DESIGN THINKING AND INNOVATION REPORT

entitled

“EASY GO”

Traffic Reduction using Machine Learning

Submitted in partial fulfillment of the requirements for the award of

Bachelor of Engineering degree in Computer Science and Engineering with
specialization in

by
V. NAVADEEP

Reg.No: 42612050

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

SCHOOL OF COMPUTING

SATHYABAMA
INSTITUTE OF SCIENCE AND TECHNOLOGY
(DEEMED TO BE UNIVERSITY)
Accredited with Grade by NAAC|12B Status by UGC| Approved by AICTE

JEPPIAAR NAGAR, RAJIV GANDHISALAI,

CHENNAI 600119.

OCTOBER 2023
 ACKNOWLEDGEMENT

I am pleased to acknowledge my sincere thanks to Board of Management of

SATHYABAMA for their kind encouragement in doing this project and for
completing it successfully. I am grateful to them.

I convey my thanks to Dr. T.Sasikala M.E., Ph.D., Dean, School of Computing,

Dr. S.Vigneshwari M.E., Ph.D., Head of the Department of Computer Science and
Engineering for providing me necessary support and details at the right time
during the progressive reviews.

I would like to express my sincere and deep sense of gratitude to my Internal

Guide Ms. K. Abinaya for his/her valuable guidance, suggestions and
constant encouragement which paved way for the successful completion of my
Design Project/Product.

I wish to express my thanks to all Teaching and Non-teaching staff members of

the Department of Computer Science and Engineering who were helpful in many
ways for the completion of the project.

iv
 ABSTRACT

Metropolitan cities worldwide face escalating challenges

related to traffic congestion and road safety, particularly during rush hours. To
achieve the following objectives this project introduces an innovative approach to
overcome these issues by making use of the harnessing advanced sensors, machine
learning and other technologies. The project aims to create and implements an
intelligent traffic regulation system that relies on real-time data by deploying a
network of cutting-edge sensors at key locations across the metropolitan areas. These
sensors utilize technologies like LIDAR, GPS and many more. The collected data is
processed in real-time using machine learning algorithms to make informed decisions
regarding traffic signal timing, speed limits, and incident response. By dynamically
adjusting traffic signals and speed limits based on real-time traffic conditions, we can
reduce travel times during rush hours.

v
 TABLE OF CONTENTS

S.NO CHAPTER TITLE PAGE

NO. NO.
1 1 ABSTRACT 8
INTRODUCTION
2 1.1 INTRODUCTION
1.2 OBJECTIVES AND
3 METHODS
4 2 EXISTING SYSTEM
5 3 LIMITATIONS OF THE 10
EXISTING SYSTEMS

6 3.1 LIMITED COVERAGE

AND AVAILABILITY
7 3.2 DELAYED
RESPONSE TO
CHANGES

8 3.3 LACK OF
CONNECTIVITY

9 4 PROPOSED
SYSTEM
10 4.1 ARCHITECTURE 11
11 4.2 ALGORITHMS 15
AND METHODS
USED
12 4.3 LANGUAGE USED
FOR DESIGNING 17

13 4.4 EXPERIMENTAL 19
RESULTS

14 5 CONCLUSION
15 6 REFERENCE
20

vi
 LIST OF FIGURES

Figure Figure Name Page No.

No.

FIG 2.1 STATIC SPEED DISPLAY SYSTEM 9

FIG 4.1.1 COMPONENTS OF SYSTEM 12

FIG 4.1.2 VEHICLE DETECTION 13

FIG 4.1.3 SPEED LIMIT CALCULATION 13

FIG 4.2.1 FLOWCHART REPRESENTATION 16

FIG 4.3.1 CODE REPRESENTATION 18

vii
 CHAPTER 1

1. ABSTRACT INTRODUCTION
1.1 Introduction
By using the Artificial Intelligence, machine learning and integrating sensors to
track vehicle speed and other critical aspects, this initiative seeks to enhance traffic
management and improve overall urban livability.
This project introduces an innovative approach to address these issues by harnessing
advanced sensor technology. We aim to monitor and regulate traffic during peak times in
urban areas effectively.
1.2 Objectives and methods:
The primary objective of this project is to develop and implement an intelligent
traffic regulation system that leverages real-time data from strategically positioned sensors
and gets processed by using machine Learning models. These sensors will monitor critical
aspects such as vehicle speed, congestion patterns and road conditions. Our goal is to
optimize traffic flow, and enhance the road safety.
For this we deploy a network of cutting-edge sensors like 3D Cameras and
Depth Sensors, radar sensors, LiDAR, GPS etc... and the data collected through this sensor
is used to build an efficient machine leaning model that uses the CNN and algorithms like
decision trees, Naïve Bayse, Support Vector Machine (SVM), and Bayes Net and etc. Then
this algorithms analysis the data collected and displays the speed to by maintained in that
specific area.

2. EXISTING SYSTEM:
Several existing systems focus on traffic speed display and control to enhance road safety
and traffic management. Here are a few examples:

Traffic Light Control Systems:

Advanced traffic light control systems adjust signal timings based on traffic flow and
demand, improving traffic flow and reducing congestion.
A traffic light control system is an automated system used to manage the operation and
timing of traffic lights at intersections and various road junctions. The primary goal of such
a system is to efficiently control the flow of vehicular and pedestrian traffic to enhance road
safety, minimize congestion, and improve overall traffic management.

ISA systems use GPS and mapping technology to inform

drivers of the current speed limit and can provide warnings or even actively intervene to
prevent the vehicle from exceeding the speed limit.
8
Static Speed Display System:
A static speed display control is a system that governs the operation and
management of static speed display signs placed along roads and highways. These signs
typically provide information about the current speed limit to approaching drivers,
encouraging them to adhere to the prescribed speed for that particular road segment. The
control system ensures that the displayed speed limits are accurate, up to date, and
synchronized with traffic regulations.

FIG 2.1

Speed Limiting Devices (SLDs):

SLDs are devices installed in vehicles to limit their maximum speed, ensuring
compliance with speed limits and enhancing road safety. Variable Speed Limits (VSLs):
VSL systems dynamically adjust speed limits on highways to match traffic flow and
conditions, aiming to reduce congestion and enhance safety.
Cooperative Intelligent Transport Systems (C-ITS):
C-ITS use vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I)
communication to provide real-time traffic information and speed limit recommendations
to driver.

Ramp Metering Systems:

These systems control the rate at which vehicles enter highways by dynamically
adjusting traffic signal lights on on-ramps, optimizing traffic flow and preventing
congestion.

9
3. LIMITATIONS OF THE EXISTING SYSTEMS
While existing traffic speed limit systems have made significant advancements in managing
and regulating traffic speed, they still face several limitations. Here are some common
limitations of current traffic speed limit systems:

3.1 LIMITED COVERAGE AND AVAILABILITY

Dynamic speed limit systems may not be widespread, limiting their
effectiveness to specific areas or highways. Many roads still rely on traditional static speed
limit signs. We can also see Lack of Real-Time Adaptability which means that many speed
limit signs and systems are static and not adaptable in real time. They do not adjust
dynamically based on changing traffic conditions, weather, or accidents.
Deploying dynamic speed limit systems, especially with advanced sensors and
realtime data processing capabilities, can be expensive, limiting widespread
implementation which is due to high implementation costs.

3.2 DELAYED RESPONSE TO CHANGES

Even if speed limits are adjustable, the response time to update the displayed
speed limits may be slow, impacting the system's ability to promptly address evolving
traffic situations. Inaccurate or Outdated Information would also lead to delay in responses,
where the speed limit databases may be outdated or inaccurate, leading to incorrect speed
limit displays and potential confusion for drivers.
Maintenance and Reliability Challenges is also one of the limitations in which
maintenance of sensors and other infrastructure components can be challenging, and
failures or inaccuracies in the system can lead to unsafe driving conditions .

3.3 LACK OF CONNECTIVITY

Lack of connectivity would often rely on extensive infrastructure, including
sensors, cameras, and variable message signs due to malfunctions or maintenance issues
which can disrupt the system and connectivity.
If any construction or changes are done in the road patters then data is to be
updated according to the availability and these are the few aspects on which they have to
work on.

Future advancements in technology and integration with connected and autonomous

vehicles could help overcome some of these limitations. However, addressing these
challenges is crucial to designing more effective, adaptable, and widely accessible traffic
speed limit systems.

10
4.PROPOSED SYSTEM

4.1 ARCHITECTURE:
Designing an architecture for a dynamic traffic speed display and control system
involves integrating various components to provide real-time speed information to drivers
and dynamically adjust speed limits based on traffic conditions. Here's a highlevel
architecture for such a system:

Data Sources:
We can get the data from the sensors.

Traffic Sensors to be used:

In the context of dynamic speed displays and traffic management, "cutting
edge" methods and technologies represent the latest and most advanced approaches that
utilize stateof-the-art solutions to achieve optimal traffic flow, safety, and efficiency.
Cutting-edge methods may involve utilizing artificial intelligence, machine learning
algorithms, real-time data analytics, advanced sensors, high-resolution displays, and
intelligent communication systems. These technologies work in synergy to provide up-to-
the-minute information, optimizing speed limits dynamically based on traffic conditions,
weather, road incidents, and other relevant factors.
Deploy a network of traffic sensors, including radar, LiDAR, cameras, and
inductive loops and Include weather sensors to monitor weather conditions, such as rain,
snow, fog, and visibility, which can impact safe driving speeds at key locations to
continuously monitor traffic speed and conditions.

Cutting Edge Sensors:

LiDAR (Light Detection and Ranging)
Radar Speed Sensors
Video Cameras
Infrared Sensors
GPS (Global Positioning System) etc.

11
 FIG 4.1.1

Communication Infrastructure:

Establish a robust communication infrastructure, such as a high-speed data network

or cellular connectivity, to transmit real-time data from the sensors to the central control
system. Then the central control system process and analyze the incoming data in real-
time using advanced analytics, sensors and machine learning algorithms to assess
traffic conditions and weather impacts. Utilizing the wireless communication
technology, traffic management systems can transmit data and speed limit updates to
the dynamic speed displays. This allows for instantaneous adjustments to speed limits
based on traffic conditions.
GPS technology can also be used to determine a vehicle's location and adjust the
displayed speed limit on dynamic speed displays according to the specific area or road
segment the vehicle is traveling on.

12
 FIG 4.1.2
Speed Limit Calculation:
Calculate optimal speed limits based on real-time traffic data, road conditions and
give conditions to the machine learning model to act accordingly. The machine learning
model uses the suitable algorithm to calculate and displays the speed that the vehicles
should maintain. Variable Message Signs (VMS) are electronic displays typically placed
above or alongside the road that can change the displayed speed limit based on real-time
traffic conditions, accidents, or weather. They often use LED technology for better
visibility and can be remotely controlled to update the displayed information to drivers.
We have to equip vehicles with in-vehicle displays that receive and display
realtime speed limit information from the VMS, ensuring drivers are aware of the current
recommended speed limit.

FIG 4.1.3

13
ARCHITECTURE OF DYNAMIC SPEED DISPLAY SYSTEM:

Initially we have to built a Machine Learning Model by analyzing the data

regarding the number of road patterns, city expandability, route details, recent updates in
roads and other data required for the model. Then using the algorithms in Machine Learning
we give an efficient model for calculation of speed which is to be displayed and to be
followed by the drivers. The speed to be maintained is calculated by the methods mentioned
above and by considering many aspects like peak hours of traffic, traffic congestion in
specific areas and many more...
All the vehicles should maintain that specific speed which is been projected on the
screen. If there is any vehicle not following the speed which is to be maintained then the
sensors placed at key locations calculates the speed of vehicle and captures the vehicle
number and the number is analyzed with the data present. And the corresponding license
holder receives the message about the violation of traffic rules.

This is how a Machine Learning algorithm is built to design an architecture for

"DYNAMIC SPEED DISPLAY SYSYTEM" in smart traffic management in the
metropolitan cities. We have to implement security protocols to protect the system from
cyber threats and unauthorized access, ensuring the integrity and reliability of the
information provided to drivers.

This architecture aims to provide drivers with real-time speed limit information
based on traffic and environmental conditions, promoting safer driving practices and
reducing traffic congestion. Continuous monitoring, data analysis, and system feedback are
essential for its effectiveness and success in improving road safety and traffic management.
4.2 ALGORITHMS AND METHODS USED

Proposed method:
A new approach to classifying network traffic using deep learning and a
metaheuristic algorithm is presented. A general view of the proposed method is shown in
Fig. 3. The proposed method contains four phases described as follows.

1. Preprocessing of the traffic flows with entropy and variance of entropy.

2. Automated feature extraction using 1D-CNN.
3. Efficient feature selection using the ant-lion meta-heuristic algorithm.
4. New instances classification using a combined fuzzy-SOM based.

14
This section provides an overview of the main methods used to classify network traffic. In
general, these methods can be categorized into two main groups: (1) machine learning
methods and (2) deep learning methods.
(1) Machine Learning Model:
Advanced algorithms can analyze real-time traffic data, weather conditions, historical
traffic patterns, and other variables to predict the most suitable speed limits for a given time
and adjust the dynamic speed display accordingly.
The essential and beneficial points of the proposed method are:

Preprocessing traffic data using entropy and entropy variance.

Automated feature extraction from the hidden layers of the one-dimensional
convolutional neural network (1D-CNN).
Efficient feature selection with high classification accuracy using ALO.
-SOM based-clustering.

Algorithms used:
There are many studies on traditional machine learning techniques, some
of which are briefly described here. The mainly used various machine learning classifiers
such as C4.5 decision trees, Naïve Bayse, Support Vector Machine (SVM), and Bayes Net.

15
 FIG 4.2.1

(2) Deep Learning Models:

Deep learning is increasingly being applied in dynamic speed display systems within
traffic management to enhance the effectiveness of speed limit adjustments and improve
overall road safety and traffic flow. Here are ways in which deep learning is utilized in this
context: Traffic Pattern Analysis:
Deep learning models can analyze historical traffic patterns, considering factors
like time of day, day of the week, weather conditions, and special events. By recognizing
recurring patterns, the system can predict traffic congestion and set appropriate speed
limits in real-time.

16
 Deep learning algorithms can process real-time traffic data from various sensors
(e.g., radar, cameras, and traffic flow sensors) to identify traffic density, vehicle speeds, and
potential congested areas. The system can then dynamically adjust speed limits based on
this analysis. The models can be trained to identify unusual patterns or anomalies in traffic
behavior, such as sudden stops, wrong-way drivers, or accidents. Upon detection,
appropriate speed limits can be displayed to warn drivers and ensure safety.

Intelligent Decision Making:

Deep learning algorithms can be trained to make intelligent decisions regarding
speed limit adjustments. The model can consider a wide array of inputs such as traffic
volume, accident history, road type, and construction zones to dynamically set optimal
speed limits. It helps in proactively setting speed limits to avoid congestion and maintain a
smooth traffic flow.

4.3 LANGUAGE USED FOR DESIGNING ML MODEL

Python is a highly popular programming language for designing machine learning

(ML) models due to several key advantages. Python is known for its simple and readable
syntax, which makes it easier to write and understand code. This feature is particularly
important when developing and maintaining complex ML models.

Rich Ecosystem of Libraries:

Python has a vast and robust ecosystem of libraries and frameworks specifically
built for ML, such as TensorFlow, PyTorch, scikit-learn, Keras, and NLTK. These libraries
provide prebuilt functions, tools, and algorithms that significantly simplify the development
process. It is a versatile language that can be used for various purposes, including data
analysis, visualization, web development, and automation. This versatility allows data
scientists and ML practitioners to integrate ML seamlessly into larger applications.
Availability of Frameworks for Deep Learning:
Python has become the primary language for deep learning due to the availability of
popular deep learning frameworks like TensorFlow, PyTorch, and Keras. These
frameworks are widely used for building and training deep neural networks.

17
 Here we use TENSER FLOW and KERAS popular and powerful tools for
implementing machine learning and deep learning models in Python, especially for
traffic classification tasks. This is a simple example of a "Convolutional Neural
Network (CNN)" model. Adjust the architecture based on the specifics of your traffic
classification problem.

This is the example code in which python used in designing ml models

FIG 4.3.1
4.4 Experimental results
To implement the proposed traffic classification method, the tensorflow
framework, and the Keras library of the Python language were used. In this section, the
dataset, the model evaluation methods, and the analysis of the results and the problems
around them are discussed.

18
 5.CONCLUSION
In conclusion, implementing a dynamic traffic speed limit and display system is a promising
approach to improving road safety, traffic flow, and overall transportation efficiency. By
benefits:

Enhanced Safety:
Real-time speed limit adjustments and immediate display updates help in promoting
safer driving habits by aligning speed limits with current traffic conditions, reducing the
risk of accidents. Reduced Congestion:
Dynamically adjusting speed limits optimizes traffic flow and minimizes congestion
during peak hours, leading to shorter travel times and a smoother driving experience.

Adaptability to Changing Conditions:

The system can swiftly respond to changing traffic patterns, accidents, and adverse
weather conditions, ensuring optimal speed limits are always in place to maintain safety
and efficiency.

Improved Driver Awareness:

In-vehicle displays and electronic variable message signs provide drivers with real-time
speed limit information, fostering better compliance with speed regulations and promoting
responsible driving behavior.

By continuously evolving the system, incorporating the latest advancements in

sensor technology, artificial intelligence, and communication infrastructure, we can create
a safer, more efficient, and sustainable urban transportation network that addresses the
challenges of modern traffic and enhances the overall quality of life for commuters.

19
 6. REFERENCE:

Traffic System with Real Time Data Analysis Using IoT”, IEEE ICCTCT 2018 –Volume 2,
6145
[2] Piotr Burnos, Janusz Gajda, Piotr Piwowar, Ryszard Sroka, Marek Stencel,
Tadeusz

187 203, 2007.

[3] JOURNAL OF ADVANCED TRANSPORTATION:

https://www.hindawi.com/journals/jat/2021/8878011/fig1/

[5] Intelligent Transportation Systems

Functional Design for Effective Traffic Management
By Robert Gordon, 2015

[6] Nehal Kassem, Ahmed E. Kosba

Speed Estimation, IEEE 75th Vehicular Technology Conference, July,2012.

20