processes

Article
Traffic Control Prediction Design Based on Fuzzy Logic and
Lyapunov Approaches to Improve the Performance of
Road Intersection
Sadiqa Jafari †,‡ , Zeinab Shahbazi † and Yung-Cheol Byun *

Department of Computer Engineering, Institute of Information Science Technology, Jeju National University,
Jeju 63243, Korea; [email protected] (S.J.); [email protected] (Z.S.)
* Correspondence: [email protected]
† These authors contributed equally to this work.
‡ Current address: Department of Electronic Engineering, Jeju National University, Jeju 63243, Korea.

Abstract: Due to the increasing use of private cars for urbanization and urban transport, the travel
time of urban transportation is increasing. People spend a lot of time in the streets, and the queue
length of waiting increases accordingly; this has direct effects on fuel consumption too. Traffic flow
forecasts and traffic light schedules were studied separately in the urban traffic system. This paper
presents a new stable TS (Takagi–Sugeno) fuzzy controller for urban traffic. The state-space dynamics
are utilized to formulate both the vehicle’s average waiting time at an isolated intersection and the
length of queues. A fuzzy intelligent controller is designed for light control based upon the length of
the queue, and eventually, the system’s stability is proved using the Lyapunov theorem. Moreover,




the input variables are the length of queue and number of input or output vehicles from each lane.

 The simulation results describe the appearance of the proposed controller. An illustrative example is
Citation: Jafari, S.; Shahbazi, Z.; also given to show the proposed method’s effectiveness; the suggested method is more efficient than
Byun, Y.-C. Traffic Control Prediction both the conventional fuzzy traffic controllers and the fixed time controller.
Design Based on Fuzzy Logic and
Lyapunov Approaches to Improve Keywords: traffic light; urban traffic; fuzzy intelligent control; Lyapunov stability
the Performance of Road Intersection.
Processes 2021, 9, 2205. https://
doi.org/10.3390/pr9122205
1. Introduction
Academic Editors: Faisal Jamil and
One of the biggest issues in the world is urban traffic and transportation between cities.
Shabir Ahmad
More than a million people spend a lot of time in traffic every day, and the total time wasted
in traffic reaches millions of hours. Another disadvantage of traffic is air pollution. Urban
Received: 11 November 2021
Accepted: 3 December 2021
traffic has three main factors: human, road, and vehicle. If none of these factors are present,
Published: 7 December 2021
urban traffic is not be generated. Recently, traffic became one of the most critical factors
disrupting urban transportation. When the flow is smooth, fast, and without significant
Publisher’s Note: MDPI stays neutral
environmental impact, the city has good traffic. An example of the essential characteristics
with regard to jurisdictional claims in
of developed cities is the permanent population extension in a comparatively short region.
published maps and institutional affil- These results are the rise in the number of cars and the need to move and transport people
iations. in urban city tracks. As a result, the control and monitoring of traffic became the central
issue in many countries [1–3].
The intelligence of urban systems such as traffic lights is another factor that can play
a very influential role in reducing traffic. Also, some countries are developing different
Copyright: © 2021 by the authors.
technologies to reduce traffic and increase the volume of highways. Because the traffic
Licensee MDPI, Basel, Switzerland.
system is complicated, random, nonlinear, and discrete, it is difficult to control it using
This article is an open access article
an accurate model. The use of fixed interval control and operated control cannot be
distributed under the terms and entirely adequate [4–6]. In the last decade, numerous researchers applied fuzzy logic in
conditions of the Creative Commons traffic signal control, including type-1 and type-2 fuzzy systems [7]. Fuzzy logic relies
Attribution (CC BY) license (https:// on scientific interpretations of personal knowledge and its experiments. Fuzzy Logic
creativecommons.org/licenses/by/ controllers were vigorously implemented in several policies with uncertainties. Fuzzy
4.0/). logic, first introduced by Zadeh in [8], is based on the scientific order of personal knowledge

Processes 2021, 9, 2205. https://doi.org/10.3390/pr9122205 https://www.mdpi.com/journal/processes

Processes 2021, 9, 2205 2 of 14

and his skills. Fuzzy logic controllers were favorably utilized in various systems that have
built-in possibilities. Fuzzy logic controllers were vigorously applied in various practices
that have integrated uncertainties [9]. A compact summary of the first importance of fuzzy
logic to traffic signal control is as follows: Pappis and Mamdani in [10] studied the control
of a private traffic intersection with easy narrow east-west/north-south traffic control
with casual vehicle approaches and no turning changes. Fuzzy rules were magnified to
assess the suitability of sustaining a popular green phase across numerous durations, based
on superimposing a metric size of resolution deadline. Specific extensions were related,
and the one delivering the most critical quality of confidence was considered. This method
considered connecting the period length divides and offsets per intersection using solely
social traffic data. Regulations on the signal period length and divisions were performed
based upon the plethora level for the respective method of an intersection. The fuzzy logic
traffic lights controller performed the task completely as opposed to the the fixed-time
controller and vehicle-operated controllers, owing to its flexibility. Its versatility allowed
it to engage the number of cars perceived at the incoming intersection and the addition
of the green time. A modern fuzzy logic controller for a detached acclaimed intersection
was offered in [11]. In the big cities, traffic density reduced people’s quality-of-life. There
are reasons for creating urban traffic, however, traffic is mainly on roads, for example,
in complex road structures where vehicles often cross due to driver intentions and problems
with the traffic management system. In this article, fuzzy logic control in urban traffic,
a strategy to use the time of traffic lights with the state, is recommended. The traffic light
system is controlled to decrease queue length and status entry [12]. This method controls
the traffic light timings and phases classification to guarantee constant traffic movement
with the least amount of waiting and lowest length of queue. This implementation of
an original traffic light control scheme using fuzzy logic technology is presented in [1].
Related to the other research that concentrates on the fuzzy controller, this paper deals
with a class of stable fuzzy controllers. We assumes the traffic situation would be too hectic
for the local fuzzy controllers to control it efficiently. As such, the unique controller will
optimize the green time delays of the complete intersections using a simulated annealing
algorithm. During the study, a novel TS fuzzy controller for urban traffic is proposed.
The state-space dynamics are utilized to express the vehicle’s standard waiting time at a
separate intersection and length of queues. A fuzzy intelligent controller is designed for
light control based on the length of the queue, and the system’s stability was proved using
the Lyapunov theorem.
More precisely, the remainder of this article is as follows:
• A comprehensive model of vehicle behavior for the city’s transportation system at
intersections on two levels is provided.
• We tested and imitated the given intersection model in traffic conditions.
• We show that the fuzzy controller design reduced the number of vehicles waiting in
line and the time required to wait at red lights.
• Evaluation of the average number of cars in a row in two ways, one without a
controller and one with a controller over a certain duration.
In this investigation, we propose a new TS fuzzy controller for urban traffic, in
which we used space dynamics from a signal intersection to model. We proved the
system’s stability using the Lyapunov theorem. Finally, we designed it based on a model
fuzzy control.
The rest of the article is composed as follows: Section 2 will briefly explain traffic
control history and the reason for using fuzzy control and experts. In Section 3, the state-
space difference of the single intersection is described. Then, in Section 4, we explain the
proposed framework. Section 5 states the designation of the TS fuzzy controller for the
single intersection. In Section 6, we describe the fuzzy controller’s mathematical model.
The simulation results for two fixed-time controllers and fuzzy intelligent control classes are
mentioned in Section 7. In Section 8, the conclusions are mentioned; finally, some suggested
methods will be presented to make the fuzzy process more efficient and accurate.
Processes 2021, 9, 2205 3 of 14

2. Literature Review
2.1. Overview Control Traffic
Today’s traffic is one of the major global issues, and an urban traffic system is essential
for managing daily movement. Developing an urban street network between the network
and related traffic flow can improve the preparation capabilities of urban systems. This
study [13] proposes a new model of urban transport in megacities using a multiagent
system and pertinent method. This used to be used for complex urban traffic, based on
interconnectivity, i.e., a traffic signal with Petri Nets to show a natural behavior. The process
of urban systems can assist in the comprehension of the urban street network. This article
studies the impact of the character proportion of a square urban network on the dynamics
of traffic systems. The application of road networks was observed using the Macroscopic
Fundamental Diagram (MFD) [14]. Due to the significant increase in vehicles in modern
decades, urban traffic congestion became increasingly severe. Furthermore, urban over-
crowding creates noise pollution. There are collected dynamical models of city-scale traffic
that can help to develop model-based perimeter control methods. Controlling traffic in
urban road networks remains challenging. As such, it was suggested a nonlinear model
with a predictive perimeter control system for ordinance and commercial optimization
goals, with closed-loop stability during development, which is for the effective and reliable
control of city-scale traffic [15–18]. Traffic light control and the flow of urban traffic are
essential elements for city traffic management. This paper proposes a genetic scheduling
model for traffic light control, which has a status update feature that was developed to
customize road signs. Also, this model can improve the cycle of road signals at different
intersections dynamically [19].

2.2. Fuzzy Controller-Based Traffic Control

Traffic congestion is an as-of-yet unsolved global problem. If we suppose that new
roads and additional capacity continue to be added, this problem will become more promi-
nent in areas where unlimited access roads (such as arterial roads) and limited access roads
(such as highways) are connected. To assist with this issue, it were proposed a fuzzy model
for a complex traffic control system and a Diverging Diamond Interchange (DDI)—Ramp
Metering (RM)—for over-saturated traffic situations, which means an alternative inter-
section geometry should be used in the intersection without conflict. The recommended
fuzzy model in this paper was designed to stop a queue from the metering grade to the
DDI; in previous work, it improved traffic situations slightly [20]. Congestion is a common
critical problem, especially in large cities. There is congestion at intersections that were
connected to different ones, such as the analytical traffic light continuance, which does
not match traffic situations. Here, we propose a form of a traffic light controller using the
Takagi–Sugeno–Kang method fuzzy logic. The research points to produce a green light
continuance calculation using fuzzy logic [21].
Life’s pace is quickening, and people feel anxious about the long travel times. Reason-
able traffic control strategies could decrease the wait time for commuters. In the past, traffic
control research was focused primarily on time sequence, ignoring phase investigation.
The typical timing system based on a single loop phase structure cannot achieve optimal
control for asymmetric or small traffic flows. Therefore, this article proposes a fuzzy con-
trol method for signals in a time-changing world and a way to develop the topological
structure. The delay of vehicles was diminished, and the traffic potential of the intersection
was improved in [22]. Traffic lights play a pivotal role in urban traffic control. This essay
discusses the control artifice and a performance review of a single-point intersection of
traffic signals. Furthermore, a fuzzy model control was used to modify the signal timing
dynamically [23]. Flow traffic at all traffic intersections is used to control the traffic lights.
This study suggests a fuzzy inductive control system that uses physical logic to determine
the phase sequence of each circuit and time of green exposure based on input parameters,
which were designed to improve traffic flow at different intersections during peak hours
and assist with the transportation of emergency vehicle.
Processes 2021, 9, 2205 4 of 14

Here, we show in Table 1 a survey of methods used in the direction system.

Table 1. Related studies of urban traffic.

Author Model Objective Limit

Multitask learning
Mostafa et al. [24] Predict traffic flow Urban traffic flow
long short term memory
Forecasting urban
Juan et al. [25] Deep learning Urban traffic
traffic
Multiagent deep
Stabilize the Large-Scale
Tianshu et al. [26] reinforcement
learning procedure traffic signal
learning
S. Neelakandan et al. [27] IoT-based traffic prediction Optimization Smart city
Effective hybrid-heuristic Optimization speed of
Chun-Wei et al. [28] Urban traffic light
algorithm the search algorithm
Handle large and
Hyunjin et al. [29] Reinforcement learning Traffic signal control
complex data
Long short-term memory Real-time signal queue
Rezaur et al. [30] Urban traffic
neural network length prediction
Mingtao et al. [31] Multiagent Optimization Urban traffic signal
Hongluan et al. [32] PSO Optimization Road intersections
Tuo et al. [33] Boosted genetic algorithm Optimization Traffic control
Reduce congestion and Reduce congestion
Ilyas et al. [34] Urban traffic
improve the safety and improve
Jinbao et al. [35] Multiobjective Optimization Intersection traffic
Machine learning
Mohammad et al. [36] Optimization Traffic signal control
techniques

3. State-Space Difference of the Single Intersection

The two phases signalized intersection configuration for single intersections is shown
in Figure 1.
In this scenario, leg one and leg 3 are phase 1, and leg two and leg 4 are phase 2.
The length of the queue at an intersection represents the traffic state. The queue is consid-
ered as follows [37]:
Ui (n + 1) = Ui (n) + qi (n) − di (n)zi (n) (1)
Here, Equation (1) one shows the index of the traffic streams (i = 1 TO m), the dis-
cretized time intervals (n = 0 TO n − 1), the length of the queue of i stream at the onset of
the n time interval (U ), the number of vehicles that join the i queue in the n time interval
qi (n), the number of cars that depart di (n), and traffic lights (z = 1 is green and z = 0 is
red), which were typically distributed random signals. If T is seen as a discretized time
interval, and if it is quickly sufficient, it can be assumed that vehicle-arrivals in each time
interval are uniform. Consequently, Wi (n) the total waiting time of vehicles is reached with
Equation (2) [37].

Wi (n + 1) = Wi (n) + TUi (n) − 1/2Tdi (n)zi (n) + 1/2Tqi (n) (2)

The formulation to facilitate state-space comparisons and the optimization goal can be
edited in matrix arrangement, as seen in [37]:

K (n + 1) = FK (n) + E(n) Z + C (n)
(3)
Y (n) = CK (n)

In Equation (3), there are the vector of state variables, which include:

K (n) = [U1 (n)U2 (n) . . . U M (n)W1 (n)W2 (n) . . . WM (n)] T (4)

and
Z (n) = [ Z1 (n) Z2 (n) . . . Z M (n)] T (5)
Processes 2021, 9, 2205 5 of 14

At urban intersections, the traffic control signal is a traffic light. The vehicles during
the green phase can enter or depart from the queue; during the red phase, the vehicles can
only join the queue. The parameters used in the suggested approach are shown in Table 2.

Figure 1. Two-phase cross signal.

Table 2. Notations used in this paper.

Component Description
Ui (n) Queue length
qi ( n ) The number of vehicles entering the queue
di ( n ) The number of vehicles leaving the queue
Zi (n) Control signal
Wi (n) Waiting time
T Sampling time
K (n) The vector of variables of mode
Z (n) Control signal
IM Identity matrix
Fi , Ei , Ci Metrics
α Traffic condition
P Positive definite

4. The Proposed Framework

One of the characteristics of an optimal traffic control system is flexibility in environ-
mental changes; however, the most important characteristic is intelligence. The correct
and accurate operation of fuzzy system is based on the knowledge and experience of the
expert operating the control system. With the fuzzy method, the experiment involved
a nonlinear function as a fuzzy model of a system. Based on the fuzzy model, special
controller systems can be designed, which can be very simple. With the fuzzy logic method,
the obtained linguistic scores given by the professionals were converted to appropriate
fuzzy set numbers. The fuzzy logic structure allows the professionals to capture the experts’
opinions in linguistic terms and evaluate the overall ratio level. Also, by using fuzzy
sets, the comparison process becomes more confident. Further maintenance support can
Processes 2021, 9, 2205 6 of 14

develop using a fuzzy-based organization. The primary supervisor of traffic is an expert

police officer, and their knowledge can be transferred to write the formula form. Fuzzy
mathematics includes fuzzy sets and related mathematical operations. In classical sets,
the boundaries of the collection are defined precisely and definitively and an element is
either a member of the set or not. Hence, the classical set’s basic premise is that the exact
definition is the set’s boundary, and the elements’ membership have no more than two
models. Still, there are many cases where it is impossible to define an exact and definite
boundary for the reader. The membership of elements in fuzzy sets is also expressed by
the degree of membership, which is a number between zero and one [38].

Fuzzy Logic-Systems
Fuzzy logic is evolved, while classical logic is generalized. In classical logic, which
is a dual value logic, any statement can be true or false. Therefore, approximate and
inaccurate judgments are possible by applying fuzzy logic [38]. There are systems whose
input information can be incorrect, i.e., the information input into a fuzzy system can
be inaccurate. One of the most well-known and inaccurate processes in fuzzy systems
is the use of fuzzy rule databases. In the fuzzy law database, each law is defined by an
“if-then” structure. Fuzzy logic is a great way to use quality information to design system
controllers and gained a special place in the control of laboratory and industrial processes
in the last decade. There is no specific way to design a fuzzy controller when controlling
the process based on personal information. The fuzzy logic control method is an expert
method, and the performance of a fuzzy controller depends largely on the experience of
the user. Individual experiments were used as a nonlinear function as a fuzzy model of a
system. Based on the fuzzy model, unique controller systems can be designed, and the idea
of controller design can be very simple. It includes all rules and regulations established
by experts to control the decisions of the decision-making system. Under the new fuzzy
method, the rules and regulations can be modified and minimized to achieve the best
results with minimal rules.
Figure 2 is a fuzzy structure logic that includes four main sections: rules, fuzzifier,
defuzzifier, and intelligent. As shown in Figure 2, a general structure of fuzzy systems is
described because several choices were provided for each block, and various compounds
creates different fuzzy systems [38].
There are systems whose input information can be inaccurate; input information in
a fuzzy system can be incorrect. An example of the most well-known and inaccurate
processes in fuzzy systems is a fuzzy rule database. In the fuzzy command database,
special law was restricted by an “if-then” installation.

Figure 2. Configuration of fuzzy system.

Processes 2021, 9, 2205 7 of 14

All the rules and conditions are governed by the decision-making system, if stated by
an expert. According to the new fuzzy method, the rules and regulations can be modified
and minimized to achieve the best results with minimal rules.
Below are the characteristics of fuzzy logic follows [39]:
• Flexible implementation; computational seriousness of machine learning methods.
• Ability to simulate human logic and thinking.
• Ability to create two solutions or answers to a problem.
• A good solution to a problem with an expected answer.
• A practical approach that uses flexible states in logic for reasoning.
• Ability to generate nonlinear functions using random complex numbers.
• Strong dependence on the researcher in creating fuzzy logic models.

5. Designing Stable TS Fuzzy Controller for Single Intersection

In this section, we outline how the fuzzy controller was developed. The fuzzy system
has three inputs and one output, and we used Takagi–Sugeno–Kang fuzzy system. The in-
puts include the length of the queue for vehicles and the number of vehicles entering or
leaving each leg, which in phase 1 is as follows:

(U1 (n) + U3 (n), U2 (n) + U4 (n), d1 + d3 , q1 + q3 ) (6)

For phase two, it is as follows:

(U1 (n) + U3 (n), U2 (n) + U4 (n), d2 + d4 , q2 + q4 ) (7)

The output control signals respectively are Z1 and Z2 . The format of fuzzy control rules is
as follows:

I f (U1 (n) + U3 (n)is(k1 ), U2 (n) + U4 (n)is (8)

(k2 ), (di + d j )is(k3 ) and(qi + q j )then( Zi is(y))
In fuzzy controller, y = f (u) is a constant function that receives values of 0 or 1
(0 means stop and 1 means move). Moreover, for input variables, we consider the linguistic
variables in three fuzzy sets, low ( L), medium ( M), and high ( H ), in the length of the
queue, and low ( L) and high ( H ) for the number of vehicles that enter or leave the leg.
Altogether, there are 36 rules. The membership function of the input variable is a triangle.
For the queue length, the interval is [0100], and the number of vehicles entering or leaving
in each leg interval was considered [05]. (Ui (n)) and ( Zi ) were defined for i = 1, 2, 3, 4.
An example of the fuzzy rules that were used in two controllers are shown in Table 3.

Table 3. One example of fuzzy rules.

Input Fuzzy Rules

If (U1(n) + U3(n) is H,(U2(n) + U4(n) is H),
A sample of rules of Z1
(d1 + d3 is H) and (q1 + q3 is L), then (Z1 is 0)
If (U1(n) + U3(n) is M, (U2(n) + U4(n) is M),
A sample of rules of Z2
(d2 + d4 is H) and (q2 + q4 is L), then (Z2 is 0)

6. Mathematical Model of the Controlled System

With the TS fuzzy method used here, the dynamic model considered a system from
an intersection; then, we used TS fuzzy for the stability of state-space equations at a
single intersection.

Theorem 1. Consider the dynamical system given in (3), then the controller structure given
in (9) and (10) makes the states of the system uniformly bounded, as well as all signals involved in
the closed-loop system.
Processes 2021, 9, 2205 8 of 14

Proof. For the green phase, the length of the queue and the waiting time equations were
defined as follows:

Ui (n + 1) = Ui (n) + qi (n) − di (n)
(9)
Wi (n + 1) = Wi (n) + TUi (n) + 1/2Tqi (n) − 1/2Tdi


Ug
which K g = and state vector was considered as follows:
Wg

K g (n + 1) = Fg (n)K g(n) + Eg qi + Eg0 di (10)

The red phase in the traffic light, stability of the queue, and the waiting time equations
were defined as follows:

Ui (n + 1) = Ui (n) + qi (n)
(11)
Wi (n + 1) = Wi (n) + TUi (n) + 1/2Tqi (n)
 
Ur
which Kr = and state vector is as follows:
Wr

Kr (n + 1) = Fr (n)Kr(n) + Er qi (12)

A Lyapunov function can be used to prove stability at an equilibrium point in phase

space if it is defined in the phase space. Lyapunov functions can determine a system’s
stability or instability. In addition, this method has the advantage that it does not necessitate
knowledge of the actual solution. By means of this calculation, the stability of nonrough
equilibrium points can be studied. In this segment, the Lyapunov function must find
and prove that the equations are stable because the equations’ discretion is the Lyapunov
function as follows:  T   
Kg P1 0 K g
V (K ) = (13)
Kr 0 P2 Kr
which P1 , P2 are definitely symmetrically positive. The difference of the Lyapunov function:

∆V (n) = V (n + 1) − V (n) = K (n + 1) PK (n + 1) − K (n) PK (n) (14)

To prove the stability of the closed-loop system using Equation (14), it can be rewritten
as follows:

∆V (n) = K gT (n + 1) P1 K g (n + 1) + KrT (n + 1) P2 Kr (n + 1) − K gT (n) P1 K g (n)

(15)
−KrT (n) P2 Kr (n)

The Equation (16) is represented as follows:

∆V (n) = ( Fg K g (n) + Eg qi + Eg0 di ) T P1 ( Fg K g (n) + Eg qi + Eg0 di ) + ( Fr Kr (n) + Er qi ) T

(16)
P2 ( Fr Kr (n) + Er qi ) − FgT (n) P1 Fg (n) − FrT (n) P1 Fr (n)

After some mathematical manipulation, the above equation is as follows:

∆V (n) = K gT (n)[ FgT P1 Fg − P1 ]K g (n) + qiT [ EgT P1 Eg + ErT P2 Er ]qi + diT Eg0T P1 Eg0 di
(17)
+KrT (n)[ FrT P2 (n) Fr − P2 ]Kr (n)

To simplify the formula, we use the following definitions:

F T Pi F − I = −Ui (18)

Furthermore,
λmin ( p)k Z k2 ≤ Z T PZ ≤ λmax ( p)k Z k2 (19)
Processes 2021, 9, 2205 9 of 14

Consequently, we used (18) and (19); thus, we have the following equation:

V (n) = −K gT U1 K g − K gT U2 K g + ( Eg0 di ) T P1 ( Eg0 di ) + ( Eg qi ) T P1 ( Eg qi )

(20)
+( Er qi )T P2 ( Er qi )

To prove the stability of the closed-loop system, the following definitions are considered:
2
( Eg0 di )T P1 ( Eg0 di ) ≤ λmax ( p1 ) Eg0 di ≤ λmax ( p1 ) Eg0 kdi k2 ≤ αkdi k2 ( Eg qi )T P1 ( Eg qi )
2
+( Er qi )T P2 ( Er qi ) ≤ λmax ( p1 ) Eg kqi k2 + λmax ( p2 )k Er k2 kqi k2 ≤ βkqi k2 − K gT U1 K g (21)
2 2
−KrT U2 Kr ≤ −λmin1 (U1 ) K g − λmin2 (U2 )kKr k2 ≤ −λmin1 K g − λmin2 kKr k2

and furthermore, the Equation (22) can be assumed without loss of generality.

kqi k ≤ qmax
(22)
kdi k ≤ dmax

By using the Equations (21) and (22), δV can be summarized as follows:

2
∆V ≤ −λmin1 K g − λmin2 kKr k2 + αkdi k2
2 (23)
+ βkqi k2 ≤ −λmin1 K g − λmin2 kKr k2 + ξ 1 + ξ 2

The differences with the Lyapunov system are uniformly in the compact set mentioned
in Equation (23).  
kK g k≥ξ 1
Ω = K| λ and Xr ≥ ξ ξ2
(24)
min1 min2

Finally, the proof is complete.

7. Simulation Results
To show the effectiveness of the suggested controller, the proposed method is applied
to the system mentioned in Equation (1), in which attention to queue length and delay,
as well as available vehicles in the traffic were optimized. It is assumed that T = 6(s) is
a sampling time. α shows values in traffic conditions. Regarding this Table, we can see
different traffic conditions in the urban traffic for i = 1, 2, 3, 4; the α parameter is between 0
and 1 so that its variations conform to Table 4. The result of the simulation for fixed-time
control and fuzzy intelligent control were discussed as follows [37].

Table 4. Different rate traffic by modifications of α.

Traffic Condition α
Nonimpregnation α ≥ 0.7
Impregnation 0.4 ≤ α ≤ 0.6
Super-impregnation 0.1 ≤ α ≤ 0.3
Unstable 0=α

7.1. Fixed-Time Control

In this section, we apply a model fixed-time controller to compare with the proposed
model. The aim of the simulation is the decrement of the queue of length. In the fixed-time
control method, the simulation results for the length of the queue and waiting time of
vehicles without a controller are shown as follows. Figure 3 shows the number of vehicles
in legs 1, 2, 3, and 4 in the intersection without a controller. The performance without a
controller is not promising. For example, Figure 3 describes the number of vehicles in the
second queue without a lower controller than other legs. Figure 3 displays the number of
vehicles in the third queue without a controller that initially increased and then decreased.
Processes 2021, 9, 2205 10 of 14

Another example from Figure 3 is the fourth leg: this leg’s traffic is greater than that of the
other legs, so it was therefore improved using the fuzzy controller.

Figure 3. Number of vehicles lined up at intersections without controllers.

7.2. Fuzzy Intelligent Controller

In this part, we show how to develop a fuzzy system in a signal intersection. The out-
put of the fuzzy controller as the control variable ( Zi ) of the traffic system, which controls
input, is a green or red phase in each leg.
Figure 4 showed the system’s total output at each leg in an intersection by using model
intelligent fuzzy control. To see how robust our method is in urban traffic for reducing
queue length, we consider the same case, whose results are shown in Figure 4, but it is an
on-based, fixed-time controller. Figure 4 shows that the controller performs satisfactorily.
The length of the queue of vehicles in the first leg reduces compared to that of fixed-time
control. Also, the number of vehicles in the second queue decreases compared to that of
fixed-time control. Moreover, the number of vehicles in the third leg was reduced compared
to that of fixed-time control. With attention to results, the length of the queue of vehicles in
the fourth leg reduced compared to that of fixed-time control.

Figure 4. Number of vehicles in queue length of an intersection using fuzzy controller.

8. Conclusions
In this article, a fuzzy model of an urban traffic network was designed for a single
intersection. The length of queues and the average waiting time for the vehicles are the
model’s state variables. Moreover, both the vehicle’s average waiting time at an isolated
intersection and the length of queues are considered controller inputs. Additionally,
the effectiveness of the suggested controller is verified by simulation results. As shown in
Processes 2021, 9, 2205 11 of 14

Table 5 and Figure 5, the percentage of improvement with attention to simulation results
using the proposed method decreases the vehicles in each intersection phase to the fixed-
time control.

Table 5. The performance of vehicles queue length results based on fuzzy intelligent and fixed-
time control.

Queue Length Fixed-Time Control Fuzzy Intelligent Control Total Improvement Percentage
U1 45 9 80.00
U2 15 4 73.33
U3 30 5 83.33
U4 260 23 91.15
SUM 350 41 88.29

Figure 5. Comparison of results in length of queue of vehicles using a fuzzy model.

9. Discussion and Future Direction

Recently, the application of fuzzy sets in automatic control attracted much attention.
In this section, we discussed the fuzzy model, whose model knowledge was categorized as
specific and implicit, according to formation and transmission models. Specific knowledge
is the knowledge stored in documents and computers, such as statistical information on
the changes of traffic parameters. At the fuzzy model intelligent control, the initial work
was performed with the goal of imitating an experienced and specialized human operator
(Knowledge Base). A continuous addition is found between discrete outputs that are
regularly obtained (fuzzy logic section). An example of this is the knowledge of traffic
police in manual control of traffic lights of many intersections. We need both kinds to
achieve a fuzzy control system for synchronizing the traffic light in a given intersection.
Therefore, specific knowledge of the field values related to the design parameters of the
fuzzy membership functions and the implicit understanding of the decision-construction
rules for the change of the green phase should be possible. Table 6 shows the comparison
of the fuzzy controller with the recent state-of-arts in this field. The advantages and
disadvantages of fuzzy logic model are listed below.
The advantages of fuzzy logic include:
• The logic of fuzzy is close to human reasoning.
• Fuzzy logic programs are fast and inexpensive.
• It can be easily designed.
• It is widely used in decisive control and forecasting systems.
Disadvantages of fuzzy logic include:
• Experts determine the principles of logic in fuzzy.
• Analyzing a system using fuzzy logic is not easy, as the reaction is unpredictable.
• Features can be time-consuming and incorrectly fixed.
Processes 2021, 9, 2205 12 of 14

Fuzzy logic is one of the most robust implementations of fuzzy sets, where the vari-
ables are in non-numerical language. Fuzzy logic was the opposite of binary or Aristotle’s
logic, which sees everything in two ways: yes or no; black or white; zero or one. These
logical changes range from 0 to 1. In this case, we proposed a fuzzy model controller to
schedule traffic lights in a single intersection in urban traffic to reduce the queue length
with attention to delay time of cars in the traffic. However, in the proposed model, mem-
bership function parameters are not optimal, and therefore, for better efficiency and results
in the future, an example of tasks that can be performed to continue working on the issue
of urban traffic control are:
• Fuzzy membership function parameters can be selected optimally with methods.
• The functions of the memberships used triangular, which can be derived from other
membership functions, such as Gauss, etc.
• Investigate classical control techniques such as optimal control, adaptive control,
robust control, etc., to design a coordinated traffic signal control system for several
adjacent intersections.

Table 6. Comparison of fuzzy controller with other works.

Author Model Problem Solution Advantage

Reducing traffic
congestion
Sung-Soo et al. [40] Congestion diffusion Urban traffic Prediction
and improving
transportation
Combining LSTM
Jinjun et al. [41] Traffic flow Predictive Optimize
with genetic algorithm
Management systems
Genetic algorithms and Improve the
Daming et al. [42] Urban traffic and alleviating urban
image perception efficiency
traffic congestion
Model predictive Traffic at
Johanna et al. [43] Predictive Optimization
control intersections
Reducing the
Optimization
Jafari et al. [44] MPC Urban Traffic queue length
Prediction
of vehicle
Reduce the
Fuzzy intelligent length of queue Optimization
Proposed model Urban traffic
controller and average Prediction
waiting time

Author Contributions: Writing original draft, Data curation, S.J., Z.S.; review and editing, S.J.;
investigation, Z.S.; methodology, Z.S.; project administration, S.J.; supervision, Y.-C.B.; Supervision,
Y.-C.B.; validation, Z.S.; visualization, Z.S. All authors have read and agreed to the published version
of the manuscript.
Data Availability Statement: Not Available.
Acknowledgments: This work was supported by Korea Institute for Advancement of Technology
(KIAT) grant funded by the Korea Government (MOTIE) (P0016977, The Establishment Project of
Industry-University Fusion District).
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Zhang, L.; Zhao, Q.; Wang, L.; Zhang, L. Urban intersection signal control based on time-space resource scheduling. IEEE Access
2021, 9, 49281–49291. [CrossRef]
2. Shahbazi, Z.; Byun, Y.C. A framework of vehicular security and demand service prediction based on data analysis integrated
with blockchain approach. Sensors 2021, 21, 3314. [CrossRef]
3. Shahbazi, Z.; Byun, Y.C. A Procedure for Tracing Supply Chains for Perishable Food Based on Blockchain, Machine Learning and
Fuzzy Logic. Electronics 2021, 10, 41. [CrossRef]
4. Shahbazi, Z.; Byun, Y.C. Integration of Blockchain, IoT and Machine Learning for Multistage Quality Control and Enhancing
Security in Smart Manufacturing. Sensors 2021, 21, 1467. [CrossRef] [PubMed]
Processes 2021, 9, 2205 13 of 14

5. Shahbazi, Z.; Byun, Y.C. Smart Manufacturing Real-Time Analysis Based on Blockchain and Machine Learning Approaches.
Appl. Sci. 2021, 11, 3535. [CrossRef]
6. Shahbazi, Z.; Byun, Y.C. Blockchain and Machine Learning for Intelligent Multiple Factor-Based Ride-Hailing Services. CMC-
Comput. Mater. Contin. 2022, 70, 4429–4446. [CrossRef]
7. Jiang, T.; Wang, Z.; Chen, F. Urban Traffic Signals Timing at Four-Phase Signalized Intersection Based on Optimized Two-Stage
Fuzzy Control Scheme. Math. Probl. Eng. 2021, 2021. [CrossRef]
8. Zadeh, L.A. Outline of a new approach to the analysis of complex systems and decision processes. IEEE Trans. Syst. Man Cybern.
1973, 28–44. [CrossRef]
9. Chatterjee, K.; De, A.; Chan, F.T. Real time traffic delay optimization using shadowed type-2 fuzzy rule base. Appl. Soft Comput.
2019, 74, 226–241. [CrossRef]
10. Pappis, C.P.; Mamdani, E.H. A fuzzy logic controller for a trafc junction. IEEE Trans. Syst. Man Cybern. 1977, 7, 707–717.
[CrossRef]
11. Vogel, A.; Oremović, I.; Šimić, R.; Ivanjko, E. Improving traffic light control by means of fuzzy logic. In Proceedings of the 2018
International Symposium ELMAR, Zadar, Croatia, 16–19 September 2018; pp. 51–56.
12. Tunc, I.; Yesilyurt, A.Y.; Soylemez, M.T. Different Fuzzy Logic Control Strategies for Traffic Signal Timing Control with State
Inputs. IFAC-PapersOnLine 2021, 54, 265–270. [CrossRef]
13. Geronimo, M.F.; Martinez, E.G.H.; Vazquez, E.D.F.; Godoy, J.J.F.; Anaya, G.F. A multiagent systems with Petri Net approach for
simulation of urban traffic networks. Comput. Environ. Urban Syst. 2021, 89, 101662. [CrossRef]
14. Wu, C.Y.; Hu, M.B.; Jiang, R.; Hao, Q.Y. Effects of road network structure on the performance of urban traffic systems. Phys. A
Stat. Mech. Its Appl. 2021, 563, 125361. [CrossRef]
15. Sirmatel, I.I.; Geroliminis, N. Stabilization of city-scale road traffic networks via macroscopic fundamental diagram-based model
predictive perimeter control. Control Eng. Pract. 2021, 109, 104750. [CrossRef]
16. Jamil, F.; Kim, D. Payment mechanism for electronic charging using blockchain in smart vehicle. Korea 2019, 30, 31.
17. Ahmad, S.; Jamil, F.; Khudoyberdiev, A.; Kim, D. Accident risk prediction and avoidance in intelligent semi-autonomous vehicles
based on road safety data and driver biological behaviours. J. Intell. Fuzzy Syst. 2020, 38, 4591–4601. [CrossRef]
18. Ahmad, S.; Jamil, F.; Iqbal, N.; Kim, D. Optimal route recommendation for waste carrier vehicles for efficient waste collection: A
step forward towards sustainable cities. IEEE Access 2020, 8, 77875–77887. [CrossRef]
19. Wang, H.; Hu, P.; Wang, H. A genetic timing scheduling model for urban traffic signal control. Inf. Sci. 2021, 576, 475–483.
[CrossRef]
20. Jovanović, A.; Kukić, K.; Stevanović, A. A fuzzy logic simulation model for controlling an oversaturated diverge diamond
interchange and ramp metering system. Math. Comput. Simul. 2021, 182, 165–181. [CrossRef]
21. Kartikasari, R.Y.; Prakarsa, G.; Pradeka, D. Optimization of Traffic Light Control Using Fuzzy Logic Sugeno Method. Int. J. Glob.
Oper. Res. 2020, 1, 51–61. [CrossRef]
22. Zhou, C.; Mo, H.; Chen, X.; Wen, H. Fuzzy Control Under Time-Varying Universe and Phase Optimization in Traffic Lights
(ICSSE 2020). Int. J. Fuzzy Syst. 2021, 23, 692–703. [CrossRef]
23. Yi-Fei, W.; Zheng, G. Research on polling based traffic signal control strategy with fuzzy control. In Proceedings of the 2018 IEEE
4th International Conference on Computer and Communications (ICCC), Chengdu, China, 7–10 December 2018; pp. 500–504.
24. Karimzadeh, M.; Schwegler, S.M.; Zhao, Z.; Braun, T.; Sargento, S. MTL-LSTM: Multi-Task Learning-based LSTM for Urban
Traffic Flow Forecasting. In Proceedings of the 2021 International Wireless Communications and Mobile Computing (IWCMC),
Harbin, China, 28 June–2 July 2021; pp. 564–569.
25. Vázquez, J.J.; Arjona, J.; Linares, M.; Casanovas-Garcia, J. A comparison of deep learning methods for urban traffic forecasting
using floating car data. Transp. Res. Procedia 2020, 47, 195–202. [CrossRef]
26. Chu, T.; Wang, J.; Codecà, L.; Li, Z. Multi-agent deep reinforcement learning for large-scale traffic signal control. IEEE Trans.
Intell. Transp. Syst. 2019, 21, 1086–1095. [CrossRef]
27. Neelakandan, S.; Berlin, M.; Tripathi, S.; Devi, V.B.; Bhardwaj, I.; Arulkumar, N. IoT-based traffic prediction and traffic signal
control system for smart city. Soft Comput. 2021, 25, 12241–12248. [CrossRef]
28. Tsai, C.W.; Teng, T.C.; Liao, J.T.; Chiang, M.C. An effective hybrid-heuristic algorithm for urban traffic light scheduling. Neural
Comput. Appl. 2021, 33, 17535–17549. [CrossRef]
29. Joo, H.; Ahmed, S.H.; Lim, Y. Traffic signal control for smart cities using reinforcement learning. Comput. Commun. 2020,
154, 324–330. [CrossRef]
30. Rahman, R.; Hasan, S. Real-time signal queue length prediction using long short-term memory neural network. Neural Comput.
Appl. 2021, 33, 3311–3324. [CrossRef]
31. Wang, J.; Lv, W.; Jiang, Y.; Qin, S.; Li, J. A multi-agent based cellular automata model for intersection traffic control simulation.
Phys. A Stat. Mech. Its Appl. 2021, 584, 126356. [CrossRef]
32. Zhao, H.; Han, G.; Niu, X. The signal control optimization of road intersections with slow traffic based on improved PSO. Mob.
Netw. Appl. 2020, 25, 623–631. [CrossRef]
33. Mao, T.; Mihăită, A.S.; Chen, F.; Vu, H.L. Boosted Genetic Algorithm using Machine Learning for traffic control optimization.
IEEE Trans. Intell. Transp. Syst. 2021. [CrossRef]
Processes 2021, 9, 2205 14 of 14

34. Khelafa, I.; Ballouk, A.; Baghdad, A. Control algorithm for the urban traffic using a realtime simulation. Int. J. Electr. Comput.
Eng. 2021, 11, 3934–3942. [CrossRef]
35. Mou, J. Intersection traffic control based on multi-objective optimization. IEEE Access 2020, 8, 61615–61620. [CrossRef]
36. Ali, M.; Devi, G.L.; Neelapu, R. Intelligent Traffic Signal Control System Using Machine Learning Techniques. In Microelectronics,
Electromagnetics and Telecommunications; Springer: Berlin, Germany, 2021; pp. 611–619.
37. Azimirad, E.; Pariz, N.; Sistani, M.B.N. A novel fuzzy model and control of single intersection at urban traffic network. IEEE Syst.
J. 2010, 4, 107–111. [CrossRef]
38. Park, I.S.; Park, C.e.; Kwon, N.K.; Park, P. Dynamic output-feedback control for singular interval-valued fuzzy systems: Linear
matrix inequality approach. Inf. Sci. 2021, 576, 393–406. [CrossRef]
39. Bernal, E.; Lagunes, M.L.; Castillo, O.; Soria, J.; Valdez, F. Optimization of type-2 fuzzy logic controller design using the GSO and
FA algorithms. Int. J. Fuzzy Syst. 2021, 23, 42–57. [CrossRef]
40. Kim, S.S.; Chung, M.; Kim, Y.K. Urban traffic prediction using congestion diffusion model. In Proceedings of the 2020 IEEE
International Conference on Consumer Electronics-Asia (ICCE-Asia), Seoul, Korea, 1–3 November 2020; pp. 1–4.
41. Tang, J.; Zeng, J.; Wang, Y.; Yuan, H.; Liu, F.; Huang, H. Traffic flow prediction on urban road network based on License Plate
Recognition data: Combining attention-LSTM with Genetic Algorithm. Transp. A Transp. Sci. 2021, 17, 1217–1243. [CrossRef]
42. Li, D.; Deng, L.; Cai, Z. Intelligent vehicle network system and smart city management based on genetic algorithms and image
perception. Mech. Syst. Signal Process. 2020, 141, 106623. [CrossRef]
43. Bethge, J.; Morabito, B.; Rewald, H.; Ahsan, A.; Sorgatz, S.; Findeisen, R. Modelling Human Driving Behavior for Constrained
Model Predictive Control in Mixed Traffic at Intersections. IFAC-PapersOnLine 2020, 53, 14356–14362. [CrossRef]
44. Jafari, S.; Shahbazi, Z.; Byun, Y.C. Improving the Performance of Single-Intersection Urban Traffic Networks Based on a Model
Predictive Controller. Sustainability 2021, 13, 5630. [CrossRef]