摘要
为提高智能交通信号在实际应用中的效率,提出一种改进的控制方法,将交通信号图像输入卷积神经网络的输入层,通过卷积层与采样层的卷积计算、残差计算以及梯度计算识别交通信号,将识别交通信号结果选取自适应跳跃式信号控制方法实现智能交通信号控制,用双层管道模型表示智能交通信号控制区域车辆流动情况,通过内外层管道路侧单元获取车辆通行信息,利用车辆权重之和及影响绿灯时间分配车辆数量实现智能交通信号控制。实验结果表明,文中方法可有效控制交通信号,不同时段车辆通过率均高于25%,获得了理想的交通信号控制结果。
In order to improve the efficiency of intelligent traffic signal in practical application,an impr-oved control method is presented.Intelligent traffic signal control is researched based on convolutional neural network input the traffic signal image to the input layer of the convolutional neural network,the traffic signal is identified through the convolution calculation,residual calculation and gradient calculation of the convolution layer and the sampling layer.The identified Traffic signal results select adaptive jump-type signal control method to achieve intelligent traffic signal control.A double-layer pipe model is used to represent vehicle flow in the area controlled by intelligent traffic signals.Vehicle traffic information is obtained through the inner and outer layer road side units.The sum of vehicle weight and affect the green light time allocation of the number of vehicles are used to achieve intelligent traffic signal control.The experimental results show that the method can effectively control traffic signals,and the vehicle passing rate in different periods is higher than 25%,which has better control effect.
作者
王丽君
史二娜
WANG Li-jun;SHI Er-na(Shaanxi Institute of International Trade&Commerce,Xianyang 712046,Shaanxi Province,China;Xi’an Traffic Engineering Institute,Xi’an 710300,China)
出处
《信息技术》
2020年第10期56-60,共5页
Information Technology
基金
陕西省教育厅专项科学研究(18JK1041)
物联网与智能技术科技创新团队建设项目(SSY18TD05)。
关键词
卷积神经网络
智能交通
信号控制
采样层
convolution neural network
intelligence traffic
signal control
sampling layer
