非均衡交通流五岔路口交通信号智能控制研究-《郑州大学学报(工学版)》

文章信息/Info

Title:: Research on Intelligent Control of Traffic Signal for Five-road Intersection of Unbalanced Traffic Flow

Author(s):: Dong Hailong; Tang Minan; Cheng Haipeng; School of Automation and Electrical Engineering, Lanzhou Jiaotong University, Lanzhou, Gansu, 730070

Keywords:: five-road intersection; traffic signalfuzzy neural network; artificial fish swarm algorithm; average delay of vehicles

摘要:: 为了缓解交通拥堵,减少车辆在交叉口的等待时间,针对五岔路口非均衡交通流,采用基于人工鱼群算法的交通信号控制方法,实现了五岔路口多相位变相序的智能控制.该方法首先由红灯相位的车辆排队长度和红灯持续时间得到该红灯相位的通行紧急度,把通行紧急度最高的相位作为下一绿灯相位;其次,将当前绿灯相位车辆排队长度和筛选出的下一绿灯相位车辆排队长度与当前绿灯相位车辆排队长度的差值作为输入,利用模糊神经网络控制器实现绿灯延时.为了避免模糊神经网络陷入局部最小值,利用人工鱼群算法对模糊神经网络参数进行优化.在不同的车辆到达率情况下进行仿真研究,结果表明:该方法比传统的控制方法在自动调节信号周期方面效果更好,减少车辆平均延误7.2％左右.

Abstract:: In order to alleviate traffic congestion and reduce the time of vehicles waiting,of unbalaned intersection,this paper adopted a traffic signal control method based on artificial fish swarm algorithm,which could achieve multi-phase and variable phase sequence intelligent control at five-road intersection.Firstly,the passable urgency of a red light phase was decided according to its vehicle queuing length and the red light duration;and the highest passable urgency phase should be the next green light phase.Secondly,the current vehicle queuing length of green light phase and the difference between vehicle queuing length of the next and current green light phase were taken as the inputs.The fuzzy neural network controller was used to control delays of the green light.In order to avoid fuzzy neural network falling into local minimum,artificial fish swarm algorithm had optimized the parameters of it.After simulation studies in the case of different rates of vehicle arrival,the results showed that this method was better than the traditional control in automatically adjusting the signal cycle,which reduced the average delay of vehicles for about 7.2％.