主干路交通信号控制仿真研究

Study on Traffic Signal Control Simulation on Main Road

传统的线控大多是以单个交叉路口或两个交叉路口作为最小的研究单元,采取某些控制算法来进行信号灯的实时控制。提出了“大路口”的概念,给出了“大路口”相位的划分,利用三个相邻路口之间的相关性,建立了以整个“大路口”车辆平均延误最小为目标的优化模型,给出了目标函数及其约束条件,采取遗传算法进行优化,获得“大路口”各个相位的绿信比;利用高阶广义神经网络预测相邻两交叉路口之间的交通流状况,计算出相邻交叉路口之间的相位差和第二相位时间。通过计算机仿真研究,此算法有效地协调三个路口的信号灯,使“大路口”支路方向交通量基本维持不变,而主干路方向的交通量则有明显改善。

Study on traditional traffic signal control almost emphasized on the isolated intersection or two intersections as one basic unit in which the signal was controlled by some algorithms. The concept of ＂big unit＂ was developed which consists of three intersections and their phase partition was also proposed. The optimal model was built in order to minimize the vehicles average delay through using the relativity between adjacent intersections in the ＂big unit＂. Then the objective function and its constraints were obtained and it was optimized by GA so that the splits of all phases in ＂big unit＂ were got. The states of the traffic flows between every two intersections were pre-estimated through using the high order general neural networks. The offsets between every two intersections and the time length of the second phase were figured out also. Through computer simulation, the signals in ＂big unit＂ were controlled effectively as well as the traffic flows in main road direction were imoroved obviously, while the traffic flows in the secondary road were not changed at all.