The Research On Transit Priority Adaptive Signal Control Based On Sparse Vehicle Detection And Fuzzy Neural Network

With the social economy having achieved rapid development and the people’s living standard having experienced a substantial increasing period, China’s car ownership is increasing year after year. However, the urban traffic congestion and environmental pollution are getting worse and worse. The further development of China’s economy and society relies more and more on road traffic system’s development. As a result, the reasonable and standardized control of road traffic must be taken seriously. However, traditional timing control strategy doesn’t meet satisfying control effect any more, what’s worse, it’s turning extremely difficult to build accurate mathematical models to implement traffic control because of urban road traffic’s randomness, complexity and uncertainty. For the sake of solving this problem and mostly improving the operational efficiency of urban traffic system, to solve the traffic congestion by intelligent traffic control strategy developed by high-techs (take adaptive control strategy for example) has become the most effective and economic method.As the global and fundamental supporting platform of urban traffic system, a sound and efficient public traffic system is the essential need of sustainable development of today’s urban transport. The urban pubic traffic is the first choice of solving the urban traffic congestion problem. An intersection is the concentration zone of urban traffic congestion and the implement of transit priority signal control into signal intersections is an important way of achieving public traffic’s prier development.From systematic thinking point of view, this paper designed three subsystems:transit identification and location subsystem, transit traveling time prediction subsystem and transit priority signal timing subsystem, then formulated their constitute and corresponding mathematical models and intelligent algorithms, and finally selected travel delay as the traffic indicator to evaluate the whole system’s improvement effect.In transit identification and location subsystem, we achieved empirical stopping area by particle filter, obtained transit buses in the mixed fleet by sparse detection algorithm, and acquired the distance from the detected bus to the stopping line in imported lanes of a signal intersection. In transit traveling time prediction subsystem, by putting the values of traveling time and prediction variances predicted by historical and adaptive forecasting models into adaptive-historical forecasting model, we could reach the prediction objective:the traveling time of a detected transit bus reaching to the stopping line of the imported lane of a signal intersection. In transit priority signal timing subsystem, the trigger premise and boundary conditions of transit buses’prior travel were given and its realizing method is fuzzy neural network control algorithm.This paper took a typical cross-shaped intersection in Chengdu city as the case. Firstly, we obtained the basic traffic parameters for simulating use by field survey and data processing; Then designed the fuzzy values of corresponding traffic parameters and performed programming design of fuzzy neural network algorithm to verify the applicability of the implement of fuzzy neural network into transit priority adaptive control; On the basis of simulation and programming design in the former two steps, this paper finally acquired the improving effect of transit priority adaptive signal control strategy compared to traditional timing signal control strategy by contrasting the chosen traffic capacity evaluating indicators.