城市轨道交通短时客流预测体系框架及关键技术

Framework and Key Technologies of Short Term Passenger Flow Forecast of Urban Rail Transit

针对当前城市轨道交通短时客流预测系统性不强等问题,构建短时客流预测体系框架,并讨论预测过程涉及的关键技术。框架的构建以自动售检票系统(AFC)获得的数据为出发点,统计站点客流和线网客流OD矩阵两类基础客流数据;在此基础上,构建线网客流分配模型,结合视频数据、站点平面布置和列车运行时刻表三类数据,考虑乘客步行时间的影响,估计断面客流数据;接着,在分析站点客流和断面客流数据时空特性的基础上,分别预测站点和断面短时客流;利用站点客流和断面客流短时预测结果反推未来OD矩阵;同时引入GARCH模型分析预测结果的可靠性,以提高短时客流预测结果的可信度。

In view of the systemic deficiencies in current research of short term passenger flow forecast of urban rail transit, a forecast framework of short term passenger flow was constructed. Based on the collected data from the automatic fare collection （AFC） systems,the passenger flow at the stations and the network 01） matrix were collected. On this basis, an assignment model of passenger flow for the network, combined with three types of the data sources, that is ,video data, station layout and train timetable, and considering lhe impact, of the walking activities of passengers, was huilt, and the sectional passenger flow was estimated. Then, based on the analysis of the spacial and temporal characteristics of the passenger flows at station and section, the two passenger flows were forecasted respectively; Afterwards, the network passenger flowOD matrix was backstepped by the short term forecasting results of the passenger flows. Meanwhile, the （;ARCtt model was introduced to analyze and to improve the forecast reliability.