基于动态阈值的交通流故障数据实时识别方法  

Real-time identification of traffic erroneous data based on dynamic threshold

作　　者：陆化普[1] 孙智源[1] 屈闻聪 

机构地区：[1]清华大学,北京100084

出　　处：《土木工程学报》2015年第11期126-132,共7页China Civil Engineering Journal

基　　金："十二五"国家科技支撑计划(2014BAG01B04-03);北京市科技计划(Z121100000312101);清华大学苏州汽车研究院(吴江)返校经费课题(2015WJ-B-02)

摘　　要：为了提高交通流数据的准确性和可靠性,提出了动态阈值模型,用以实现交通流故障数据的实时识别。以北京市道路交通网络为研究区域,考虑研究对象路段的交通流实时数据与历史数据、上下游相关路段交通流数据的相关性,以及交通流参数之间的相关性,将传统的交通阈值理论推广到动态阈值模型。面向实时的城市道路交通流检测,基于动态阈值模型,计算交通流参数的实时阈值,以识别交通流故障数据。实例研究表明:该文的动态阈值模型可快速、有效的识别出交通流故障数据。A model of dynamic threshold is developed to identify traffic erroneous data timely,so that the accuracy and reliability of traffic data can be improved. With Beijing traffic network taken as research area,the traditional traffic threshold theory is extended to the dynamic threshold model. The relationship between real-time data and historical data of an investigated objective road section,the relationship between its upstream and downstream of the investigated objective road section,and the relationship between different traffic flow parameters are considered in the proposed model. To identify traffic erroneous data,the threshold of real-time traffic data is calculated based on the dynamic threshold model. The case study shows that real-time traffic erroneous data can be effectively identified based the proposed method.

关 键 词：交通工程 动态交通流数据 故障数据 动态阈值 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]