基于模糊神经网络的信号交叉口交通量预测  

Traffic Flow Prediction of Signalized Intersections Using Fuzzy_Neural Network Approach

作　　者：尹宏宾[1] 徐建闽[1] 黄仕进[2] 

机构地区：[1]华南理工大学交通学院,广东广州510640 [2]香港大学

出　　处：《华南理工大学学报（自然科学版）》2000年第6期11-15,共5页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目! (6 99740 1 6 );广东省自然科学基金资助项目! (990 5 76 )

摘　　要：预测精度高 ,实时性强的交叉口交通量预测算法可以极大地提高城市交通控制的效率 .文中提出了基于模糊神经网络技术的信号交叉口交通量预测方法 .该方法以模糊神经网络为核心 ,应用在线滚动学习模型实现交叉口交通量预测 ,并应用了交通量微观仿真系统对模型进行检验 .仿真结果表明该模型比传统方法精度高、收敛速度快 .High accuracy on_line algorithms of traffic flow prediction can effectively improve performance of the urban traffic adaptive control system. In this paper the on_line rolling prediction model of traffic flow at signalized intersections based on fuzzy logic system and neural network is proposed. The method makes use of on_line rolling learn characteristics of fuzzy_neural network. Therefore, no complex computation is involved in this model process and it can also learn constantly from historical data. The simulation results have proved the effectiveness of the model, which has gread potential for application to the urban traffic control system.

关 键 词：交叉口 交通量预测 模糊神经网络 城市交通控制 

分 类 号：U491.113[交通运输工程—交通运输规划与管理] U491.14[交通运输工程—道路与铁道工程]