交通运输系统工程与信息 ›› 2023, Vol. 23 ›› Issue (6): 307-318.DOI: 10.16097/j.cnki.1009-6744.2023.06.030

• 工程应用与案例分析 • 上一篇    

基于智能刷卡数据的乘客上车站点估计研究

高万晨,路世昌*,李丹   

  1. 辽宁工程技术大学,工商管理学院,辽宁 葫芦岛 125000
  • 收稿日期:2023-08-21 修回日期:2023-09-13 接受日期:2023-09-14 出版日期:2023-12-25 发布日期:2023-12-23
  • 作者简介:高万晨(1989- ),男,辽宁大连人,博士生。
  • 基金资助:
    辽宁省教育厅人文社会科学研究项目 (21-A817)。

Estimation Algorithms of Passenger Boarding Stops Based on Smart Card Data

GAO Wan-chen,LU Shi-chang*,LI Dan   

  1. School of Business Administration, Liaoning Technical University, Huludao 125000, Liaoning, China
  • Received:2023-08-21 Revised:2023-09-13 Accepted:2023-09-14 Online:2023-12-25 Published:2023-12-23
  • Supported by:
     Humanities and Social Sciences Research Project of the Department of Education of Liaoning Province (21-A817)。

摘要: 针对城市公交自动收费系统中缺少乘客上车站点的问题,本文首先设计两阶段、改进K近邻和改进模糊C均值聚类这3种估计算法,并将估计结果与传统的时间窗算法进行对比;其次,采用熵率方法确定不同算法估计乘客上车站点的准确率;最后,以珠海市18路公交的智能刷卡数据为例,验证所提出算法的有效性。研究结果表明,3种算法均能实现所有乘客上车站点的全部匹配,与传统的时间窗算法相比,匹配率高约36.3%。就3个维度样本数据的平均熵率而言,乘客上车站点估计的准确率从高到低分别为两阶段算法,改进K近邻算法,改进模糊C均值聚类算法;两阶段算法与改进 K 近邻算法准确率相差不大,选择熵率最小的算法确定乘客最终的上车站点。本文研究方法可以应用于城市公交系统。

关键词: 城市交通, 上车站点估计, 算法, 智能刷卡数据, 熵率

Abstract: This paper design three estimation algorithms: the two-stage algorithm, the improved K-nearest neighbor algorithm, and the improved fuzzy C-means clustering algorithm, to address the issue of missing passenger boarding stops in urban public transit automatic fare collection systems. We compare the estimation results of these algorithms with the traditional time window algorithm, and evaluate their accuracy using the entropy rate method. The proposed algorithms are validated using smart card data from bus No. 18 in Zhuhai City. The results indicate that all three algorithms can successfully match all passenger boarding stops, and the matching rate is approximately 36.3% higher than that of the traditional time window algorithm. The accuracy of the passenger boarding stop estimation, based on the average entropy rate of the three- dimensional sample data, is ranked as follows: the two- stage algorithm, improved K-nearest-neighbor algorithm, and improved fuzzy C-means clustering algorithm. The difference in accuracy between the two-stage algorithm and the improved K-nearest neighbor algorithm is minor. The algorithm with the lowest entropy rate is selected to determine the final boarding stops of passengers, making it suitable for implementation in urban public transit system.

Key words: urban traffic, boarding stop estimation, algorithm, smart card data, entropy rate

中图分类号: