Abstract—This paper presents a fuzzy traffic controller for
controlling a multilane-multiple traffic intersection in urban
city environment. The developed controller applied to the
model is based on the cooperation and distributed mechanisms
with the neighbors’ intersection. The phase sequence and phase
length extension are used in decision making to provide
efficient and more optimal traffic control. The input variables
are normalized and the proposed fuzzy traffic controller is
suitable to be used to any case study without modification to
the controller modules. The simulation illustrates that the fuzzy
traffic controller proposed in this study show a remarkable
improvement over the existing fuzzy traffic controller without
input normalization.
I. INTRODUCTION
RAFFIC control in urban networks is a nonlinear
process varying dynamic characteristics throughout the
day. The fundamental principle of urban traffic control is to
respond to dynamic changes of the traffic demand and
synchronizing traffic signal with the neighbor intersections.
As the junctions in a road network are interconnected, the
decisions made at one junction inevitably affect those made
at the adjacent junctions.
In multilane-multiple intersection, the traffic signal
control operates dependently to neighbor intersection. This
means that the control algorithm can be much complicated
than algorithms for isolated intersection. The system
controlling the intersection has a lowest degree of freedom
to choose a control strategy for multilane-multiple
intersection. In this case, a control strategy is choosing must
be considering the traffic from upstream intersection and
congestion occur in the downstream intersection. Traffic
control system is nonlinear, fuzzy and nondeterministic, and
traditional methods of modeling and control can’t work very
well. As a perfect urban area traffic signals control system, it
is supposed to respond to traffic demand and online-
Manuscript received April 15, 2009. This work was supported in part by
the Universiti Putra Malaysia and CAIRO, Universiti Teknologi Malaysia.
Azura Che Soh, is currently doing PhD in Centre for Artificial
Intelligence and Robotics (CAIRO) of Universiti Teknologi Malaysia, Jalan
Semarak, 54100 Kuala Lumpur.(e-mail:azura@eng.upm.edu.my).
Marzuki Khalid, is a professor and Deputy Vice Chancellor (Research &
Innovation) of Universiti Teknologi Malaysia, Skudai, Johor.
(e-mail:marzuki@utmkl.utm.my).
Mohammad Hamiruce Marhaban is an associate professor with the
Department of Electrical and Electronics Engineering, Universiti Putra
Malaysia,43400, Serdang, Selangor.(e-mail:hamiruce@eng.upm.edu.my).
Rubiyah Yusuf, is a professor and director of CAIRO, Universiti
Teknologi Malaysia, Jalan Semarak, 54100 Kuala Lumpur.
(e-mail:rubiyah@utmkl.utm.my).
optimize timing plans in time, and then implement real-time
control. Thus, the traffic system must have adaptive
characteristic and automatic control strategy.
Fuzzy controllers have proven effective in controlling a
single traffic intersection [1]-[3], even when the intersection
is somewhat complex. In certain instances, however, even if
local controllers perform well, there is clearly no guarantee
that they will continue to do so when the intersections are
coupled with irregular traffic flow.
Research on controlling a set of traffic intersections can
be found in [4]-[6]. Jee-Hyong et. al. [4] applied the fuzzy
logic approach for dynamic multiple traffic intersection in
urban areas. The controller changes not only the phase
lengths but also the phase sequences adaptively based on the
traffic situation. References [5] and [6] were based on the
distributed fuzzy controller which is adaptive to the traffic
conditions. However, they dealt with simple traffic
conditions, such as, regularly located intersections. They did
not change the sequence of phases in their research.
Changing of the phases sequence may confuse drivers, but
in the view point of the performance of traffic controllers, it
is much sought. If the sequence of phases is fixed and traffic
conditions are often changed, the traffic signal cannot
effectively deal with the traffic conditions which may result
in deterioration of the of the traffic controllers.
In the case above [4]-[6], the controller could easily
handle changes in traffic flow, but required different
parameter settings for each junction. The fuzzy membership
functions of the input fuzzy variables of the controller are
different for each lane depending on the lengths of the
traffic. The ratio value for this input is required as the length
of the lane is different. Using the ratio value not only
decreases the complexity of the fuzzy logic system but also
improves computation time.
This paper presents the development of intelligent traffic
light system based on the fuzzy logic approach applied to
multiple-multilane intersection in an urban area. The
controller is developed based on the cooperation mechanism
and synchronizing mechanism with the neighboring
intersections. The phase sequence and phase length
extension are used in the decision making of the fuzzy
controller. In this controller, the modification of input
variables in the ratio values was purposely done to include
all the related variables to one specific variable.
Improving Fuzzy Traffic Controller for Multilane-Multiple
Intersection
Azura Che Soh, Marzuki Khalid, Mohammad Hamiruce Marhaban, Rubiyah Yusuf, Member, IEEE
T
2009 IEEE International Conference on Control and Automation
Christchurch, New Zealand, December 9-11, 2009
FrMT5.2
978-1-4244-4707-7/09/$25.00 ©2009 IEEE 1819