Traffic signal control based on genetic neural network algorithm

Urban traffic signal control system is very complex, so it is very difficult to built a precise mathematical model. This paper presents a control algorithm which is alterable in phase-cycle and based on back propagation neural network method. After considering the lengths of each phase motorcade, this method determine how much time the current phase of the green light to extend and change the length of phase cycle. Meanwhile, the convergence rate of network is improved by using genetic algorithm to optimize network weights and threshold. Simulation results demonstrate that this algorithm can reduce the average junction waiting time and total waiting queue length effectively. The average delay of vehicles can be decreased in the application of this algorithm.