考虑交通流非线性特性的交通信号迭代学习控制策略

An Iterative Learning Control Strategy for Traffic Signals Considering Nonlinear Characteristics of Traffic Flow

现实中城市交通流的运行具有很强的非线性特性,采用简单的线性模型难以全面描述交通流的实际运行过程.本文在考虑城市交通流非线性动态特性的基础上,提出了一种非线性交通流排队模型,并基于宏观交通流固有的周期性特征,设计了交叉口信号的迭代学习控制策略.通过对交叉口信号的迭代学习控制,使交叉口各进口道的车辆排队长度尽可能趋于均衡,提高交叉口信号有效绿灯时间的利用率,从而提高路网的通行效率.最后通过严格的数学推导证明了该方法的收敛性,仿真研究及实验结果验证了所提方法的有效性.

The operation of urban traffic flow has strong non-linear characteristics in real traffic situation,so it is difficult to describe the actual operation process of traffic flow comprehensively by using simple linear model.Considering the nonlinear dynamic characteristics of urban traffic flow,a nonlinear traffic flow queuing model is proposed,and an iterative learning control strategy for intersection signals is designed based on the inherent periodic characteristics of macroscopic traffic flow.By the iterative learning control of intersection signals,the vehicle queue lengths in the import way are balanced as far as possible,thus the utilization rate of effective green light time at intersections and the efficiency of road network are effectively improved.Finally,the convergence of the method is proved by strict mathematical deduction,and the effectiveness of the proposed method is verified by the simulation research and experimental results.