交叉路口的综合智能信号控制方案研究

针对城市道路交叉路口的交通信号控制,构造了一种能实时调整控制结构和参数的综合智能信号控制方案.方案中的结构评价单元根据每一相位结束时的交通状况,决定是否在标准相位的基础上进行控制结构调整.参数评价单元则根据评价周期内的信号控制效果,利用所获得的交通流数据对神经网络信号控制器进行"滚动"训练.两个神经网络总是交替处于学习和工作状态.仿真结果表明,该控制方案能很好地综合"结构调节"和"参数调节"的优点,因而能更好地适应路口的实际交通状况,从而达到有效提高路口通行能力的目的.

Study on the Approach of Composite Intelligent Intersection Traffic Signal Control

A composite method of intelligent intersection traffic signal control is introduced, whose structures and parameters can change online according to the traffic conditions. The structure-evaluation unit determines whether the control structure is changed by the end of each traffic phase. And the parameter-evaluation unit can give rolling training to neural network controllers with the traffic flow data obtained during evaluation cycles according to the effects of controlling. The two neural networks arc always alternatively in the states of learning or working during the process of self-learning according to the decision of the parameter-evaluation unit on the traffic conditions of intersection. Simulation results reveal that the traffic signal control can integrate the advantages of both structure-regulation and parameter-regulation, the proposed approach can better fit the actual traffic conditions than old ones, and thus has achieved the goal of improving the traffic capacity of intersections.