基于AFC数据的城市轨道交通乘客目的站实时预测

Passenger Destination Real-time Forecast of Urban Rail Transit Based on AFC Data

为解决轨道交通只有在乘客刷卡离开路网,且获得实际列车运行图后,才可通过清分模型系统仿真推演其全出行链的问题,在实时接入自动售检票系统(AFC)刷卡数据情况下,建立基于乘客出行OD规律、乘客职住地规律及重点去向车站规律的级联目的站预测模型,为进站乘客快速预测目的站,进而与清分模型系统结合实现进站乘客在网分布的实时动态推演。通过开展乘客目的站预测,提高客流实时仿真推演系统对各项客流指标的预测准确度,不仅方便运营管理人员进行科学高效的网络化运营调度指挥及客运管理,还可为乘客出行提供个性化引导。

The whole travel chain of rail transit can be simulated and deduced through the clearing model system only after passengers swipe their cards while leaving the railway network and the actual train working diagram is complete.In order to eliminate the lag,this study establishes a cascade destination prediction model based on the passengers’travel regularities,destination distribution law and key destination station under the condition of real-time access to AFC card swiping data,so as to quickly predict the destination station for passengers.In combination with the clearing model system,the real-time dynamic deduction of the passenger distribution on the network is realized.The system is not only convenient for operation management personnel to carry out scientific and efficient networked operation,dispatching command and passenger transport management,but also can provide personalized guidance for passenger travel.