基于IC卡历史数据的公交客流统计方法

Approach on Public Traffic Passenger Flow Statistics Based on Historical IC Data

随着公交IC卡的普遍使用,基于公交IC卡数据的公交客流统计方法逐步成为主流的统计方法。因国内绝大部分城市均采用上车刷卡的方式,所以需根据用户出行规律进行下车站点的推测。为降低运算复杂程度,现有的IC卡分析算法多采用基于后续公交站点吸引权的概率算法,这种算法既不能判断单个乘客的下车站点,对于总体客流又存在数据准确性问题。为此,本文对于能够形成出行闭环的单日多次出行采用传统方法推测其下车地点;对于未形成闭环的单日出行,则分析相关单个乘客历史类似天日的IC卡刷卡数据,统计出其最可能的下车站点,进而得出总体客流。采用本算法对重庆市某段时间的所有IC卡数据进行处理和统计后,经人工计数的数据对比表明:相较于传统的出行闭环算法和站点吸引权算法,本算法对于公交客流的统计更加准确。

With popularity of public traffic IC cards, passenger flow statistics tends to be based on analysis of IC card data. Most Chinese cities only let passengers swipe cards while they get on buses, so we need estimate their get-off bus stations according to their trip rules. To cut down computational complexity, most of current IC card analysis algorithms are probabilistic algorithms, which fail to deduce each passenger＇s get-off bus station in addition to their relatively low data accuracy. Thus, for passengers with trip-closed loop, we apply traditional passenger trip rule to obtain their get-off bus stations; otherwise, we analyze their historical trip data in similar days so as to predict the most probable get-off bus stations. By above method, we can get the total passenger flow of a city, Verified by monthly IC card data of Chongqing, the algorithm can get more accurate passenger flow compared with independent traditional trip-closed loop one and probabilistic ones.