There are many intersections in the urban traffic systems. So it is important to effectively control traffic signals at intersections. In order to do many factors should be considered, such as traffic density, velocity, neighboring intersections and so on. But controlling traffic signals with such factors is not easy because of the complexity and dynamics of traffic patterns.
There is some research on traffic signal control based on fuzzy logic because fuzzy logic is adequate for qualitative modeling of complex systems. But most research focuses on simple traffic situations.
This paper presents a traffic fuzzy controller for a set of intersections, i.e. intersection group. In the developed system, each intersection has its own traffic controller which manages phase sequence and phase length dynamically according to its own and it's neighbor's traffic situations. The controller consists of three modules - the urgency module, the stop module and the decision module. The urgency module calculates the urgency degree of all possible phases and selects the phase having the maximum degree of urgency as the next phase. The urgency degree indicates how urgent a phase is. The output of the stop module is the degree to which the green phase can stop. The decision module takes the outputs of the urgency module and the stop module as inputs and decides whether to quit the green phase. To carry out the performance evaluation of the controller, a simulator for the intersection group has been developed.
The developed fuzzy traffic controller has been compared with two conventional methods: the vehicle-actuated and the vehicle-adjusted methods. The developed fuzzy traffic controller shows good performance in case where traffic patterns change over time and in uneven traffic situations.
