International Journal of Trend in Scientific Research and Development (IJTSRD)

Volume: 3 | Issue: 4 | May-Jun 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 - 6470

A two Stage Fuzzy Logic Adaptive Traffic Signal Control for an

Isolated Intersection Based on Real Data using SUMO Simulator
Taha Mahmood1, Muzamil Eltejani Mohammed Ali2, Akif Durdu2
1Electrical
and Electronic Engineering, Selçuk University, Konya, Turkey
2Robotics Automation Control Laboratory (RAC-LAB), Konya Technical University, Konya, Turkey

How to cite this paper: Taha Mahmood ABSTRACT

| Muzamil Eltejani Mohammed Ali | Akif In this paper, a two-stage fuzzy logic system has been proposed to control an
Durdu "A two Stage Fuzzy Logic isolated intersection adaptively. The aim of this work is to minimize the average
Adaptive Traffic Signal Control for an waiting time for a different traffic flow rates in real time means. In the first stage,
Isolated Intersection Based on Real Data the system consists of two modules named next phase selection module and the
using SUMO Simulator" Published in green phase extension module. In the second stage the system consists of the
International Journal of Trend in decision named module. The study was performed using SUMO traffic simulator.
Scientific Research and Development A comparison is made between a fuzzy logic controller and a conventional fixed-
(ijtsrd), ISSN: 2456- time controller. As a result, fuzzy logic controller has shown better performance.
6470, Volume-3 |
Issue-4, June 2019, Keywords: Adaptive traffic control, Fuzzy logic, isolated intersection, SUMO
pp.656-659, URL: INTRODUCTION
https://www.ijtsrd.c Today, the number of vehicles is increasing rapidly and then traffic congestion are
om/papers/ijtsrd23 a major problem in big cities. In recent years, the management of traffic has
873.pdf IJTSRD23873 become an important part of the modern city. The demand for traffic light control
is increased with the rising users and pedestrians on the road. Traffic jams have
Copyright © 2019 by author(s) and many negative impacts on the environment, social and economic life, time,
International Journal of Trend in production and fuel consumption [1,2]. The high proportion of vehicle-usage, the
Scientific Research and Development insufficiency of space to build a new transport infrastructure and a lack of public
Journal. This is an Open Access article funds further complicate the problem. In this condition, it is necessary to improve
distributed under the quality of service for road-users and to provide smart and economical
the terms of the solutions. In a relatively inexpensive way to reduce congestion problems, traffic
Creative Commons signal system method was used to ensure the best contingent use of the existing
Attribution License (CC BY 4.0) road network. Traffic signaling system strategies are generally divide into three
(http://creativecommons.org/licenses/ categories; traffic controlled by people, the traditional and sensitive traffic control
by/4.0) systems. At the first, the traffic officer uses hand gestures and sometimes verbal
communication to control traffic, and in the light of his [4]. They developed a logical traffic system that takes into
instinct, decides who must be go or stop. The second way account two-way intersections and is able to adjust changes
relies on the concept of fixed time, which cannot be changed in traffic signal intervals based on traffic level, the results
according to variable traffic demand, so that the times are show that the proposed system provides better performance
distributed on different phase of traffic light based on the in terms of total waiting time, total moving time, and vehicle
traffic census [3]. The third way uses sensors such as loop queue [5]. They defines the ATAK system as an traffic
detectors or cameras to calculate and modify signals in real management tool at signalized intersections, the results of
time, the traffic density, speed, flow, queue length and ATAK showing better performance were compared with the
vehicle waiting times are measured immediately on the lanes multi-planned traffic control system [6]. The design of an
and stimulate the lights to respond as required. adaptive traffic light controller using fuzzy logic control
Sugeno Method is used to determine the length of green time
In this paper, we propose a two-stage traffic signal system at at an intersection and the results show that the traffic light
an isolated intersection, using a technique based on fuzzy using fuzzy logic control performs better than using fixed
logic to control traffic signals. In this study, SUMO version time control [7]. They developed algorithms to deal with
0.32.0 was used to simulate the traffic light. The average information uncertainty in dense traffic flows. Adaptive
waiting time is used to evaluate the performance of the fuzzy fuzzy logic management was established for traffic control
logic compared to the fixed time. systems, the proposed method allowed it to more effectively
control the intersection than the fixed time traffic signals [8].
Related Works: With proposed WeightSpreadsheetPad technology, the
Recently, many researchers have focused on finding a new system with fuzzy logic controller and PIC controller for
way to control signals that adjusted the timing and phase of intelligent traffic light improved overall performance
the lights to reduce traffic congestions. They present a novel compared to traditional traffic control [9]. They proposed a
fuzzy model for an isolated signalized intersection, results fuzzy logic controller NS-II to control intersection in real-
show that the performance of the proposed model is better time, the proposed method has a better performance in
that of fixed time control in terms of average waiting time terms of average total delay time and queue length

@ IJTSRD | Unique Paper ID - IJTSRD23873 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 656
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
compared to type 1, type 2 and type 2-DE [10]. A two stage green phase, as the output is the degree of urgency for
traffic light system fuzzy logic-based for real-time traffic selected phase. The module compares the degree of urgency
monitoring has been proposed to dynamically manage both of all input phases and selects the most urgent. Each input
the phase and green time of traffic lights for an isolated and output has its own membership functions. Inputs have
signalized intersection, show better performance compared six membership functions (Zero, Very Short, Short, Medium,
to conventional traffic control in terms of average delay time Long, Very Long) with waiting time and queue length. The
[2]. output (urgency degree) has five membership functions
(Zero, Low, Medium, High, Very High) (see Fig 3).
FUZZY LOGIC CONTROL SYSTEM
Fuzzy logic traffic light control is an alternative to existing
traditional traffic signal control, which can be used for a
wide range of traffic arrangements at a road-intersection.
General structure of fuzzy traffic lights control system is
shown in Fig 1. The lanes is arranged together with the four
phases, the first and the second phase have three, the third
phase has two and the fourth phase has one lanes. For each
lane, a tracker is placed on the road, the detectors are used
to indicate the presence and count the vehicles at the same
time. This system controls the traffic phases and provides
better evaluation of the changing traffic patterns.

Waiting time

Figure.1. General structure of fuzzy traffic control system

Queue length
The two-stage traffic light system based fuzzy logic will not
only decide whether to extend an existing green phase but
also decide which red phase will be set as the green phase
and then determine the extension time of the green phase.
Therefore, the phase sequence in this system is unclear. Fig 2
shows a schematic diagram of the controller.

Urgency degree
Figure.3.The fuzzy sets of Waiting time, Queue length and
Urgency degree

Table I shows a part of the 30 fuzzy rules of this module. For

example, if the waiting time is (Zero) and the queue length
(Longe) means the Urgency (Medium).
Figure.2. Schematic diagram of the controller
Table I. Some rules of the Next Phase Module.
The first stage consists of two modules:
Waiting Queue Urgency
1. Next Phase Module Rules
time Length degree
This module selects a candidate for the next green phase. It
R1 Z VS Z
observes the traffic conditions of all the red phases except
R2 L M H
the green phase and selects the most urgent red phase
among them. The inputs of this module are the (waiting R3 VS L M
times and queue length) of all phases except the current ----- -------- -------- ----------
R30 VL VL VH

@ IJTSRD | Unique Paper ID - IJTSRD23873 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 657
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
2. Extension time module Second Stage Decision module
The extension time module observes the traffic conditions of The decision module decides whether to change the green
the green phase. Produces the stop degree according to the phase or not according to outputs of the first stages. The
observation result. The degree of stop indicates that the inputs of the module are the outputs of the first stage
controller must stop or extend the time of the green phase (Urgency degree and Stop or Extend). The decision is output
according to the remaining and passing vehicles number. and have two membership functions (No,Yes) (see Figure 5).
The inputs of this module are; remaining and passing
vehicles number. Output is the stop or extend time of the
current green light phase. The current green phase;
remaining and passing vehicles numbers have four
membership functions. Membership function of the
remaining vehicles number (Zero, Short, Medium, Long),
membership function of the passing vehicles (Zero, Low,
Medium, High). The output of the module (Stop) has three
membership functions (No, Maybe, Yes) (see Fig 4).
Figure5. The fuzzy sets of decision
Table III shows a part of the 10 fuzzy rules of this module.
For example, if the Urgency degree is (Very High) and Stop
or Extend (Yes) means the Decision (Yes).

Table III. Some Rules of the Decision Module

Rules Urgency degree Stop or Extend Decision
R1 M MB Y
R2 Z Y N
Remaining vehicles number ----- -------- ----------- ----------
R10 VH Y Y

Simulation
In this study SUMO (Simulation of urban mobility) 0.32.0
version is used for simulations. SUMO is a microscopic
simulator operating at each vehicle level. To obtain more
control over the simulation process, interaction can be
established by using SUMO's TraCI protocol during
simulation. SUMO is not just a traffic simulator, is an
application package that allows the user to import a road
Passing vehicles number network and define the corresponding traffic request.
The data of this study is from heavy traffic congestion real
four-way intersection located in urban areas in Kilis
(Turkey) obtained in 24 hours by Mosas company (see Fig.
6). The data shows how many vehicles are going from one
lane to another every 5 minutes. Using the SUMO program,
the cycle time is set to 150 seconds to achieve the lowest
average waiting time in the fixed time traffic control system.
That is, we give phase1, phase2, phase3, phase4, respectively
25, 40, 18, 55 seconds. In the fuzzy traffic control system, a
two-stage three-model traffic light control system was used
to achieve the lowest average waiting time. This system
adjusts the cycle and phase times according to the traffic
Stop or Extend
conditions based on the data from the detectors. The
Figure.4.The fuzzy sets of Remaining, Passing vehicles
performance of the fuzzy logic traffic signal controller
number and Stop or Extend
indicates a 61.363% improvement when compared with the
performance of the traditional traffic controller (see Table
Table II shows a part of the 12 fuzzy rules of this module. For
IV). The isolated intersection was evaluated by simulation.
example, if the remaining vehicles number is (Short) and the
passing vehicles number (High) means Stop or Extend (No).

Table II. Some Rules of Extension Time Module

Remaining Passing
Stop or
Rules vehicles vehicles
Extend
number number
R1 Z M MB
R2 S H N
R3 M L N
----- -------- ------------ ----------
R12 L H N Fig.6. Kilis intersection

@ IJTSRD | Unique Paper ID - IJTSRD23873 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 658
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Table IV. Comparison of results [4] Azimirad, E., Pariz, N., & Sistani, M. B. N. (2010). A
Control Average Vehicle novel fuzzy model and control of single intersection at
Improvement
methods Waiting Time (s) urban traffic network. IEEE Systems Journal, 4(1), 107-
Fixed Time 44 111.
% 61.363
Fuzzy Logic 17 [5] Zaied, A. N. H., & Al Othman, W. (2011). Development o
f a fuzzy logic traffic system for isolated signalized
Conclusion intersections in the State of Kuwait. Expert Systems
We proposed the design of a two-stage traffic signal system with Applications, 38(8), 9434-9441.
for the isolated intersection based on fuzzy logic and
compared its performance with the simulated traditional [6] Gündoğan, F., Karagoz, Z., Kocyigit, N., Karadag, A.,
traffic control method. Traffic simulation is simulated on Ceylan, H., & Murat, Y. Ş. (2014). An Evaluation of
SUMO. To control the intersection, the controllers manage Adaptive Traffic Control System in Istanbul,
traffic signals based on the traffic information they collect. Turkey. Journal of Traffic and Logistics Engineering
The controller receives information from its detectors. Using Vol, 2(3).
this information, the fuzzy rule basic system gave an optimal [7] Prasetiyo, E. E. , Wahyunggoro, O., & Sulistyo, S. (2015).
phase and a signal sequence at the intersection points. The Design and Simulation of Adaptive Traffic Light
results obtained from the simulation showed that the Controller Using Fuzzy Logic Control Sugeno
average vehicle waiting time decreased from 44 seconds to Method. International Journal of Scientific and
17 seconds, suggesting that the recommended method was Research Publications, 5(4), 1-6.
better than the fixed time control method.
[8] Yusupbekov, N. R., Marakhimov, A. R., Igamberdiev, H.
Acknowledgment Z., & Umarov, S. X. (2016). An adaptive fuzzy-logic
Authors are thankful to RAC-LAB (www.rac-lab.com) and traffic control system in conditions of saturated
MOSAŞ GROUP (Akıllı Ulaşım ve Tartım Sistemleri Şti.) for transport stream. The Scientific World Journal, 2016.
providing the database for this study. [9] Bhavsar, S., Sharma, V., & Student, B. E. (2017).
Intelligent Traffic Controller using Fuzzy
References: Logic. International Journal of Engineering
[1] Mannion, P., Duggan, J., & Howley, E. (2015). Parallel Science, 14961.
reinforcement learning for traffic signal
control. Procedia Computer Science, 52, 956-961. [10] Wen, C., Hui, Z., Tao, L., & Yuling, L. (2015). Intelligent
traffic signal controller based on type-2 fuzzy logic and
[2] Alam, J., & Pandey, M. K. (2015). Design and analysis of NSGAII. Journal of Intelligent & Fuzzy Systems, 29(6),
a two stage traffic light system using fuzzy logic. J 2611-2618.
Inform Tech Softw Eng, 5(162), 2.
[11] Lee, J. H., & Lee-Kwang, H. (1999). Distributed and
[3] Garg, H., & Kaushal, E. G. (2017). Traffic Lights Control cooperative fuzzy controllers for traffic Intersections
System for Indian Cities Using WSN and Fuzzy group. IEEE Transactions on Systems, Man, and
Control. Traffic, 4(07). Cybernetics, Part C (Applications and Reviews), 29(2),
263-271.

@ IJTSRD | Unique Paper ID - IJTSRD23873 | Volume – 3 | Issue – 4 | May-Jun 2019 Page: 659