全网络神经模糊控制在城市单路口交通实时控制中的应用

在已有城市单路口交通模糊控制方式和控制策略的基础上，提出了基于全网络化结构的神经模糊控制方法。方法考虑了影响信号灯控制策略的各种因素，根据分级并行控制思路，对车流采用不同的优先级和不同的控制策略进行协调控制，提高了系统的实时性，降低了系统的复杂性。采用6层全网络结构的神经网络进行了控制算法的实现，并利用已有数据对神经网络进行了学习训练，使网络结构和参数具有更为广泛的适用性。

Application of Neural Fuzzy Real Time Control of Full Network for Urban Traffic in Single Intersection

Based on fuzzy control method and strategy for urban traffic single intersection, full network structuring neural fuzzy control method is proposed. With consideration of various factors influencing the control strategy of traffic signal lamps in intersection, According to the idea of stepped parallel control the different priority level and strategy are applied to coordinate traffic control. The method have effectively reduced the complexity of system and raised real time operation. The control algorithm is provided by full network structuring of six level neural units and the neural network is trained by learning program of history data to make network structure and parameters more extensive applicability. The method has combined the advantage of fuzzy control and neural network technology, more effectively used fuzzy rule of expert knowledge storehouse and traffic data to mine association rule. Full analysis of the computer simulation show the method is theoretically innovative and feasible. Then the factors of influencing control results are discussed and some researches to be taken in the future are proposed.