Lb/noH/£Z/ob 
A REAL TIME TRAFFIC SIGNAL CONTROL 
SYSTEM 
A dissertation submitted to the 
Department of Electrical Engineering, University of Moratuwa 
in partial fulfillment of the requirements for the 
degree of Master of Science 
LIBRARY 
UNIVERSITY OF MORATUWA, SRI LANKA 
M O R A T U W A . 
by 
JAYAKODY ARACHCHIGE NIMANTHI NISHANI 
KUMUDU JAYAKODY 
Supervised by: Dr. Lanka Udawatta and Dr. Sisil Kumarawadu, 
Department of Electrical Engineering 
University of Moratuwa, Sri Lanka 
University of Moratuwa 
91207 
g\zo 
January 2008 
9 1 2 0 7 i 
DECLARATION 
The work submitted in this dissertation is the result of my own 
investigation, except where otherwise stated. 
It has not already been accepted for any degree, and is also not being 
concurrently submitted for any other degree. 
J.A.N.N.K Jayakody 
25.01.2008 
We endorse the declaration by the candidate. 
Dr. Lanka Udawatta 
skJ "A 
Dr. Sisil Kumarawadu 
11 
CONTENTS 
Declaration 
Abstract 
Dedication 
Acknowledgement 
List of Figures 
List of Tables 
Chapter 1 Introduction 
1.1 Background 
1.2 Motivation 
1.3 Literature Review 
1.4 Contributions of the Research 
1.5 Organization of the Report 
Chapter 2 Available Adaptive Traffic Control Systems 
2.1 Introduction 
2.2 Major Methodologies 
2.3 Split Cycle Offset Optimization Technique 
2.4 Sydney Coordinated Adaptive Traffic System 
2.5 Real time Hierarchical Optimized Distributed & Effective 
System 
2.6 OP AC - Optimized Poicies for Adaptive Control 
Chapter 3 Decentralized Intelligent control Model i s 
3.1 Introduction ^ 
3.2 Detection in Adaptive Control 1 g 
3.2.1 Function of Detection j g 
3.2.2 Detection Methods j9 
iii 
3.2.3 Inductive Loop Detectors 20 
3.3 Influence Detection 21 
3.3.1 Structure of the Intelligent Control 21 
3.3.2 Formulating the Influence Function for Traffic Signal Control 22 
Chapter 4 Fuzzy Control Strategy Development 25 
4.1 Introduction 25 
4.2 Fuzzy Inference System 25 
4.2.1 Fuzzy Basics 25 
4.2.2 Fuzzification 26 
4.2.3 Rule Base 3 0 
4.2.4 Determination of the Single Output 33 
4.3 Control Strategy 34 
4.3.1 System Architecture 34 
4.3.2 Decentralized Approach 35 
4.3.3 Prediction Method 36 
4.3.4 Control Mechanism 37 
Chapter 5 Simulation and Results 41 
5.1 Introduction 4] 
5.2 Simulation Model Development 41 
5.2.1 Simulink Basics 41 
5.2.2 Traffic Simulation Model of an Intersection 42 
5.2.3 Simulation Results 44 
5.3 Comparison with existing Fixed time Traffic Control System 46 
5.3.1 Delay Calculation 46 
5.3.2 Delay Calculation for the Approach A 47 
IV 
5.3.3 Delay Calculation for the Approach B 
5.3.4 Delay Calculation for the Approach C 
5.3.5 Delay Calculation for the Approach D 
49 
51 
53 
Chapter 6 
References 
Conclusions and Future work 
6.1 Conclusions 
6.2 Future Work 
56 
56 
56 
58 
v 
Abstract 
Traffic congestion problem in Colombo city is getting worse since traditional traffic 
control system could not fulfill the need. Since the existing system is a fixed time 
fixed cycle control system, it cannot fit with dynamic traffic environment. 
In this research, a decentralized control strategy to control a traffic network grid is 
presented. Single controller is to control traffic signals of all approaches at one 
intersection and each approach green time is given by its separate Fuzzy Inference 
System. Vehicle arrival data are to be collected by lane detectors. Inductive Loop 
Detectors are proposed for this purpose. Herein, a methodology is developed to 
decide green time of each approach based on the arrival data by the Fuzzy Inference 
System and the Cycle time. Influence to the particular intersection is identified and is 
factorized as an input to the Fuzzy Inference System. Later, the green time is decided 
by the FIS. Results for this mechanism are shown for one intersection on a simulated 
environment modeled by Matlab. 
Calculations have been done based on the real data obtained for fifteen occasions. 
Results for three sets of data from both existing fixed time system and the intelligent 
model have been compared based on the calculations done for the total vehicle delay 
time, expected at the passing the particular intersection. It shows 51.6% of minimized 
total vehicle seconds delay by the intelligent traffic control model over the fixed time 
control system. 
vi 
To my parents 
vii 
Acknowledgement 
This is to pay my warm gratitude to the Department of Electrical Engineering, 
University of Moratuwa, SriLanka for giving me the opportunity to read for the M.Sc 
in Industrial Automation and to do this research during the period of October 2005 to 
January 2008. 
I would like to extend my warm thanks with respect to my research supervisors Dr. 
Lanka Udawatta and Dr. Sisil Kumarawadu, who have given endless support and 
guidance while I was doing the research and following the program. Their breadth of 
knowledge and the advices given me at the presentations made me highly impressed 
to do the research in depth and to develop myself by enriching academic, research 
oriented and professional maturity. Their countless advices and the guidance were 
invaluable. 
I would also like to thank Professor Ranjith Perera, the Head of the Department. He 
has given me support and encouragement, and his advice and feedback about my 
research at presentations have greatly enhanced and strengthened the study. I thank 
him for all the time and energy he has paid for my work. 
I also wish to extend thanks to my friends and colleagues who were invaluable in 
completing this study, and to the management of Electro-Serv (pvt) ltd who has given 
me the support by releasing me from the duty to follow this program. 
I am indebted to my parents and to my brother Suranga Jayakody for the guidance 
they have given to me. I recall their constant support, encouragement, love and 
guidance given me every time. 
Finally, I would like to share my research experience with all of you. 
Nimanthi Jayakody 
University of Moratuwa, Sri Lanka 
January 2008 
viii 
List of Figures 
Figure Page 
3.1 Block diagram of the Inductive Loop Detector 20 
3.2 Part of the traffic network 22 
3.3 Intersection A 23 
4.1 Trapezoidal membership function for the input Influence 27 
4.2 Triangular membership function for the output Split 28 
4.3 Variation of Split Vs. Influence 29 
4.4 Graphical representation of rule evaluation 31 
4.5 Block diagram of the control system 35 
4.6 Master Slave control model 36 
4.7 Flow chart for proposed adaptive control mechanism 39 
5.1 Traffic simulation model developed by Matlab 43 
5.2 Arrivals at an intersection over the cycle time 47 
IV 
List of Tables 
Table Page 
4.1 Membership Values of the Input Influence 27 
4.2 Membership Values of the Output Split 28 
5.1 Detector Counts at Approach Counters 44 
5.2 Influences calculated for the detector data 44 
5.3 Splits determined by the Fuzzy Inference Systems 45 
5.4 Delay Calculation for the Intelligent Model at Split A 48 
5.5 Delay Calculation for the Fixed time Control Model at 
Split A 49 
5.6 Delay Calculation for the Intelligent Model at Split B 50 
5.7 Delay Calculation for the Fixed time Control Model at 
Split B 50 
5.8 Delay Calculation for the Intelligent Model at Split C 52 
5.9 Delay Calculation for the Fixed time Control Model at 
Split C 52 
5.10 Delay Calculation for the Intelligent Model at Split D 54 
5.11 Delay Calculation for the Fixed time Control Model at 
Split D 54 
5.12 Improvement obtained by the Intelligent Control Model 55 
x