Multi-Layer Traffic Signal Control Model Based on Fuzzy Control and Genetic Algorithm

Abstract

This paper proposes a multi-layer traffic signal fuzzy control model based on genetic algorithm (GA) for isolated intersections. The model includes three layers. The first layer is traffic demand prediction. By using a comprehensive index, traffic demand intensities (TDI), the paper estimates the traffic demand on each approach lane during green time. The second layer, called phase sequence fuzzy controller, is implemented to optimize signal phases according to traffic flow conditions. The third layer is green time fuzzy controller. TDI and the phase sequence are used to determine whether the current signal phase will be extended or terminated. In this research, genetic algorithm will be adopted to determine the membership function of this multi-layer fuzzy control model. The performance of this control model will be compared to the model without membership function optimization at a simulated four-approach intersection, which will show the control model presented outperforms the traditional model in reducing average total delay at an intersection.