城市路口交通状态判别方法研究  

Identification Method of Traffic State in Urban Intersection

作　　者：任其亮[1] 王世能[1] 王坤[1] 詹家凤 曾柯[1] 

机构地区：[1]重庆交通大学交通运输学院,重庆400074

出　　处：《重庆交通大学学报（自然科学版）》2015年第6期111-115,122,共6页Journal of Chongqing Jiaotong University(Natural Science)

基　　金：重庆市教委自然科学研究项目(KJ130422);中国博士后科研基金(20110490809)

摘　　要：为客观、有效地评价城市路口交通状态,选择平均最大排队长度、饱和度、平均车辆延误、速度比等4个对城市路口交通状态变化最为敏感且容易获取的参数作为评价指标,选取Relief F算法对所选取的评价指标进行权值判定,设计了基于模糊FCM聚类的综合判别模型,并借助VISSIM交通仿真软件对路口交通流状态进行了仿真。结果表明:改进的W-FCM判别方法相比于传统的神经网络判别法,总体误判代价降低了47.5%,判别精度提升至96.7%,改进的判别方法具有较强的可行性。In order to objectively and effectively evaluate the traffic state of urban intersection, the parameters were chosen as the evaluation indexes, including the average maximum queue length, saturation, average vehicle delay and velocity ratio, which were the most sensitive evaluating indicators to the traffic state of urban interaction and also easy to be obtained. Relief F algorithm was selected to determine the weights of the selected evaluation indexes, and a comprehensive identification model based on fuzzy clustering FCM was designed. And then the traffic flow state at urban intersection was simulated by VISSIM traffic simulation software.. The results show that compared with the traditional neural network method, the total error cost of the improved W-FCM method is reduced by 47.5% and the accuracy is improved to 96.7%. The improved identification method has a strong feasibility.

关 键 词：交通运输工程 城市路口 交通状态 排队长度 W-FCM聚类算法 

分 类 号：U491[交通运输工程—交通运输规划与管理]