基于小波变换和频谱分析的交叉口群路径分级方法  

Routes classification method at intersections group using wavelet transform and spectrum analysis

作　　者：李岩[1] 过秀成[1] 杨洁[1] 何赏璐[1] 刘迎[1] 

机构地区：[1]东南大学交通学院,南京210096

出　　处：《东南大学学报（自然科学版）》2012年第1期168-172,共5页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金资助项目(50422283);建设部软科学研究资助项目(2008-K5-14);江苏省普通高校研究生科研创新计划资助项目(CX10B_072Z)

摘　　要：为确定交叉口群的关键路径,建立了基于小波变换和频谱分析的路径分级模型.根据交叉口群交通关联性强的特点,通过小波变换对交通检测数据的降噪,以突显其短时变化特性;计算上下游交叉口进口流向交叉谱的一致性和位相,确定上下游流向的关联性和统计滞后关系,得出流向的特征向量,并应用模糊识别方法将路径分级,确定交叉口群的关键路径.采用南京市广州路交叉口群实测交通数据验证模型的有效性.结果表明,模型可实现交叉口群的路径分级,并识别其关键路径,为交叉口群交通信号协调控制奠定基础.A route classification method for intersection group using wavelet transform and spectrum analysis is proposed.According to the strong transportation relationship of the intersection group,the de-noise function of wavelet domain is applied to extract the short time pattern of each movement.Then the magnitude squared coherence and phase of the cross spectrum calculated by two entry movements between upstream and downstream intersections at one route are adopted to procure the eigenvectors of the route.Fuzzy cluster is then used to classify all the routes into several clusters by the eigenvectors,which makes the critical route easily acquired.Results of using the proposed model to an intersections group at Guangzhou Road in Nanjing show that the routes can be divided into three clusters,which match the field observation.The route classified by the proposed method can be used for optimization of the traffic signal control for related intersections group with satisfying effectiveness,robustness and reliability.

关 键 词：路径分级 小波变换 频谱分析 交叉口群 数据挖掘 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]