基于Schatten p范数的城市快速路异常交通状态估计

Abnormal Traffic State Estimation of Urban Expressway Based on Schatten P-norm

道路车辆速度是交通状态判别中最为重要的评价指标。对于一些GPS数据量稀疏的快速路路段,车辆临时停车、走应急车道等行为会出现浮动车数据反映的交通速度与实际路段车速不符的情况,从而无法对路段交通状态进行准确的判断。针对这类数据稀疏路段,为了能够准确地判断路段交通状态,以出租车数据为研究对象,提出了一个基于浮动车数据的判别异常交通状态算法。该算法利用GPS轨迹数据,基于地图匹配技术,将车辆运动轨迹与道路几何特征相结合,生成时空速度矩阵。使用Schatten p范数矩阵解决数据的局限性,在数据可用率较低的情况下减小速度估计误差。最后以北京市中关村地区出租车GPS数据为研究对象进行方法验证。通过与均值法、奇异值分解算法进行对比分析,结果表明,该算法在连续数据缺少的情况下仍有良好的性能,提高路网交通状态识别精度。

Road vehicle speed is the most important evaluation index in traffic state identification.For some expressways with sparse GPS data,the traffic speed reflected by the floating vehicle data is inconsistent with the actual speed of the road section due to the temporary parking and emergency lane of vehicles,which makes it impossible to accurately judge the traffic status of the road section.For this kind of data sparse road,in order to accurately judge the traffic status of the road,this paper takes taxi data as the research object,and proposes a traffic state identification algorithm based on floating car data.Based on the GPS trajectory data and map matching technology,the algorithm combines the vehicle trajectory and road geometric features to generate the space-time velocity matrix.Schatten p norm matrix is used to solve the limitation of data and reduce the speed estimation error when the data availability is low.Finally,this paper takes the taxi GPS data in Zhongguancun area of Beijing as the research object to verify the method.