交叉口多相位神经网络智能控制  

The intelligent control of multi-phase intersectionand its artificial neural network

作　　者：常玉林[1] 张鹏[1] 宫镇[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2004年第6期485-487,共3页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(50078015);国家博士后基金资助项目

摘　　要：针对城市交叉口交通流的特点和普遍存在的拥挤问题,以F-B配时理论为基础,提出了一种基于神经网络的交叉口多相位智能控制,并以镇江市大市口为例,在调查了大量的交通流数据的基础上,描述了建立网络的方法,给出了学习算法,仿真结果表明,该神经网络控制器具有良好的学习和泛化能力,实时性良好.In light of the characteristics of the traffic flow and problems of congestion widely existed (in urban) intersections, and based on the theory of F-B time allocation, a multi-phase intersection intelligent controlling calculation is presented. Based on the traffic flow data in Dashikou of Zhenjiang, the (artificial) neural network is built, and the study calculation method is set up. The simulation result shows that the introduced artificial neural network has the capability of study and generalization with real time (feature.)

关 键 词：人工神经网络 BP网络 交叉口 实时控制 

分 类 号：U491.4[交通运输工程—交通运输规划与管理]