Fuzzy Neural Network Realized by ANFIS and Its Application to the Traffic Signal Timing Optimization

A parameter-optimization algorithm is proposed to optimize parameters of traditional fuzzy controllers with ANFIS(adaptive neuro-fuzzy inference system),a fuzzy neural network toolkit in Matlab.This algorithm improves the shape and distribution of membership functions of the controllers,and is applied to multi-phase traffic signal timing for the urban isolated intersection.All simulation results prove that the proposed algorithm can reduce average vehicle delay time and the vehicles can pass through intersections more smoothly.