Design and-analysis-of-a-two-stage-traffic-light-system-using-fuzzy-logic-2165-7866-1000162

Volume 5 • Issue 3 • 1000162J Inform Tech Softw Eng
ISSN: 2165-7866 JITSE, an open access journal
J
ournalofInformati
on
Technology &
S
oftwareEngineerin
g
ISSN: 2165-7866
Alam and Pandey, J Inform Tech Softw Eng 2015, 5:3
http://dx.doi.org/10.4172/2165-7866.1000162
Research Article Open Access
Information Technology & Software Engineering
Design and Analysis of a Two Stage Traffic Light System Using Fuzzy Logic
Javed Alam1
* and Pandey MK2
1
Research Scholar, Dept. of CS,Mewar University, Gangrar, Chittorgarh, Rajasthan, India
2
Director, Computer Science and Applications, AIMCA, Haldwani, Uttrakhand, India
*Corresponding author: Javed Alam, Research Scholar, Dept of CS, Mewar
University, Gangrar, Chittorgarh, Rajasthan, India, Tel: 09557592020; E-mail:
javedalam4u@gmail.com
Received September 18, 2015; Accepted November 10, 2015; Published
November 30, 2015
Citation: Alam J, Pandey MK (2015) Design and Analysis of a Two Stage Traffic
Light System Using Fuzzy Logic. J Inform Tech Softw Eng 5: 162. doi:10.4172/2165-
7866.1000162
Copyright: © 2015 Alam J, et al. This is an open-access article distributed under
the terms of the Creative Commons Attribution License, which permits unrestricted
use, distribution, and reproduction in any medium, provided the original author and
source are credited.
Keywords: Traffic congestion; Traffic urgency; TSTLS; TUDM;
DTDM
Introduction
Traffic congestion of streets and roads constitutes a critical
problem which is aggravated by the rise in the number of vehicles and
by greater urbanization. The slow pace in the development of new
highways and roads and public opposition to the widening of existing
streets in some locations has forced the city managers to optimally use
the existing infrastructures in order to effectively manage the flow of
traffic. Moreover the loss of valuable time during traffic congestion
can directly affect the production, productivity, performance and the
utilization of fuel. The control of traffic light signal is one of the subjects
of intelligent or advance systems being investigated by researchers
because this kind of control has a direct impact on the effectiveness of
urban transportation systems [1].
Traffic signals operate in pre-timed, actuated or adaptive mode.
Pre-timed control consists of a series of intervals that are fixed in
duration. They repeat a preset constant cycle. In contrast to pre-timed
signals actuated signals have the capability to respond to the presence
of vehicles or pedestrians at the intersection. Vehicle actuated signals
require actuation by a vehicle on one or more approaches in order for
certain phases or traffic movements to be serviced. They are equipped
with detectors and the necessary control logic to respond to the
demands placed on them. Vehicle-actuated control uses information
on current demands and operations obtained from detectors within the
intersection to alter one or more aspects of the signal timing on a cycle-
by-cycle basis. Timing of the signals is controlled by traffic demand.
Adaptive or area traffic control system sometime called as ATCS.
It is the latest generation of traffic control system. ATCS continuously
compute optimal signal timings based on this detected volume and
simultaneously implement them. ATC system to effectively respond
to the rapid changes in dynamic traffic conditions which are designed
to have intelligent real-time dynamic traffic control systems. These
systems use data from vehicle detectors and optimize traffic signal
time in real time that’s why ATCS called as adaptive traffic signal.
The timing plan of ATCS controller changes automatically for these
purpose digital computers are used.
Over the course of performing a literature review, adaptive traffic
signal systems have been operating successfully in many countries since
the early 1970. Adaptive traffic signal control systems are normally
complicated and include prediction and estimation modules. More
than twenty Adaptive traffic signal controls are available on the market.
They are significant due to their relative acceptance in the field as well
as the relative extent of their real world implementation. The most
widely deployed control systems are discussed here.In the early 1980,
Nathan Gartner at the University of Massachusetts at Lowell proposed
a traffic control system called asOptimized Policies for Adaptive
Control for the Federal Highway Administration. Optimized Policies
for Adaptive Control sometime called as OPAC. The Split Cycle Offset
Optimization Technique (SCOOT) was also developed in the early
1980 by the Transport Research Laboratory in the United Kingdom.
The Sydney Coordinated Adaptive Traffic System (SCATS) is slightly
newer, having been created in the early 1990 by the Roads and Traffic
Authority of New South Wales, Australia. A major difference between
SCATS and SCOOT is that SCATS does not have a traffic model or
a traffic signal control plan optimizer. Timing of signals is controlled
by digital computer based control logic. It has ability to modify signal
timings on a cycle-by-cycle basis using traffic flow information collected
at the intersection. It is not model based but has a library of plans that
it selects from and therefore banks extensively on available traffic data.
The Real-time Hierarchical Optimized Distributed Effective System
(RHODES) is the newest of these four systems, having been produced
in the mid-1990 at the University of Arizona at Tucson.
A decentralized adaptive traffic signal control method known as
ALLONS-D (Adaptive Limited Look-ahead Optimization of Network
Signals – Decentralized) presented by Porche [2] in his dissertation
Abstract
Traffic congestion is a growing problem in many countries especially in large urban areas. The reason behind this
is the increase in number of vehicles all over the world. Due to this vehicles do not run efficiently which increasestravel
time, noise pollution, carbon dioxide emissions and fuel usages.In large urban cities, traffic signal controller plays an
important role to improve the efficiency of vehicles, traffic congestion and hence reduces travel time, noise pollution,
carbon dioxide emission and fuel used. In this paper, a two stage traffic light system for real-time traffic monitoring
has been proposed to dynamically manage both the phase and green time of traffic lights for an isolated signalized
intersection with the objective of minimizing the average vehicle delay in different traffic flow rate. There are two
different modules namely traffic urgency decision module (TUDM) and extension time decision module (ETDM). In the
first stage TUDM, calculates urgency for all red phases. On the bases of urgency degree, proposed system selects
the red light phase with large traffic urgency as the next phase to switch. In second stage ETDM, calculatesgreen light
time i.e., extension time of the phase which has higher urgencyaccording to the number of vehicles. Software has been
developed in MATLAB to simulate the situation of an isolated signalized intersection based on fuzzy logic. Simulation
results verify the performance of our proposed two stage traffic light system using fuzzy logic.

Citation: Alam J, Pandey MK (2015) Design and Analysis of a Two Stage Traffic Light System Using Fuzzy Logic. J Inform Tech Softw Eng 5: 162.
doi:10.4172/2165-7866.1000162
Page 2 of 9
intersections by Chui [7]. In this research only two way streets are
evaluated without considering any left or right turn. A two stage
traffic light controller proposed by Trabia, et al. [8]. In the first stage,
observed approach traffic flows are used to estimate relative traffic
intensities. These traffic intensities are then used in the second stage
to decide current green signal should be extended or terminated for
through movements without considering any left or right turn. The
isolatedintersection model proposed by Soh AC, et al. [9], used consists
of two lanes in eachphase. There are two inputs i.e., vehicles queue
lengthand waiting times for each phase. The maximum values of these
inputs are selected for controllerto optimized control of traffic flows.
A fuzzy logic traffic system proposed by Zaied ANH, et al. [10], that
considers the two two-way intersections and is able to adjust changes in
time intervals of a traffic signal based on traffic situation level.
Indrabayul, et al. [11], proposed an adaptive timely traffic light is
as solution for congestion in typical area in Indonesia. Makassar City,
particularly in the most complex junction is observed for months using
static cameras. The condition is mapped into fuzzy logic to have a better
time transition of traffic light as opposed to the current conventional
traffic light system. Fuzzy logic based traffic light shows significant
number of potential reduced in congestion.
Shahraki, et al. [1] a new fuzzy logic based algorithm is proposed
in this research not only can reduce the waiting time and the number
of vehicles behind a traffic light at an intersection but also can consider
the traffic situations at adjacent intersections as well. The fuzzy control
system comprises three stages. These three stages include the next
green phase, green phase extender, and the decision stage. The inputs
are applied through the green phase selector. The next green phase
stage selects the most urgent phase from the phases waiting to become
green. If necessary, the green phase extender increases the duration
of the green light. In the decision making stage, by deciding either to
increase the green light duration or to change to another phase, the
most urgent stage is selected from the two stages of next green phase
and green phase extender.
Collotta, et al. [12], a novel approach to dynamically manage the
traffic lights cycles and phases in an isolated intersection. The proposed
solution is a traffic lights dynamic control system that combines
Wireless Sensor Network for real time traffic monitoring with multiple
fuzzy logic controllers, one for each phase that work in parallel. Each
fuzzycontrolleraddressesvehiclesturningmovementsanddynamically
manages both the phase and the green time of traffic lights.
Wu, et al. [13], a dynamic control technique for traffic lights is
presented, which is based on the queue detection in the left and straight
lanes assuming that the vehicles in the right lane are not in conflict
with the others. For queue detection purposes, two induction coils are
used, the first one to detect oncoming vehicles, the second to measure
the vehicles that leave the intersection. The work considers 12 phases,
scheduled according to the priority of each phase that depends on the
queue lengths of the specific phase lanes. The additional green time
is then calculated using a fuzzy logic controller that processes two
parameters, i.e., the queue length of the lane with the green light and
that of the lanes with the red light. The phase duration depends on the
traffic flow that the phase should serve and in this respect the main
limitation of the works presented by Shahraki, et al. [1] and Wu, et
al. [13] is that the green time extension is calculated by a single fuzzy
controller for all the phases whereas for better performance, fault-
tolerance and flexibility, as explained before a controller for each phase
would be needed to determine the green time duration of the specific
phase. The same problems in Shahraki, et al. [1] and Wu, et al. [13] also
based on a depth-first branch and bound algorithm.More recently, Yu
and Recker [3] developed a stochastic adaptive traffic signal control
model. The authors formulated traffic signal control as a Markov
Decision Process (MDP) and solved it by dynamic programming.
Although dynamic programming algorithm can be used to solve this
MDP problem and is guaranteed to find the optimal policy [4], it needs
a well-defined state-transition probability function. In practice this
state transition probability function is difficultto define. In the case
of intersection traffic control state transition probability function are
affected by the arrival of actual traffic and often time is different. Thus
it is even more difficult to give a precise estimate. An intersection traffic
signal control application in addition to the number of states is usually
very large. The dynamic programming algorithm to calculate the time
could make a serious problem.
However this type of methods still has the problems that under
certain circumstances, the excessive computation requirement makes
some systems based on dynamic programming and Markov decision
process require accurate traffic arrival information for the next one
or two minutes to determine the best control plans. This information
is very difficult to obtain. These systems take ordinary variable in
computation. Therefore it is necessary to improve the traffic controller
for effective traffic management and better traffic flow, we use linguistic
variable in place of ordinary variable.
Fuzzy logic enables the implementation of rules very similarly to
what goes on in the human thinking process. In other words, fuzzy
controllers have the ability to take decision even with incomplete
information. More and more sophisticated controllers are being
developed for traffic control. These algorithms are continually
improving the safety and efficiency by reducing the waiting delay of
vehicles on signals. This increases the tempo of travel and thus makes
signals more effective and traffic flow smooth. The key motivation
towards fuzzy logic in traffic signal control is the existence of
uncertainties in signal control. Decisions are taken based on imprecise
information and the effect of evaluation is not well known.
In this paper we discuss the design and analysis of atwo stage traffic
lightsystem for isolated intersectionusing fuzzy logic basedtechnology
which has the capability of mimicking human intelligence for
controlling traffic light. We used fuzzy logic tools available with
MATLAB and developed software to simulate the situation of traffic
at an isolated junction. The simulated model used for the analysis of
efficiency of traffic light controller. The average vehicle delays will be
used to evaluate the performance of a two stage traffic light system
using fuzzy logic. The software can also be used as an exercise for
undergraduate and graduate students to understand the concept of
fuzzy logic and its application to a real life environment. The rules and
membership functions of the fuzzy logic controller can be selected and
changed their outputs can be compared in terms of several different
representations.
Related Work
In this section, we discuss different research work in the field of
traffic light system using fuzzy logic. The first attempt made to design
fuzzy traffic controller was in 70s by Pappis and Mamdani[5]. Kelsey
and Bisset also designed a simulator for signal controlling of an isolated
intersection with one lane. Same work was also done by Niittymaki
and Pursula [6]. They observed that Fuzzy Controller reduces the
vehicle delay when traffic volume was heavy. Niittymaki and Kikuchi
developed Fuzzy based algorithm for pedestrians, crossing the road.
Initially fuzzy logic was used to control traffic in multiple

doi:10.4172/2165-7866.1000162
Page 3 of 9
characterize the approach presented in Zaied and Othman [10] whose
aim is to dynamically regulate through a fuzzy controller, the green
time duration of an 8-phase traffic lights in an isolated intersection. In
this case, 8 inputs (i.e., the queues) and 16 outputs (i.e.,, the calculated
green time for each allowed direction) have to be handled. This means
that the fuzzy controller is characterized by multiple input variables.
Traffic Engineering and Traffic Congestion Estimation
Traffic signals are designed to ensure safe and orderly flow of
traffic protect pedestrians and vehicles at busy intersections and reduce
the severity and frequency of accidents between vehicles entering
intersections. In other words traffic signals are one of the most effective
and flexible active control of traffic and is widely used in several cities
worldwide. The conflicts arising from movements of traffic in different
directions are addressed by time sharing principle. The advantages
of traffic signal include an orderly movement of traffic an increased
capacity of the intersection and require only simple geometric design.
However the disadvantages of the signalized intersection are large
stopped delays and complexity in the design and implementation.
Although the overall delay may be lesser than a rotary for a high volume
a user may experience relatively high stopped delay [14].
Cycle length
According to Mathew [14], a signal cycle is one complete rotation
through all of the indications or phases that are provided. Cycle length
is the time that it takes a signal to complete one full cycle of indications
or phases.
Interval
It indicates the change from one stage to another stage. Intervalis
oftwo types i.e., change interval and clearance interval
Change interval: The change interval some time called as yellow
time. It is provided after green time for movement. The purpose is
to inform a driver approaching the intersection during the end of a
green time about the coming of a red signal. They normally have a
value of 3 to 6 seconds. Institute of transportation engineers (ITE) has
recommended following methodology for computing the appropriate
length of change interval
n
v
Y t
2 (g a)
= +
+
Where t is the reaction time (about 1.0 sec), v is the velocity of the
approaching vehicles, g is the acceleration due to gravity (9.8 m/sec2
),
n is the grade of the approach in decimals and a is the deceleration of
Change interval. It can also be approximately computed as y = SSD/v,
where SSD is the stopping sight distance and v is the speed of the
vehicle.
Clearance interval: The clearance interval some time called as all-
red time. It will facilitate a vehicle just crossed the stop line at the turn
of red to clear the intersection without being collided by a vehicle from
the next phase. Institute of transportation engineers (ITE) recommends
the following policy for the design of all read time, given as
AR
w L
If no pedestrians
v
w L P
R max , If pedestrians crossing
v v
P L
If protected
v
+


+ 
=   
 
 +


Where w is the width of the intersection from stop line to the
farthest conflicting traffic, L is the length of the vehicle, v is the speed of
the vehicle, and P is the width of the intersection from STOP line to the
farthest conflicting pedestrian cross-walk.
Phase
A phase follows the time of green interval plus the change and
clearance intervals. Therefore during green interval, non-conflicting
movements are assigned into each phase. The objective of phase design
is to separate the conflicting movements in an intersection into various
phases so that movements in a phase should have no conflicts. If all the
movements are to be separated with no conflicts then a large number
of phases are required. In such a situation the objective is to design
phases with minimum conflicts or with less severe conflicts as discuss
by Mathew [14].
Lost time
Lost time is the interval in which there is no effectivemovement at
the intersection.
Delay
There are a number of measures used for capacity analysis and
simulations for signalized intersection. The most common measures the
average delays per vehicle, average queue length and number of stops.
The delay is a measure that is most directly related to the experience
of the driver. It is the measure of extra time consumption through the
intersection. Any time is a useful measure of the queue length and the
intersection of an adjacent upstream intersection will begin to hinder
the discharge is important in determining. The number of stops made
in particular in the air quality model, is an important input parameter.
It is directly perceived by the driver. Apart from these three, the effects
of delay signalized intersections are the most frequently used measures.
Due to the delay estimation is complicated by the arrival of vehicles at
random, over saturated flow conditions, lost time due to stopping of
vehicles etc. although delay can be quantified in several different ways.
Thus, as the most frequently used forms of delay are following-
• Stopped time delay
• Approach delay
• Travel time delay
• Time-in-queue delay
• Control delay
Traffic congestion
Traffic congestion means there are more people trying to use a given
transportation facility during a specific period of time than the facility
can handle with what are considered to be acceptable levels of delay
or inconvenience. Delays at particular locations in a transportation
network are certainly provoking to those using the system but these
delays are part of a much larger picture of how a transportation
system allows people and goods to move around a metropolitan area
[15]. Congestion management are designed to improve the operating
efficiency of the existing transportation system by increasing the use
of alternative transportation modesaltering trip patterns through the
application of measures and improving traffic flow through measures
such as route guidance systems, traffic signal improvements and
incident management.
In TSTLS, four-phase signal are designed as shown in Figure 1. In
a cycle, each approach goes through two time intervals, green interval
and red interval. The green interval has three timing parameters
namely lost time TLOST
, minimum time TMIN
, and maximum time TMAX
.

doi:10.4172/2165-7866.1000162
Page 4 of 9
Lost time TLOST
represents the green time not utilized by the vehicles
waiting in queue. It consists of startup lost time at the beginning of a
green phase and clearance lost time at the end of the phase [16]. In Our
proposedTSTLS, we assume that the entire lost time per phase occurs
only at the beginning of the green phase. When a phase is initiated
minimum green time to be provide and it represented by Minimum
time TMIN
. This ensures that the green signal stays long enough for
safe passage of a single vehicle to clear the intersection. The maximum
green time that can be provided to any phase and represented by
Maximum time TMAX
. In fixed-cycle traffic light system each cycle can
be approximated to a periodic task with period T characterized by a
fixed green time Tg
, yellow time Ty
and red time Tr
described as
T = Tg
+ Ty
+ Tr
Whenever, a phase has the least Urgency than all other phases in
a cycle. This phase can be skip. Therefore green time not provided to
this phase because traffic system is based on urgency. If this phase will
be skip into this cycle and next one or more consecutive cycle then
congestion may occur. In our proposed system this type of congestion
can be estimated and hence minimum green time TMIN
will be provided
to this phase in each cycle.
Description of the Proposed Two Stagetraffic
Lightsystem (TSTLS)
Investigation on the driving behaviors indicates that it is dangerous
to change dynamic phase compositionbecause this may disturb the
drivers’ mental status and may get nervous. In our proposed two stage
traffic light system using fuzzy logic model it is assumed that phase
composition is predetermined and the phase sequence as well as signal
timing are changeable.
From the review of literature related with traffic light system it has
reported that fuzzy logic controllers perform better than pre-timed,
actuated and adaptive controllers. However the phase changes in
sequential order without considering the urgency of the red phases. In
this paper, a two stage traffic light system using fuzzy logic will not
only decide whether to extend or terminate a current green phase but
also decide which red phase will be set as green phase then determine
the extension time of green phase. Therefore in this traffic light system
the phase sequence is uncertain.The average vehicle delays will be used
to evaluate the performance of the two stage traffic light system using
fuzzy logic [17].
The general structure of an isolatedintersection is illustrated as in
Figure 1. Each lane equipped with two electromagnetic. The first sensor
is located behind each traffic light and second sensor is located behind
the first sensor at distance S. The first sensor counts the number of
cars passing the traffic light and the second sensor counts the number
of cars coming to the intersection. The number of cars waiting at the
traffic light is determined by the difference of the reading between the
two sensors. Each traffic light in front of it is a proximity sensor and
can only sense the presence of a car waiting at the junction, which is
contrary to the traditional control systems.
The isolated intersection considered as in Figure 1 is characterized
by four phases as in Figure 2 with eight lanes. Each phase has two lanes.
As we discuss earlier the objective of phase design is to separate the
conflicting movements in an intersection into various phases so that
movements in a phase should have no conflicts. For example, when
Phase 1 is enabled, only the cars of Lane EL1 of the Road directionE
and Lane WL1 of the Road direction W can go straight or turn left,
while all the other lanes will have the red light to stop [18]. Table 1
summarizes the notation adopted in two stage traffic light system.As
shown in Figure 1 the length of each phase is obtained as the maximum
of the queue lengths intended as the number of vehicles during the
green/red light in their respective lanes as shown
Figure 1: General Structure of intersection.
Phase-1
Phase-2
Phase-3
Phase-4
Figure 2: Phases of Traffic light.
Symbol Lane Movement Allowed
EL 1 East Lane- 1 Straight and left
EL 2 East Lane- 2 Right
WL 1 West Lane- 1 Straight and left
WL 2 West Lane- 2 Right
NL 1 North Lane- 1 Straight and left
NL 2 North Lane- 2 Right
SL 1 South Lane- 1 Straight and left
SL 2 South Lane- 2 Right
Table 1: Notation.

doi:10.4172/2165-7866.1000162
Page 5 of 9
Phase1= maximum (WL1, EL1)
Phase2 = maximum (WL2, EL2)
Phase3 = maximum (NL1, SL1)
Phase4 = maximum (NL2, SL2)
Vehicle arrival distribution
There are four phases or approaches in this isolated intersection
model with twelve total movements and a server traffic light. Each phase
consists of three movements which are through movement, right and
left turn movements. This model is based on an M/M/1 queuing theory
using three basic concepts of queuing theory i.e., customer, queue,
and server. The first and the second M in M/M/1 stand for memory
less distribution of inter-arrival times and service times respectively.
The “1” indicates that the isolated intersection has single server, which
means one traffic signal, to service single signal phase at one time [19].
In this model First-In-First-Out (FIFO) principle is applied to the
vehicles queue. The vehicles are known as customers while services
time is the time for the vehicles to depart and to cross the intersection.
Traffic at the intersection arrival and service time are independent
random variables with Poisson distribution.Therefore, the arrival of
vehicles at the intersection with arrival rate λ and the average of the
inter-arrival times between vehicles are 1/ λ. The arrival of the vehicles
in a Poisson process and the time period in order to reach the number
of vehicles is as follows Poisson distribution.
( ) .
( ) x 0, 1, 2, 3,........
!
X t
t e
P x
x
λ
λ −
= =
Where P(x) is the probability of vehicles x during counting interval
t and λ is the average arrival rate per unit time, t is the duration of each
counting interval.The vehicle arrival rate is 0~1 per second.
Put n= λt, where n is the average number of vehicles during
counting interval. The equation can be written as follows
!
.
)(
x
en
xP
nX −
=
Computation of the average vehicle delay (AVD)
In this research paper the average delay of vehicles is the
performance evaluation for traffic signal control of intersection. If the
value of average vehicle delay is small then the traffic signal control
effect is better. The amount of queuing vehicles in red light phase at
time 𝑡i.e.,(𝑡) can be calculated using the following ways-
1
1
r
1 1
Q ( )
k t
gi it
i t
t Q q
= =
 
= +  
 
∑ ∑
where 𝑖 is one of the red light phase branches, 𝑄 𝑔𝑖
is the amount of
queuing vehicles in the red light phase branch 𝑖 when the current green
signal ends,t1
is the time interval during the red light, 𝑞𝑖t1
is the amount
of arriving vehicles in the red light phase branch 𝑖 during time interval
t1,
𝑘is the number of red phase branch. In our proposed system 𝑘 = 6.
The total vehicle delay in red light phase i.e., Dr
can be calculated as-
1
1
r
1 1 1
D
rt k t
gi it
t i t
Q q
= = =
  
= +   
   
∑ ∑ ∑
Wheretr
is the red light time of red light phase.
The amount of queuing vehicles at time 𝑡seconds during green light
i.e.,Qg
(t) can be calculated as follows:
∑ ∑= =








−+=
k
i
f
t
t
itrii tSqQzt
1 1
g .)(Q
1
1
Where 𝑄𝑟𝑖
is the amount of arriving vehicles in the green light
phasebranch 𝑖during last red light phase, 𝑘is the vehicles flowin green
light phase, Sf
is saturation flow, 𝑞𝑖t1
is the amount of arrivingvehicles
in the green light phase branch 𝑖during time t1
. In our research paper
there are four phases in anintersection.The value 𝑧𝑖
can show whether
there are queuing vehicles ornot in current green light phase branch 𝑖.
If 0.
11
1
>−+ ∑=
tSqQ f
t
t
itri , then 𝑧𝑖
= 1, otherwise, 𝑧𝑖
= 0.The total vehicle
delay of green light phase i.e.,, Dg
can becalculated as follows:
∑ ∑ ∑= = = 















−+=
Et
t
k
i
s
t
t
itrii tqqQz
1 1 1
g .D
1
1
Where tE
is the green light time i.e., extension time provided by
extension time decision module during green light phase. The total
vehicle delay in the 𝑥cycle i.e.,Dx
can be calculated as follows:
grx DDD +=
The average vehicle delay (AVD) can be calculated as follows:
T
cyclex
A
∑∈
=
xD
AVD
Where 𝐴T
is the total amount of arriving vehicles from all directions
in an intersection during all cycles.
Traffic urgency decision module
Traffic urgency decision module (TUDM) calculates urgency for
all red phases as shown in Figure 3. On the bases of urgency degree,
proposed system selects the next red light phase to switch. There are two
input variables namely Queue, Timer
and one output variable namely
Urgency. The input variable Queue count the number of vehicles of
current red light phase and variable Timer
count the duration of red
light since the last end of the green light. The output variable Urgency
is traffic urgency of red light phases. Therefore for instance in the case
of Phase 1, TUDM will determine the urgency (U1) by processing the
queue length and duration of red light since the last end of the green
light of phase 1. Similarly, for phase 2, TUDM will determine the
urgency (U2) by processing the queue length and duration of red light
since the last end of the green light of phase 2 and so on.
Figure 3: Proposed Traffic urgency decision module.

doi:10.4172/2165-7866.1000162
Page 6 of 9
Extension time decision module
Extension time decision module (ETDM),calculate green light
time i.e., extension time of the phase which has higher urgency can be
calculated according to the number of vehicles as shown in Figure 4.
There are two input variables namely Queue-Lane-1, Queue-Lane-2 and
one output variable Extension-Time. The input variable Queue-Lane-
1count the number of vehicles of lane 1 and input variable Queue-Lane-
2count number of vehicles of other side i.e., lane 2 of red light phase
which has big traffic urgency. The output variable Extension-Time is
the extension time of current green light phase. Therefore for instance
in the case of Phase 1, ETDM will determine the green light duration
i.e., extension time (E1) by processing the queue length of the Lane
EL1 of the Road direction E and Lane WL1 of the Road direction W.
Similarly, for phase 2, ETDM will determine the extension time (E2)
by processing the queue length of the Lane EL2 of the Road direction E
and Lane WL2 of the Road direction W and so on.
Fuzzy Parameters and Their Membership Functions
Design
For a two stage traffic lightsystem, there are five membership
functions such as Zero, Small, Medium, Large and Very-Large for each
of the input as well asoutput fuzzy variable of the system. There are two
input variables Queueand Timer
and one output variableUrgencyfor
traffic urgency decision module as shown in Figure 5-7 respectively.
For extension time decision module, we designed two input variables
namely Queue-Lane-1, Queue-Lane-2 and one output variable
Extension-Time as shown in Figure 8-10 respectively. Each input and
output fuzzy variable is design using triangular membership function.
Fuzzy Rules and Defuzzification
The inference mechanism in the fuzzy logic controller resembles
that of the human reasoning process. Fuzzy logic technology is
associated with artificial intelligence. For example, a traffic policeman
managing a junction say, one from the east and one from the north he
would use his expert opinion in controlling the traffic more or less in
the following way:
IF traffic from the east side of the city is VERY-LARGEand traffic
from the north sides is SMALLthen allow movement of traffic to the
east side LONGER
In two stage traffic light system using fuzzy logic, we develop two
different modules namely traffic urgency decision module (TUDM)
and extension time decision module (ETDM). In TUDM 25 fuzzy rules
have been found and for ETDM 16 rules have been found.The someFigure 4: Proposed extension time decision module.
Figure 5: Queue membership.
Figure 6: Timer
membership.
Figure 7: Urgency membership.
Figure 8: Queue-Lane-1 membership.

doi:10.4172/2165-7866.1000162
Page 7 of 9
fuzzy rules are used for designingTUDM and ETDM shown in the
Figure 11 and 12 respectively.
In the fuzzy logic controller once the appropriate rules are fired,
the degree of membership of the output fuzzy variable i.e.,, Urgency is
determinedbyencodingtheantecedentfuzzysubsetsinthiscaseQueue,
Timer
and the output fuzzy variable i.e., Extension-time is determined
by encoding the antecedent fuzzy subsets, in this case Queue-Lane-1,
Queue-Lane-2. In two stage traffic light system using fuzzy logic the
max-min implication technique is used. Using this technique the final
output membership function for each rule is the fuzzy set assigned to
that output by clipping the degree of truth values of the membership
functions of the associated antecedents. Once the membership degree
of each output fuzzy variable is determined all of the rules that are
being fired are then combined and the actual crisp output is obtained
through defuzzification. The procedure of converting each aggregated
fuzzy output set into a single crisp value is called defuzzification. In
traffic urgency decision module and extension time decision modules
we use centroid defuzzification method.
Simulation Result and Discussion
After a two stage traffic light system was carefully designed, we
test the system and discuss the impact of the input variables on the
output variable.With thehelp ofsimulation we show the effect of the
two inputsQueue, Timer
and Queue-Lane-1, Queue-Lane-2to resulted
Urgency andExtension-Timerespectively.
As shown in Figure 13, the Urgency (z-axis) is small when the value
of Queue (x-axis) and Timer
(y-axis) have a small value. The urgency
grows fastly and gets a maximum value when the queue side is being
too many and the timer
density become small. On the other hand,
urgency grows fastly and gets a maximum value when the queue side is
being small and the timer
density become too large.
As shown in Figure 14 as well as Table 2, the extension time (z-axis)
Figure 9: Queue-Lane-2 membership.
Figure 10: Extension-Time membership.
Figure 11: Fuzzy rules for TUDM develop in MATLAB.
Figure 12: Fuzzy rules for ETDM develop in MATLAB.
Figure 13: Input variables Queue, Timer
Vs output variable Urgency using
centroid defuzzification method.
No of Vehicles
(Maximum = 60)
Extension Time
(in Seconds)
Queue Lane-1 Queue Lane- 2 ETDM Actual Method
3 3 15 10
5 5 18 12
10 10 22 16
15 15 25 21
20 20 26 26
25 25 31 31
30 30 35 36
35 35 38 40
40 40 42 45
10 20 26 26
10 25 31 31
10 30 35 36
10 35 38 40
10 40 42 45
20 10 26 26
25 10 31 31
30 10 35 36
35 10 38 40
40 10 42 45
50 10 46 55
60 10 54 64
Table 2: Comparison of ETDM and Actual method.

doi:10.4172/2165-7866.1000162
Page 8 of 9
Figure 14: Input variables Queue-Lane-1, Queue-Lane-2 Vs output
variable Extension-Time using centroid defuzzification method.
Figure 15: Delay Measures.
is small when the value of queue-lane-1 (x-axis) and queue-lane-2
(y-axis) have a small value. The extension time grow slowly and have
a long value when both the queue-lane-1 side and the queue-lane-2 go
to medium to large value. If one of the queue-lane sides constant and
other side increase then extension time also increase this is equivalent
to actual method.
In this research paper, the vehicle arrival rate is Poisson distribution
which is divided into three types such as low traffic rate, middle traffic
rate and high traffic rate. The ranges of each vehicle rate are as 0∼0.3
car per second, 0.3∼0.6 car per second and 0.6∼ 1 car per second
respectively. Table 3 summarizes all the simulation parameters used in
TSTLS (Figure 15).
As shown in Table 2 result of extension time decision module and
actual method are very close. In maximum situation extension time
taken by ETDM is similar to actual method. But ETDM take less time
than fixed time controller without making unstable conditions in the
traffic flow. Extension time of actual method are found with the help of
the following formula-
Average Distance 2.67 20
Extension Time
Average Speed 2.78
n´ +
= =
Wheren is no of vehiclesand average distance is 180 meter (160+20)
the distance from last vehicle in the queue to cover the intersection and
average speed is 10 km/h or 2.78 m/sec (Figure 16 and 17).
The average vehicle delay of pre-timed control system and TSTLS
are shown in following tables. The average vehicle delay of low traffic
rate, middle traffic rate and high traffic rate is shown in Tables 4, 5 and
6 respectively.
Table 7 shows theimprovement percentage of TSTLS with pre-
timed control system which is taken with different traffic rate. The
average vehicle delay in low traffic rate of TSTLS is reduced by 48.91%
compared to pre-timed control system. The average vehicle delay in
middle traffic rate of TSTLS is reduced by 58.65% compared to pre-
timed control system and the average vehicle delay in high traffic rate
of TSTLS is reduced by 39.73% compared to pre-timed control system.
Conclusion and Future Works
The two stage traffic lightsystem using fuzzy logic performed better
than the fixed time system or even vehicle actuated system due to its
flexibility. The flexibility involves the number of vehicles sensed at the
incoming junction and in first stage, determines the maximum urgency
Maximum detectable queue per lane (Vehicles) 60
Maximum Length queue per lane (meters) 160
Maximum Length between parallel stop line (meters) 20
Maximum queue per phase (vehicles) 120
Lost time per phase, TLOST
(Seconds) 4
Saturation flow rate (Vehicles/hour/lane) 2400
Maximum allowable green phase, TMAX
(Seconds) 60
Maximum allowable green phase, TMIN
(Seconds) 8
Low arrival rate (Vehicles/minute) 0-18
Medium arrival rate (Vehicles/minute) 18-36
High arrival rate (Vehicles/minute) 36-60
Fixed yellow time, TY
(Seconds) 4
Table 3: Simulation parameters for intersection.
Figure 16: The whole design structure of TUDM using Mamdani Method.
Figure 17: The whole design structure of ETDM using Mamdani Method.
No. of Simulation Pre-timed control system Two stage traffic light
system (TSTLS)
1 61.98 31.11
2 73.56 39.35
3 69.63 36.69
4 57.33 27.38
5 64.85 32.73
Average 65.47 33.45
Table 4: Average vehicle delay of low arrival rate.

doi:10.4172/2165-7866.1000162
Page 9 of 9
degree ofred phases then in second stage, determines the extension
time of the green phase. In the fixed time system, being an open loop
system phase sequence is not changeable and the green time is not
extended whatever the density of carat the junction. In addition to the
fuzzy variables as mentioned, the fuzzy system also has an advantage
of performing according to linguistics rules in the manner of how a
human would use.
It can be observed from the result that a two stage traffic light
system (TSTLS) provide better performance in terms of average vehicle
delay than pre-timed control system. Therefore TSTLS, improve the
efficiency of vehicles, traffic congestion and hence reduce travel time,
noise pollution, carbon dioxide emission, fuel used and save the time
of human being.
One direction for future research on a proposed two stage traffic
light system is that to enhance the system with a neural network
which is able to forecast the traffic conditions i.e., to predict the traffic
conditions at different times of the day or on different days of the week.
This combination would allow the fuzzy control system to make its
decision taking into account not only the current traffic situation as
detected by the sensor but also the probable short term evolution of the
traffic conditions. In this way the choice of the phase would depend on
the number of vehicles in the queue while the green time duration of
the traffic lights would be determined based on the traffic flow forecast
by the neural network.
Moreover,prospect research direction is to provide the system with
the ability of detecting emergency situation such as the presence in
the queue of ambulance, VIP vehicle and fire trucks etc. through non-
expensive sensors and of implementing suitable contingency actions
so as to prioritize the phase that hosts those vehicles. A further area
of investigation refers to the adoption of low power mechanisms to
reduce the sensor node power consumption as shown in Collotta, et al.
[20] and hence increase the network lifetime.
system (TSTLS)
1 167.56 78.63
2 136.38 59.34
3 156.74 67.21
4 143.23 51.86
5 121.81 42.97
Average 145.14 60.01
Table 5: Average vehicle delay of medium arrival rate.
system (TSTLS)
1 302.78 186.34
2 290.56 173.67
3 287.45 168.96
4 308.67 189.48
5 293.39 175.18
Average 296.57 178.73
Table 6: Average vehicle delay of high arrival rate.
Arrival rate
Low Medium High
Pre-timed control system 65.47 145.14 296.57
TSTLS 33.45 60.01 178.73
Improvement
TSTLS vs. Pre-timed Control
48.91% 58.65% 39.73%
Table 7: Improvement TSTLS vs. Pre-timed control system.
References
1. Shahraki AA, Shahraki MN, Mosavi MR (2013) Design and Simulation of a
Fuzzy Controller for a Busy Intersection. IEEE: 1-6.
2. Porche IR (1997) Dynamic Traffic Control: Decentralized and Coordinated
Methods. Ph.D. Dissertation, Department of Electrical Engineering and
Computer Science, The University of Michigan: 930-935.
3. Yu XH, Recker WW (2006) Stochastic Adaptive Control Model for Traffic Signal
Systems. Transportation Research Part C 14: 263-282.
4. Gosavi A (2003) Simulation-Based Optimization: Parametric Optimization
Techniques and Reinforcement Learning, Springer.
5. Pappis CP, Mamdani EH (1977) A Fuzzy Logic Controller for a Traffic Junction.
IEEE Transactions on Systems, Man, and Cybernetics 7: 707-717.
6. Niittymaki J, Pursula M (2000) Signal Control Using Fuzzy Logic. Elsevier fuzzy
sets and systems 116: 11-22.
7. Chiu S (1992) Adaptive Traffic Signal Control Using Fuzzy Logic. Proceedings
of the IEEE Intelligent Vehicles Symposium: 98-107.
8. Trabia MB, Kaseko MS, Ande M (1999) A two-stage fuzzy logic controller for
traffic signals. Elsevier: Transportation Research Part C 7: 353-367.
9. Soh AC, Rhung LG, Sarkan HM (2010) MATLAB Simulation of Fuzzy Traffic
Controller for Multilane Isolated Intersection. International Journal on Computer
Science and Engineering 2: 924-933.
10. Zaied ANH, Othman WA (2011) Development of a fuzzy logic traffic system
for isolated signalized intersections in the State of Kuwait. Elsevier: Expert
Systems with Applications 38: 9434- 9441.
11. IndrabayulAreni IS, Makobombang NN, Sidehabi SS (2014) A fuzzy logic
approach for timely adaptive traffic light based on traffic load. IEEE: 170-174.
12. Collotta M, Bello LL, Pau G (2015) A novel approach for dynamic traffic lights
management based on Wireless Sensor Networks and multiple fuzzy logic
controllers. Elsevier: Expert Systems with Applications, pp.5403- 5415, March
2015.
13. Wu L, Zhang X, Shi Z (2010) An intelligent fuzzy control for crossroads traffic
light, 2nd
WRI global congress-GCIS 3: 28-32.
14. Mathew T V (2014) Design Principles of Traffic Signal.
15. Zhang W, YUE WL (2001) Alternative solutions for urban traffic congestion,
Proceedings of the Eastern Asia Society for Transportation Studies 3: 327-342.
16. Barto AG, Mahadevan S (2003) Recent Advances in Hierarchical Reinforcement
Learning. Kluwer Academic Publishers 13: 343-379.
17. Wei H, Yong W, Xuanqin W, Yan W (2001) A cooperative fuzzy control
method for traffic Lights. IEEE Intelligent Transportation Systems Conference
Proceedings-Oakland (CA), USA: 185-188.
18. Nakatsuyama M, Nagahashi H, Nishizuka N (1984) Fuzzy Logic Phase
Controller for Traffic Junctions in the One-Way Arterial Road. in Proc. of 9th
IFAC-World Congress, Budapest, Hungary: 2865-2870.
19. Niittymaki J, Nevala R, Turunen E (2003) Fuzzy Traffic Signal Control and a
New Inference Method Maximal Fuzzy Similarity. Researchgate: 716-728.
20. Collotta M, Pau G, Salerno V, Scata G (2011) A fuzzy based algorithm
tomanage power consumption in industrial wireless sensor networks. Industrial
informatics 9th IEEE international conference, India: 151-156.
Citation: Alam J, Pandey MK (2015) Design and Analysis of a Two Stage
Traffic Light System Using Fuzzy Logic. J Inform Tech Softw Eng 5: 162.
doi:10.4172/2165-7866.1000162

Design and-analysis-of-a-two-stage-traffic-light-system-using-fuzzy-logic-2165-7866-1000162

More Related Content

What's hot (20)

Similar to Design and-analysis-of-a-two-stage-traffic-light-system-using-fuzzy-logic-2165-7866-1000162 (20)

Recently uploaded (20)

Design and-analysis-of-a-two-stage-traffic-light-system-using-fuzzy-logic-2165-7866-1000162