交通流在间歇流状态下的概率变点模型  

A probability change-point model for traffic flow in the intermittent condition

作　　者：夏媛媛[1] 马立云[2] 王晓原[2] 李真[1] 

机构地区：[1]青岛科技大学人事处,山东青岛266061 [2]山东理工大学交通与车辆工程学院智能交通研究所,山东淄博255091

出　　处：《山东理工大学学报（自然科学版）》2011年第4期12-16,共5页Journal of Shandong University of Technology:Natural Science Edition

基　　金：国家自然科学基金资助项目(61074140)

摘　　要：以往研究交通流突变点的算法,大部分忽略了对交通流关键参数量变模式的辨识.依据交通流理论,结合非参数变点统计方法,以在淄博市华光路采集的间歇流数据为例,建立了对交通流模式量变识别的负二项分布概率变点模型,对交通流变点存在与否进行假设检验,并对变点发生(时间)位置的搜索算法作了详细研究,最后结合实际数据对累计次数法进行了标定和检验,验证了本方法的有效性和可行性.For the former algorithms about the change points of traffic flow,recognition of quantitative change of key parameters of traffic flow is mostly neglected.Based on traffic flow theory and nonparametric statistical method of change-point,a nonlinear probability change-points model used to identify quantitative change was firstly established with intermittent flow of Huaguang Road of Zibo city as an example.The hypothesis-testing problem of change-point and the algorithm used to search the change-point were discussed.Finally,the method is calibrated and tested with the field data,which verify the validity and the feasibility of the theory.

关 键 词：交通流 模式识别 累计次数法 统计变点理论 

分 类 号：U491[交通运输工程—交通运输规划与管理]