Intelligent Real-time Optimization Control On Traffic Light

In recent years, the process of urbanization and society economy has been developed at full speed. Under the influence, the urban traffic demand and the motor vehicles have been increased rapidly; traffic congestion has been increased rapidly; air and noise pollution have been presented ascending trend; the vicious traffic accidents happened frequently. Traffic problems have been plaguing the development of many large cities and traffic jam has becoming the "bottleneck" of urban economic and social development. Building convenient, safe, efficient, low-pollution, conducive to the ecological, environmental protection and diversified urban public transport system has becoming a new urban transportation goal. Improving the intersection signals control strategy and easing traffic congestion problem stare in the face.View of the city intersection signal control issues, paper centering on how to achieve rational and effective control of traffic signals at junctions, achieve an in-depth and meticulous research on intersection traffic signal control theory and the strategy.First, the thesis has analyzed the vehicle rate of flow relationship of two adjacent intersections. Consider that with the characteristics of nonlinear and uncertain of the traffic system, the paper applies the general regression neural network to forecast the traffic volume. GRNN comprehensive consideration of the impact of every sample book over overall sample book space, both not lose a sample book burning, and may reduce the impact of strange data over forecast result. Against the shortcoming of uncertain function effect center of GRNN, use the adjacent clustering method to identified centers, to cut down arithmetic amounts, and improve the forecast speed; In addition, to determine the important parameter of GRNN, apply simple and effective method of the steepest descent, compare with the enumeration method has prior to more intelligence. There make neural network forecast good results, forecast speed, and provide a new method to forecast the vehicle flow.On the basis of the forecast of the traffic flow, the flow forecast is introduced to the control of traffic intersections lights, made the traffic lights control based on the vehicle forecast, enhance the initiative of signal control. Control strategy is built on the order of the variable phase, namely by the current traffic situation from the phase rotation sequence, the length of delay of the green light control by it at the same time. Thereby it will enhance the intersection vehicles passing through rate, and reduce the vehicles delay time. On this basis, consider the new idea that the speed signals suggestion will be apply at the entrance to the intersection. It's a very good practical, and serviceable in actual traffic system.