneously — gesture-
based movement control and visual processing using onboard AI. The block diagram clearly shows how
9
different parts of the system interact with each other to make the rover function properly.
Gesture-Based Movement Control
The user operates the system through a Smartphone Application, specially designed to capture simple
hand gestures or screen touches.
Once the gesture is detected, the mobile app sends a wireless signal using Bluetooth communication,
through the HC-05 Bluetooth Module.
The HC-05 Module receives the signal and passes it to the Arduino UNO microcontroller. The Arduino
UNO acts as the brain for movement control.
The Arduino reads the incoming signal and decides the required action — whether the rover should
move forward, backward, left, right, or stop.
According to the command, the Arduino sends signals to the L298N Motor Driver, which is responsible
for controlling the direction and speed of the DC Motors.
The DC Motors then rotate the wheels of the rover in the desired direction as per the user’s gesture,
allowing smooth and wireless movement of the vehicle.
Visual Processing with Onboard AI
At the same time, the rover carries an ESP32-CAM Module, which is fitted with a small camera
(OV2640).
The ESP32-CAM keeps capturing live images of the surroundings continuously.
These images are processed directly on the device using a pre-trained AI Model, which runs on the
ESP32-CAM itself. This model has been trained using Edge Impulse, a platform for building lightweight
AI models.
With the help of this onboard AI, the rover can detect objects like obstacles, people, or specific items in
its path.
If an object is detected, the system can take pre-programmed actions, such as sending alerts or stopping
the rover to avoid collisions.
Importantly, all this processing happens within the rover itself, without needing the internet or cloud
services. This makes the system independent and suitable for areas with no network.
Power Supply
The entire system, including the Arduino, ESP32-CAM, Motor Driver, and Bluetooth Module, is
powered by a Lithium-ion Battery Pack, ensuring that the rover works wirelessly and can move freely
without being plugged in.
3.3 Circuit Diagram
The circuit diagram is shown below :-
10
Figure 3.3: Circuit Diagram
Explanation of Circuit Diagram
The given circuit diagram represents the electrical connections of the Gesture Controlled Rover
with Visual Processing. Each component is carefully interconnected to ensure proper
communication, power supply, and control of the rover's movement and visual operations. Power
Supply
The entire system is powered by a Lithium-ion Battery Pack (3 x 3.7V), which supplies sufficient
voltage to operate all components including the Arduino, ESP32-CAM, Motor Driver, and
Bluetooth module. Proper power and ground lines are shared among all modules to maintain
stability. Arduino UNO and Motor Driver (L298N)
The Arduino UNO acts as the central controller for interpreting gesture [Link] wires
from the Arduino connect to the L298N Motor Driver, which controls two DC Motors based on
received signals.
The ENA, IN2, IN4, IN3, and QUT1/QUT2 pins are connected to control motor direction and
speed.
The motors are connected to the output terminals of the L298N, providing mechanical movement
to the rover wheels. HC-05 Bluetooth Module
11
The HC-05 Module establishes a wireless link between the user's smartphone and the Arduino
UNO.
TX and RX lines are connected to the Arduino for transmitting and receiving gesture commands.
Power and ground connections are shared from the Arduino supply.
Here at first we construct the circuit as shown in Figure 2.3 Then through the data cable we insert the
commands in the microcontroller ATMEGA 328P (UNO). These commands help the
microcontroller to interface with the Bluetooth module HC05 and also with the motor driver L293D.
Here the Bluetooth module act as a receiver which receives the instruction from the smart phone
(remote or transmitter). Then the microcontroller decides the operation for the instruction which
is coming from the smart phone. The functions of the given instructions are operated by the
microcontroller. The instructions are sent by the smart phone. We can easily control the
movements of the dc motor.
3.4 Arduino IDE
The Arduino integrated development environment (IDE) could be a cross-platform application
written in Java, and derives from the IDE for the process programing language and also the Wiring
projects. It’s designed to introduce programming to artists and different newcomers unfamiliar with
code development. It includes a code editor with options like syntax highlighting, brace matching,
and automatic indentation, and is additionally capable of compilation and uploading programs to the
board with one click. A program or code written for Arduino is named a "sketch".
12
Arduino programs are written in C or C++. The Arduino IDE comes with a code library referred
to as "Wiring" from the first Wiring project that makes several common input/output operations a
lot of easier. The users would like solely to outline 2 functions to create a possible cyclic
government program:
setup (): a function run once at the beginning of a program that may initialize setting
loop (): a function referred to as repeatedly till the board powers off+ programming
languages
13
3.5 Arduino Uno
The Arduino Uno is an open-source microcontroller board based on the Microchip ATmega328P
microcontroller and developed by [Link]. The board is equipped with sets of digital and
Analog input/output (I/O) pins that may be interfaced to various expansion boards (shields) and
other circuits. [1]. The board has 14 digital I/O pins (six capable of PWM output), 6 Analog I/O
pins, and is programmable with the Arduino IDE (Integrated Development Environment), via a
type B USB cable. It can be powered by the USB cable or by an external 9-volt battery, though it
accepts voltages between 7 and 20 volts. The Uno board is the first in a series of USB-based
Arduino boards; it and version 1.0 of the Arduino IDE were the reference versions of Arduino,
which have now evolved to newer releases. The ATmega328 on the board comes pre-
programmed with a bootloader that allows uploading new code to it without the use of an
external hardware programmer. There are many versions of Arduino boards introduced in the
market like Arduino Uno, Arduino Due, Arduino Leonardo, Arduino Mega, however, most
common versions are Arduino Uno and Arduino Mega.
Figure 3.5: Arduino Uno
14
3.6 Arduino Uno Features and Pin Configuration
Uno means one in Italian, as uno is the first microcontroller of arduino family.
Features of Arduino Uno:-
• The operating voltage is 5V
• The recommended input voltage will range from 7v to 12V
• The input voltage ranges from 6v to 20V
• Digital input/output pins are 14
• Analog I/P pins are 6
• DC Current for each input/output pin is 40 mA
• DC Current for 3.3V Pin is 50 mA
• Flash Memory is 32 KB
• SRAM is 2 KB
• EEPROM is 1 KB
• CLK Speed is 16 MHz
Power Supply
The Arduino Uno power supply can be done with the help of a USB cable or an external power
supply. The external power supplies mainly include AC to DC adapter otherwise a battery. The
adapter can be connected to the Arduino Uno by plugging into the power jack of the Arduino
board. Similarly, the battery leads can be connected to the Vin pin and the GND pin of the
POWER connector. The suggested voltage range will be 7 volts to 12 volts.
Input & Output
The 14 digital pins on the Arduino Uno can be used as input & output with the help of the functions
like pinMode(), digitalWrite(), & Digital Read().
Pin1 (TX) & Pin0 (RX) (Serial): This pin is used to transmit & receive TTL serial data, and these
are connected to the ATmega8U2 USB to TTL Serial chip equivalent pins.
15
Pin 2 & Pin 3 (External Interrupts): External pins can be connected to activate an interrupt over
a low value, change in value.
Pins 3, 5, 6, 9, 10, & 11 (PWM): This pin gives 8-bit PWM o/p by the function of analogWrite().
SPI Pins (Pin-10 (SS), Pin-11 (MOSI), Pin-12 (MISO), Pin-13 (SCK): These pins maintain SPI-
communication, even though offered by the fundamental hardware, is not presently included
within the Arduino language.
Pin-13(LED): The inbuilt LED can be connected to pin-13 (digital pin). As the HIGH-value pin,
the light emitting diode is activated, whenever the pin is LOW.
Pin-4 (SDA) & Pin-5 (SCL) (I2C): It supports TWI-communication with the help of the Wire
library.
AREF (Reference Voltage): The reference voltage is for the analog i/ps with analogReference().
Reset Pin: This pin is used for reset (RST) the microcontroller.
Memory
The memory of this Atmega328 Arduino microcontroller includes flash memory-32 KB for storing
code, SRAM-2 KB EEPROM-1 KB.
Communication
The Arduino Uno ATmega328 offers UART TTL-serial communication, and it is accessible on
digital pins like TX (1) and RX (0). The software of an Arduino has a serial monitor that permits
easy data. There are two LEDs on the board like RX & TX which will blink whenever data is being
broadcasted through the USB.
16
3.7 Arduino Uno Features and Pin Configuration
Uno means one in Italian, as uno is the first microcontroller of arduino family.
Features of Arduino Uno:-
• The operating voltage is 5V
• The recommended input voltage will range from 7v to 12V
• The input voltage ranges from 6v to 20V
• Digital input/output pins are 14
• Analog I/P pins are 6
• DC Current for each input/output pin is 40 mA
• DC Current for 3.3V Pin is 50 mA
• Flash Memory is 32 KB
• SRAM is 2 KB
• EEPROM is 1 KB
• CLK Speed is 16 MHz
Power Supply
The Arduino Uno power supply can be done with the help of a USB cable or an external power
supply. The external power supplies mainly include AC to DC adapter otherwise a battery. The
adapter can be connected to the Arduino Uno by plugging into the power jack of the Arduino
board. Similarly, the battery leads can be connected to the Vin pin and the GND pin of the
POWER connector. The suggested voltage range will be 7 volts to 12 volts.
Input & Output
The 14 digital pins on the Arduino Uno can be used as input & output with the help of the functions
like pinMode(), digitalWrite(), & Digital Read().
Pin1 (TX) & Pin0 (RX) (Serial): This pin is used to transmit & receive TTL serial data, and these
are connected to the ATmega8U2 USB to TTL Serial chip equivalent pins.
17
Pin 2 & Pin 3 (External Interrupts): External pins can be connected to activate an interrupt over
a low value, change in value.
Pins 3, 5, 6, 9, 10, & 11 (PWM): This pin gives 8-bit PWM o/p by the function of analogWrite().
SPI Pins (Pin-10 (SS), Pin-11 (MOSI), Pin-12 (MISO), Pin-13 (SCK): These pins maintain SPI-
communication, even though offered by the fundamental hardware, is not presently included
within the Arduino language.
Pin-13(LED): The inbuilt LED can be connected to pin-13 (digital pin). As the HIGH-value pin,
the light emitting diode is activated, whenever the pin is LOW.
Pin-4 (SDA) & Pin-5 (SCL) (I2C): It supports TWI-communication with the help of the Wire
library.
AREF (Reference Voltage): The reference voltage is for the analog i/ps with analogReference().
Reset Pin: This pin is used for reset (RST) the microcontroller.
Memory
The memory of this Atmega328 Arduino microcontroller includes flash memory-32 KB for storing
code, SRAM-2 KB EEPROM-1 KB.
Communication
The Arduino Uno ATmega328 offers UART TTL-serial communication, and it is accessible on
digital pins like TX (1) and RX (0). The software of an Arduino has a serial monitor that permits
easy data. There are two LEDs on the board like RX & TX which will blink whenever data is being
broadcasted through the USB.
18
Figure 3.6: Arduino Uno connections
19
3.8 HC-05 Bluetooth Module
HC-05 is a Bluetooth module which is designed for wireless communication. This module can be
used in a master or slave configuration. It is used for many applications like wireless headset, game
controllers, wireless mouse, wireless keyboard and many more consumer applications. It has range
up to to <100m which depends upon transmitter and receiver, atmosphere, geographic & urban
conditions. It is IEEE 802.15.1 standardized protocol, through which one can build wireless
Personal Area Network (PAN). It uses frequency hopping spread spectrum (FHSS) radio
technology to send data over air. It uses serial communication to communicate with devices. It
communicates with microcontroller using serialport (USART).
Figure 3.7: HC-05 Bluetooth Module
20
3.9 Bluetooth Communication Between Devices
First Send data from Smartphone terminal to HC-05 Bluetooth module and see this data on PC
serial terminal and vice versa. To communicate smartphone with HC-05 Bluetooth module,
smartphone requires Bluetooth terminal application for transmitting and receiving data. You can
find Bluetooth terminal applications for android and windows in respective app store. Before
establishing communication between two Bluetooth devices, 1st we need to pair HC-05 module
to smartphone for communication. First, search for new Bluetooth device from your phone. You
will find
Bluetooth device with ― HC-05 name. Second, click on connect/pair device option; default pin
for HC-05 is 1234 or 0000. In smart phone, open Bluetooth terminal application and connect to
paired device HC-05. It is simple to communicate we just have to type in the Bluetooth terminal
application of smartphone. Characters will get sent wirelessly to Bluetooth module HC-05.
Figure 3.8: Bluetooth communication between devices
21
3.10 Motor Driver Module L298N
This L298N Motor Driver Module is a high power motor driver module for driving DC and Stepper
Motors. This module consists of an L298 motor driver IC and a 78M05 5V regulator. L298N
Module can control up to 4 DC motors, or 2 DC motors with directional and speed control.
I. Features & Specifications :-
• Driver Model: L298N 2A
• Driver Chip: Double H Bridge L298N
• Motor Supply Voltage (Maximum): 46V
• Motor Supply Current (Maximum): 2A
• Logic Voltage: 5V
• Driver Voltage: 5-35V
• Driver Current:2A
• Logical Current:0-36mA
• Maximum Power (W): 25W
• Current Sense for each motor
• Heatsink for better performance
• Power-On LED indicator II.
Controlling a DC Motor :-
We can only have full control over a DC motor if we can control its speed and spinning
direction. This is possible by combining these two techniques.
1. PWM – to control speed
2. H-Bridge – to control the spinning direction
3. 1. PWM – to control speed
The speed of a DC motor can be controlled by changing its input voltage. A widely used technique
to accomplish this is Pulse Width Modulation (PWM).
21
22
2. H-Bridge – to control the spinning direction
The spinning direction of a DC motor can be controlled by changing the polarity of its input
voltage. A widely used technique to accomplish this is to use an H-bridge.
An H-bridge circuit is made up of four switches arranged in a H shape, with the motor in the
center.
Closing two specific switches at the same time reverses the polarity of the voltage applied to the
motor. This causes a change in the spinning direction of the motor.
Figure 3.9: Motor driver module L298N
23
3.11Voltage Drop of L298N
The L298N has a voltage drop of approximately 2V. This is due to the fact that internal switching
transistors have a voltage drop of approximately 1V when forward biased, and because an H- Bridge
requires the current to pass through two transistors, the total voltage drop is 2V.
So, if you connect 12V to the motor power supply terminal, the motors will receive approximately
10V. This means that a 12V DC motor will never spin at full speed.
In order to get the motor to run at its maximum speed, the motor power supply should have a
voltage that is slightly higher (+2V) than the actual voltage requirement of the motor.
Taking into account a voltage drop of 2V, if you are using 5V motors, you will need to provide
7V at the motor power supply terminal. If you have 12V motors then your motor supply voltage
should be 14V.
Figure 3.10: Voltage Drop of L298N
24
3.12 Micromors and Grippy Wheels
Mobile wheeled or tracked robots have a minimum of two motors which are used to propel and
steer the robot. Hobbyists tend to choose “skid steering” (like a tank) because of its simplicity to
design, incorporate and control. A three wheeled robot‘s third (front) wheel usually prevents the
robot from falling over. Four wheeled robots have either two or four drive motors and use skid
steering. Six wheeled robots most commonly have either two, four or six drive motors. Adding
―idle wheels (wheels not connected to a motor) often has the unfortunate consequence of
removing weight from the drive wheels resulting in slip and loss of traction, the center wheel often
loses contact with the ground.
Figure 3.11: Micromotors and Grippy Wheels
25
3.13 Jumper Wires
A jump wire (also known as jumper, jumper wire, DuPont wire) is an electrical wire, or group of
them in a cable, with a connector or pin at each end (or sometimes without them – simply "tinned"),
which is normally used to interconnect the components of a breadboard or other prototype or test
circuit, internally or with other equipment or components, without soldering.
Figure 3.12: Jumper Wires
26
3.14 Batteries
A battery is an energy source consisting of one or more electrochemical cells and terminals on
both ends called an anode (-) and a cathode (+). Here for the project, 3 3.7V Lithium-ion batteries
has been used.
Figure 3.13: Batteries
27
3.15 Switch
A switch is used for turning on the power of the vehicle. It controls the battery
connection of the vehicle.
Figure 3.14: Switch
28
3.16 Smartphone with app
The smart phone is the transmitter of this circuit. It sends the data to microcontroller through
Bluetooth module. It also helps to send the instruction of forward, backward, left, right to the
microcontroller. Actually, the smart phone is used as a remote of this system. Here we operating
the Bluetooth RC Controller application as the operating remote of this system. The advantage of
this project is that the application software designed for android phones is kept simple but attractive
with all necessary built-in functions. The novelty lies in the simplicity of the design and
functioning.
Figure 3.15.2: Application Interface
3.17 ESP32-CAM module
The ESP32-CAM is a small, powerful, and low-cost microcontroller module that comes with an
integrated OV2640 camera, Wi-Fi, Bluetooth, and support for running lightweight AI models. It is
widely used in projects like surveillance, IoT applications, object detection, wireless image
streaming, and embedded AI tasks.
This module allows you to capture live images and process them locally, making it ideal for
portable,
battery-powered smart systems like your gesture-controlled rover.
29
The ESP32-CAM is connected to the same battery supply, ensuring stable operation.
The camera captures live images and runs the AI model for object detection directly on the device.
Though the ESP32-CAM operates independently from the movement control, it shares the same power
source for seamless integration.
The two DC Motors connected to the L298N output terminals are responsible for moving the rover in
forward, backward, left, or right directions based on the Arduino's instructions.
All modules, including the Arduino, ESP32-CAM, HC-05, L298N, and Battery Pack, share a common
ground to maintain consistent electrical reference, which is essential for proper operation
Pin Name Function
3V3 Provides 3.3V regulated power to the module.
GND Ground connection (common for the entire system).
IO0 Used for flashing the firmware during programming mode.
IO2 General-purpose I/O pin, often used for onboard LED control.
IO4 / IO5 General-purpose digital I/O pins, can be used for peripheral connections or sensor inputs.
TXD0/RXD0 Serial communication pins for programming or debugging. Connects to USB-to-Serial
30
converter.
IO12 / IO13 / IO14 / IO15 General-purpose I/O pins. Some may have specific functions like camera
communication.
IO16 / IO17 GPIOs with optional use for peripheral connections or custom tasks.
IO18 / IO19 / IO21 / IO22 / IO23 GPIOs used for camera interface or external device control.
OV2640 Camera Connector Dedicated slot for connecting the 2MP OV2640 camera for image
capture.
31
CHAPTER 4
DESIGN AND IMPLEMENTATION
4.1 Wiring L298N Motor Driver Module to Arduino UNO
Initial procedure is done by connecting the motor power supply. In our experiment, we are using
DC gearbox motors, also called “TT” motors, which are often found in two-wheel-drive robots.
They are rated for 3 to 12V. We will therefore connect an external 12V power source to the VS
terminal. Because L298N has a voltage drop of about 2V, the motors will receive 10V and spin at
a slightly lower RPM. But that’s okay.
Next, we need to supply 5V to the logic circuitry of the L298N. We’ll use the on-board 5V
regulator to draw 5V from the motor power supply, so keep the 5V-EN jumper in place.
Now connect the L298N module’s Input and Enable pins (ENA, IN1, IN2, IN3, IN4 and ENB) to
the six Arduino digital output pins (9, 8, 7, 5, 4 and 3). Note that both Arduino output pins 9 and
3 are PWM-enabled.
Finally, wire one motor to terminal A (OUT1 and OUT2) and the other to terminal B
(OUT3 and OUT4). You can swap out your motor’s connections. There is technically no right or
wrong way.
Figure 4.1: Connecting Motor Driver Module to an Arduino
31
32
CHAPTER 5
Edge impulse
Edge Impulse is an easy-to-use, online platform that helps developers build machine learning (ML)
models for small, low-power devices like microcontrollers, sensors, and embedded boards. It allows
users to train, test, and deploy AI models directly onto hardware devices, without needing powerful
computers or complex coding knowledge.
With Edge Impulse, you can collect data, label it, and create AI models that perform tasks like object
detection, sound recognition, or motion detection. Once the model is ready, it can be converted into
lightweight code that runs efficiently on devices such as the ESP32-CAM.
The biggest advantage of Edge Impulse is that it allows AI models to run directly on the device, a
process called Edge AI. This means the system does not depend on cloud servers or internet connectivity
for processing. As a result, tasks like object detection can happen in real-time with low power
consumption.
In this project, Edge Impulse is used to train an AI model that helps the ESP32-CAM detect objects and
obstacles around the rover, enhancing its ability to respond intelligently to its surroundings.
33
Feature Extraction
The Edge Impulse system analyzes the labeled images to extract important features like edges, shapes,
and color patterns. This process converts images into a simplified format that the AI can understand.
Model Training
Using the labeled data, Edge Impulse trains a machine learning model, often a small Convolutional
Neural Network (CNN), to recognize different objects. The model "learns" to detect and classify objects
like bottles, chairs, or persons based on the patterns in the images.
Model Testing
After training, the model is tested on new images to check its accuracy. If it performs well, it means the
AI is ready to detect objects in real-world scenarios.
Model Optimization
The final trained model is converted into a compact, optimized version suitable for small devices like
the ESP32-CAM. This makes sure it uses minimum memory and power but still works effectively.
Image Conversion and Processing in ESP32-CAM
Image Acquisition
The ESP32-CAM module is equipped with an OV2640 camera sensor, a 2-megapixel CMOS sensor.
This sensor captures light from the environment and converts it into analog electrical signals
corresponding to pixel intensity and color information.
Analog-to-Digital Conversion (ADC)
The electrical signals generated by the image sensor are passed through an integrated Analog-to-Digital
Converter (ADC).
The ADC digitizes the signals into discrete numerical values that represent pixel intensity levels (for
grayscale) or Red-Green-Blue (RGB) color components for colored images.
Each pixel's color information is represented as binary data, typically 8 bits per color channel in standard
formats.
Analog-to-Digital Conversion (ADC)
The electrical signals generated by the image sensor are passed through an integrated Analog-to-Digital
Converter (ADC).
The ADC digitizes the signals into discrete numerical values that represent pixel intensity levels (for
grayscale) or Red-Green-Blue (RGB) color components for colored images.
Each pixel's color information is represented as binary data, typically 8 bits per color channel in standard
formats.
34
Data Representation in Binary
The complete image is now stored in the microcontroller's memory as a binary matrix, where each
element represents the intensity or color value of a pixel in binary form (0s and 1s).
For example, a pixel with an RGB value like (128, 64, 255) is internally stored as:
10000000 01000000 11111111
This binary image data can be stored temporarily in formats like JPEG, RGB arrays, or grayscale
buffers, depending on the application.
Pre-processing and Feature Extraction
Before object detection, the binary image data may undergo pre-processing, including resizing,
grayscale conversion, or normalization to reduce computational load.
Feature extraction algorithms or layers of a Convolutional Neural Network (CNN) analyze the pixel
patterns to detect edges, contours, textures, and spatial features necessary for object classification.
35
CHAPTER 6
RESULT & ANALYSIS
6.1 Resultant and Analysis
The proposed Gesture Controlled Rover with Visual Processing was developed, implemented, and tested
under various controlled conditions to evaluate its performance, functionality, and real-world usability.
The results of the system were observed based on object detection accuracy, gesture control
responsiveness, power efficiency, and overall system reliability.
1. Object Detection Performance
The AI model trained using Edge Impulse was deployed onto the ESP32-CAM for real-time object
detection. The following observations were recorded:
The system successfully detected common objects such as bottles, laptops, pens, chairs, and humans.
In well-lit indoor environments, the object detection accuracy ranged between 75% to 85%,
depending on the object type and distance.
The system struggled with detecting objects in low-light conditions or when the objects were
partially hidden.
Small or reflective objects occasionally resulted in false negatives, which can be improved with
better datasets in future upgrades.
Gesture Control Responsiveness
The gesture control feature was tested using a custom-designed smartphone application
communicating via Bluetooth (HC-05 Module):
The rover responded effectively to basic gesture commands including forward, backward, left, right,
and stop.
The average command response time from the smartphone to motor actuation was recorded to be
less than 500 milliseconds, ensuring real-time and smooth control.
Within the 10 to 15-meter Bluetooth range, the connection was stable and reliable.
Beyond the effective range, command delays and occasional disconnections were observed, as
expected with Bluetooth limitations.
3. Power Efficiency and Operation Time
The system was powered using a Lithium-ion battery pack, and its performance was observed:
The average continuous operational time was approximately 2 hours on a fully charged battery.
Power consumption remained low due to the use of optimized AI models and energy-efficient
components like the ESP32-CAM and Arduino UNO.
The system operated effectively without overheating or significant voltage fluctuations.
36
4. System Stability and Reliability
The rover performed well on smooth indoor surfaces such as tiles and flat outdoor grounds.
Minor instability was observed when navigating rough or uneven terrains, indicating the system is
more suitable for flat
areas.
The onboard AI worked independently of internet connectivity, making the system reliable for
remote or isolated locations.
Limitations Noted During Testing
Object detection struggled under poor lighting.
Gesture control limited to Bluetooth range only.
Frame rate for object detection was around 1 to 2 frames per second, suitable for basic tasks but not
high-speed scenarios.
Overall Analysis
The project successfully demonstrated the integration of gesture-based control and edge AI for real-
time visual processing in a portable robotic platform. The combination of wireless control, onboard
object detection, and energy-efficient hardware presents a practical, low-cost solution for
surveillance, educational robotics, and basic autonomous navigation tasks.
While the system performed effectively within its intended scope, further hardware upgrades and AI
model improvements are recommended to overcome the identified limitations and extend its real-
world applications.
37
Figure 6.2: Working Explanations.
6.2 Applications
• Remote Surveillance of restricted or unsafe areas.
• Disaster Rescue Operations in collapsed buildings or danger zones.
• Wildlife Monitoring in forests and protected areas.
• Farm Security for detecting animal intrusion or crop monitoring.
• Border Patrolling for security surveillance.
• Warehouse Security for inspecting large storage spaces.
• School and College Robotics Projects for learning and demonstrations.
6.3 Advantages
o The Bluetooth communication offers reliable wireless control within a reasonable range.
o The AI model running on the ESP32-CAM ensures real-time object detection with low
power consumption.
o Helps to reduce human risk by allowing machines to explore dangerous or harmful
environments in place of people.
o Supports faster search and rescue operations during disasters like building collapses or
floods.
o Allows for monitoring of hard-to-reach areas, improving safety in industries, military
zones, and disaster sites.
o Reduces the need for human presence in hazardous zones, such as chemical plants or
38
radiation-affected areas.
o Can assist elderly or differently-abled individuals by providing remote-controlled
monitoring tools.
o Improves the efficiency of security systems, helping prevent theft or unauthorized access
in public or private spaces.
o Promotes responsible use of technology, where machines are used to assist, not replace,
human workers.
39
CHAPTER 7
CONCLUSION & FUTURE SCOPE
7.1 Conclusion
The development of a Gesture Controlled Rover with Visual Processing using the ESP32-CAM
successfully demonstrates how modern embedded systems and AI technologies can be combined to
create a portable, low-cost, and intelligent robotic platform. The project effectively integrates two
important features—gesture-based wireless control and real-time object detection using onboard AI.
The system allows users to control the movement of the rover through intuitive hand gestures via a
smartphone application, eliminating the need for complex remote controllers. Simultaneously, the
ESP32-CAM module captures live images of the surroundings and processes them locally using a
lightweight machine learning model trained with Edge Impulse. This enables the rover to detect objects
such as bottles, people, laptops, or obstacles in real time.
The project was successfully tested under controlled conditions, achieving satisfactory results in terms
of object detection accuracy, system responsiveness, and power efficiency. The results validate the
feasibility of using compact hardware like the ESP32-CAM for AI-based visual processing in real-world
robotics applications.
While certain limitations such as low-light performance, limited object categories, and restricted control
range were identified, the system provides a strong foundation for future improvements. The rover’s
modular and scalable design ensures that with further enhancements like advanced AI models, longer-
range communication, orautonomous navigation, the system can be adapted for broader applications
such as surveillance, search and rescue, environmental monitoring, and educational robotics.
Overall, the project meets its primary objectives of providing an affordable, portable, and easy-to-use
robotic system that combines gesture control with visual intelligence, showcasing the potential of
accessible AI and embedded technologies for practical use in human-friendly robotic platforms.
7.2 Future Scope of the Project
The system can be upgraded to include voice control along with gesture control for easier operation.
Advanced AI models can be implemented to recognize a wider variety of objects with higher accuracy.
The ESP32-CAM can be replaced with a more powerful module like ESP32-S3 or Raspberry Pi Zero 2
W for better processing and image quality.
Long-range communication technologies like LoRa, GSM, or 5G modules can be added for wider
40
operational range.
A solar-powered charging system can be developed for extended outdoor use without manual charging.
The system can be enhanced with night vision cameras to improve object detection in low-light
conditions.
The rover can be used for healthcare purposes, such as delivering medicines or monitoring patients in
quarantine zones.
The project can be expanded into a fully autonomous robot, capable of making decisions without human
input.
More efficient battery technology can be adopted to increase operational time.
A dedicated mobile application with an improved user interface can make gesture control more reliable
and user-friendly.
REFERENCES
1 Book References
2 S. Rajesh and P. Kumar, "Introduction to Embedded Systems", McGraw Hill Education,
2020.
3 (Covers fundamentals of microcontrollers like Arduino and ESP32, ideal for understanding
system
design.)
4 K. A. Krishnamurthy, "Wireless Communication Principles and Applications", Tata McGraw
Hill, 2019.
5 (Explains Bluetooth technology and wireless control principles relevant to this project.)
6 M. A. Mazidi, "The 8051 and Embedded Systems: Using Assembly and C", Pearson
Education, 2018.
41
7 (Although 8051-based, the embedded system principles apply to Arduino and ESP32-based
projects.)
8 References
9 Espressif Systems. “ESP32-CAM Datasheet.” [Online]. Available:
[Link]
10 Arduino. “Arduino UNO Technical Specifications.” [Online]. Available:
[Link]
11 Edge Impulse. “Edge Impulse Documentation.” [Online]. Available:
[Link]
12 TensorFlow Lite. “TensorFlow Lite for Microcontrollers.” [Online]. Available:
[Link]
13 HC-05 Bluetooth Module. “Technical Specifications.” [Online]. Available:
[Link]
14 OV2640 Camera Module. “2MP Camera Sensor Specifications.” [Online]. Available:
[Link]
41
15 Microchip Technology. “ATmega328P Datasheet.” [Online]. Available:
[Link]
Microcontrollers-ATmega328P_Datasheet.pdf
16 L298N Motor Driver Module. “Specifications and Application Notes.” [Online]. Available:
[Link]
17 B. Preddy et al., “Real-Time Object Detection on Embedded Systems,” International Journal
of Embedded Systems, vol. 12, no. 4, pp. 295–303, 2021.
42
18 M. Patel and S. Singh, “Gesture Controlled Robotics: An Overview,” International Journal of
Robotics Research, vol. 9, no. 2, pp. 45–56, 2022.
19 Wang, F.Q. (2012) Research New Type of Supply Chain Management Model Based on the
Technology of Internet of Things—With Radio Frequency Identification (RFID) Technology as
an Example.
20 Shen, S.B., Fang, Q.L. and Zong, P. (2009) Research of Internet of Things and related
technology. Journal of Nanjing University of Posts and Telecommunications (Natural Science),
6, 1-11.
21 Wang, B.Y. (2009) Summary of IOT Technology Research. Journal of Electronic Measurement
and Instrument, 12, 1-7.
22 Lin, G. (2012) Research on Complex Event of Supply Chain Decision Support Based on Internet
of Things.
23 Han, Y. and Chai, Y.T. (2011) The Architecture Reference of Flexible Supply Chain Simulation
System.
Journal of System Simulation, 6, 1270-1278.
24 J. Ziegler, M. Graube, L. Urbas, RFID as universal entry point to linked data clouds, in: Proc.
IEEE
International Conference on RFID Technologies and Applications, 2012, pp. 281–286.
R. Want, An introduction to RFID technology, IEEE Pervasive Comput. 5 (2006) 25– 33.
25 M. Gireesh Kumar, K. Sripath Roy, Zigbee based indoor campus inventory tracking using RFID
module, Int. J. Eng. [Link]. (IJERA) 4 (7(1)) (2014) 132, ISSN 2248-9622.
26 Son Minh Huynh, David Parry, A.C.M. Fong, Novel RFID and ontology based home
localization system for misplaced objects, IEEE Trans. Consum. Electron. 60 (3) (2014).
27 J. Gubbi, R. Buyya, S. Marusic, M. Palaniswami, Internet of things (IoT): a vision architectural
elements and future directions, Future Gener. Comput. Syst. 29 (7) (Sep. 2013) 1645–1660.
43
42
44