Traffic Control Sub-area Dynamically Divide And Signal Coordination Optimization Control

With the development of economic, the problem of traffic congestion increasingly serious. At present, the most effective and feasible solution is making the existing road network to play out maximum utility though traffic management. The dynamically divide of traffic control sub- area and the traffic signal coordination optimization control research helps to raise the level of traffic management.This paper studied the methods of dynamic divide traffic control sub- area and traffic signal coordination optimization control. Among them, the dynamically divide of traffic control sub- area research on the basis of predecessors' research, this paper proposed based on artificial neural network traffic control sub-area dynamically divide. Artificial neural network is self-learning, distributed storage, parallel processing etc, adapt to the classification, pattern recognition, automatic control, predictive and other fields, this paper using artificial neural network forecast coordinate coefficient between adjacent intersections, and then divide sub-area with coordinate coefficient. The principles of traffic control sub-area division using cycle principle, flow principle and distance principle, BP neural network and fuzzy neural network is applied in the traffic control sub-area dynamic division, show the specific design of BP neural network and fuzzy neural network, used MATLAB for simulation experiments, the results of experiments have verified the feasibility of the two methods, fuzzy neural network method have faster convergence speed and smaller errors than BP neural network method. Based on artificial neural network method divide the traffic control sub-area dosen't need build complicated traffic system model, can better adapt to the traffic system of real-time feature.This thesis studied traffic signal coordination optimization control on the basis of traffic control sub-area dynamically division. Using multi-agent and Dyna-Q reinforcement learning algorithm realize signal coordination control between sub-areas, and simulated in traffic simulation software synchro6.0, compared the single control, the coordinated control only divide sub-area, the coordinated control between sub-areas, the experiment results show that the coordinated control between sub-areas can effectively reduce system delay time, realized the effective combination between the traffic signal coordination control and sub-area dynamically division.