13-07-2012, 12:28 PM
Design and Simulation of Adaptive Fuzzy Control on the Traffic Network
[attachment=27535]
INTRODUCTION
Transportation with newly-developed technologies
has become a part of modem human life. However, the
traffic jam also becomes more serious and the traffic
signal control is regarded as a key technology to solve it.
Many researchers choose the fuzzy logic to replace
manual operation in the rush hour [5, 7-12]. Since they
usually require too many fuzzy rules, the calculations
become more complex. Therefore, a similar result was
expected using the simplified fuzzy rules from the
simulation data form the fixed-time control.
Basically, traffic systems are time-varying with a
random nature, there are many parameters need to be
setup for the simulation software. In the present study,
the software VISSIM which is developed by University
of Karlsruhe and PTV System Software and Consulting
AG is used to build the traffic networks. The VISSIM
offers a wide variety of urban and highway applications
and it integrates public and private transportation. Even
complex traffic conditions are visualized in an
unprecedented level of detail providing realistic traffic
models. In the present study, the software VISSIM is
modified to construct a simulation environment to
analyze the proposed adaptive fuzzy approach for the
signal control logic.
SIMPLIFIED FUZZY RULES
In order to design intelligent controllers which can be
functioned similarly as a traffic police, three input
parameters are adopted in the present simulation study
as: CarsBehindRed, CarsBehindGreen and Cycle Time
[5]. These parameter setting is quite suitable and simple
for real traffic simulation but its fuzzy rules are quite
complex. In order to achieve the purpose of intelligent
control, neural fuzzy control [13-14] was proposed to
train the system in self-learning procedures to produce a
fuzzy rule base. In this paper, the adaptive fuzzy
controller with four membership functions for each of
the input and output fuzzy variable of the system, as
shown in Figs. 1-4, is proposed with less CPU
calculation and delay time to obtain the traffic flow
control. Therefore, the waiting length varied with the
time is used to simplify the fuzzy rules.
THE ADAPTIVE FUZZY CONTROL
VISSIM is used in this study to simulate general fuzzy
controllers. When the traffic flow of both directions of
east-west (E-W) and south-north (S-N) are in the rush
hour, the fuzzy controller will adopt the method of
changing signals frequently to reduce the traffic
congestion. After the light changes from read to green, it
may maintain at least 7 seconds then it is allowed to
change to red according to the fuzzy control results. For
an extreme case, if the light changes to green only for 7
seconds and it must switch back to the red according to
the fuzzy control rules, only a few of the front vehicles
will pass through the green lights and most of the back
vehicles will still blocked after the red light. Thus, the
traffic congestion would spread to its neighbors and will
result in the traffic jams.
To deal with the traffic jam during the rush hours,
fuzzy controllers are commonly operated with more
frequently signal switching; however, it is not practical
for real applications. The increasing weighting of the
threshold value for changing signals in the present fuzzy
controller is adopted when the Cycle Starting Time is
within 5 15 seconds to overcome the quick-switching
problem. For example, when the Cycle Starting Time is
within 5 15 seconds, the light will change if the
defuzzified value is larger than 0.63.
IMPLEMENT THE TRAFFIC NETWORK SIMULATION BY VISSIM
VISSIM is a powerful simulation software in traffic
analysis. By setting the provided user interface like the
vehicle types, the turning ratio, and the traffic
composition, etc., the traffic network can be thus
established. The VISSIM simulation environment is
pretty similar to the real traffic conditions in detail.
In order to compare with the fixed-time mode which
is used in most the general traffic intersections, we
employ the VisVAP to design the adaptive signal control
logic. It allows users to define their own signal control
logic in conjunction with VISSIM.
[attachment=27535]
INTRODUCTION
Transportation with newly-developed technologies
has become a part of modem human life. However, the
traffic jam also becomes more serious and the traffic
signal control is regarded as a key technology to solve it.
Many researchers choose the fuzzy logic to replace
manual operation in the rush hour [5, 7-12]. Since they
usually require too many fuzzy rules, the calculations
become more complex. Therefore, a similar result was
expected using the simplified fuzzy rules from the
simulation data form the fixed-time control.
Basically, traffic systems are time-varying with a
random nature, there are many parameters need to be
setup for the simulation software. In the present study,
the software VISSIM which is developed by University
of Karlsruhe and PTV System Software and Consulting
AG is used to build the traffic networks. The VISSIM
offers a wide variety of urban and highway applications
and it integrates public and private transportation. Even
complex traffic conditions are visualized in an
unprecedented level of detail providing realistic traffic
models. In the present study, the software VISSIM is
modified to construct a simulation environment to
analyze the proposed adaptive fuzzy approach for the
signal control logic.
SIMPLIFIED FUZZY RULES
In order to design intelligent controllers which can be
functioned similarly as a traffic police, three input
parameters are adopted in the present simulation study
as: CarsBehindRed, CarsBehindGreen and Cycle Time
[5]. These parameter setting is quite suitable and simple
for real traffic simulation but its fuzzy rules are quite
complex. In order to achieve the purpose of intelligent
control, neural fuzzy control [13-14] was proposed to
train the system in self-learning procedures to produce a
fuzzy rule base. In this paper, the adaptive fuzzy
controller with four membership functions for each of
the input and output fuzzy variable of the system, as
shown in Figs. 1-4, is proposed with less CPU
calculation and delay time to obtain the traffic flow
control. Therefore, the waiting length varied with the
time is used to simplify the fuzzy rules.
THE ADAPTIVE FUZZY CONTROL
VISSIM is used in this study to simulate general fuzzy
controllers. When the traffic flow of both directions of
east-west (E-W) and south-north (S-N) are in the rush
hour, the fuzzy controller will adopt the method of
changing signals frequently to reduce the traffic
congestion. After the light changes from read to green, it
may maintain at least 7 seconds then it is allowed to
change to red according to the fuzzy control results. For
an extreme case, if the light changes to green only for 7
seconds and it must switch back to the red according to
the fuzzy control rules, only a few of the front vehicles
will pass through the green lights and most of the back
vehicles will still blocked after the red light. Thus, the
traffic congestion would spread to its neighbors and will
result in the traffic jams.
To deal with the traffic jam during the rush hours,
fuzzy controllers are commonly operated with more
frequently signal switching; however, it is not practical
for real applications. The increasing weighting of the
threshold value for changing signals in the present fuzzy
controller is adopted when the Cycle Starting Time is
within 5 15 seconds to overcome the quick-switching
problem. For example, when the Cycle Starting Time is
within 5 15 seconds, the light will change if the
defuzzified value is larger than 0.63.
IMPLEMENT THE TRAFFIC NETWORK SIMULATION BY VISSIM
VISSIM is a powerful simulation software in traffic
analysis. By setting the provided user interface like the
vehicle types, the turning ratio, and the traffic
composition, etc., the traffic network can be thus
established. The VISSIM simulation environment is
pretty similar to the real traffic conditions in detail.
In order to compare with the fixed-time mode which
is used in most the general traffic intersections, we
employ the VisVAP to design the adaptive signal control
logic. It allows users to define their own signal control
logic in conjunction with VISSIM.