基于GPS数据的出租车路径选择行为研究

Taxi Route Choice Behavior Modeling Based on GPS Data

出租车路径选择行为研究对于剖析城市交通流时空分布规律及交通流分配具有重要意义.基于北京市出租车调度系统采集的GPS(Global Positioning System)数据,本文首先通过旅程数据提取、地图匹配及双向扫除算法,构建车辆路径选择行为分析所需的RP(Revealed Preference)数据.其次,在综合考虑道路网络条件和交通条件对驾驶员路径选择行为影响的基础上,构建基于PSL(Path-Size Logit)模型的出租车多路径概率选择模型.案例分析结果表明,出租车驾驶员更倾向于选择旅行时间短、转向次数少、主干路和次干路比例较高的路径.同时,文中所提出的路径选择行为建模方法,对于提高GPS数据的使用价值具有积极的意义.

Taxi route choice behavior analysis is of great significance to study urban traffic flow＇s spatiotemporal distribution and assignment. First, based on the GPS data collected from taxi dispatch system, RP data for taxi route choice behavior analysis are obtained by using trip-identification method, map matching and double-sweep algorithm. Second, by considering the comprehensive impacts of road network condition and traffic status on taxi drivers＇ route choice behavior, a PSL model based vehicle route choice model is developed. Finally, the results of case study show that taxi drivers incline to choose the route with less travel time, lower frequency of the left turns and right turns, and higher total ratio of arterial road and sub-arterial road. Meanwhile, the modeling approach of route choice behavior is of positive significance to improve the use value of GPS data.