Fuzzy neural network controllers and simulations for urban region traffic

This paper uses the fuzzy neural networks (FNN) theory to solve the real time region traffic distributed control problem. The region traffic is regarded as a large scale system and the subsystems are the intersections in the region. Each intersection has its own traffic controller which manages the green phase and the phase length dynamically according to its own and its neighbor´s traffic situations. The object of the controller is to make the lane unblocked and the average vehicle delay time shortest in the front of each intersection. The proposed method is compared with the vehicle actuated method which is one of the type conventional methods. The simulation show good performances in the cases of time-varying traffic patterns and heavy traffic situations.