Rolling horizon Optimization based Three stage Fuzzy Logic Controller for Urban Traffic Signals

Abstract:

A rolling horizon optimization based three stage fuzzy controller for urban traffic signals is presented. This controller introduces overlap phases of conflict free approach flows to develop the model of a three stage fuzzy controller. Aiming at online learning of the parameters of fuzzy membership functions and fuzzy rules, a rolling horizon framework is used to efficiently explore the effectiveness of fuzzy controllers under different traffic conditions. In this framework, according to the real time observed traffic data, a genetic algorithm (GA) based heuristic integrated with golden ratio (GRGA) has been employed to yield reliable solutions. Experiments are conducted on a typical blackspot isolated intersection via Paramics based online simulation under daily and abnormal traffic scenarios. Extensive simulation results demonstrate the potential of the developed controllers for adaptive traffic signal control with rapid response to the uncertainty of time variant traffic flow at intersections.