An Adaptive Two-Stage Fuzzy Controller for Traffic Signals: Architecture, Algorithms, and Simulation

This paper presents the design and evaluation of a two-stage fuzzy logic traffic signal controller with online optimization for an isolated intersection. The controller is designed to be adaptive to real-time traffic demands and has the following two critical features: (1) designing a two-stage fuzzy logic controller. The fuzzy controller uses strategic and tactical vehicle loop detectors, placed respectively upstream and at the stop-line of the intersection on each approach, to measure approach arrival flows and estimate queues. This data is used to decide whether to extend or terminate the current signal phase. These decisions are made using the proposed two-stage fuzzy logic procedure. (2) Developing an adaptive optimization framework of fuzzy controller to adjust fuzzy membership functions and controller rules with on-line learning. This study models the performance index of average delays based on traffic status identification, and at regular time intervals, employs a hybrid genetic algorithm to efficiently yield the reliable solution through reappearance of statistical traffic flow. The performance of this controller is compared to those of fixed-time, actuated, classic fuzzy and two-stage fuzzy controller for different traffic conditions.