中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊
重庆交通大学学报（自然科学版）

重庆交通大学学报（自然科学版） ›› 2015, Vol. 34 ›› Issue (1): 100-104.DOI: 10.3969/j.issn.1674-0696.2015.01.22

• • 上一篇    下一篇

基于向前搜索和K均值聚类的信号交叉口行人群集分类方法研究

于泉，王萌 ，邓小惠   

  1. 北京工业大学 交通研究中心，北京 100124
  • 收稿日期:2013-08-20 修回日期:2013-09-02 出版日期:2015-02-15 发布日期:2015-03-10
  • 作者简介:于泉(1976—)，男，山东海阳人，副教授，工学博士，主要从事交通信息与控制方面的研究。E-mail：yuquan@bjut.edu.cn。
  • 基金资助:
    国家自然科学基金项目（51108010）

Pedestrian Swarming Classification Method Based on Forward Searching and K-Means Clustering Algorithm at Signalized Intersection

Yu Quan, Wang Meng, Deng Xiaohui   

  1. Transportation Research Center, Beijing University of Technology, Beijing 100124, China
  • Received:2013-08-20 Revised:2013-09-02 Online:2015-02-15 Published:2015-03-10

PDF

摘要： 以信号交叉口行人群体为研究对象，借鉴生物界群集研究的成果，给出了行人群集的定义，同时提出群集半径的概念和向前搜索的群集半径计算方法。运用K均值聚类方法对北京市信 号交叉口60 h视频资料的737个行人群集样本进行分析，得到描述行人群集的特征参数数值。研究结果表明：行人群集半径能够较好地评价我国城市信号交叉口人行横道的服务水平，可以为 信号交叉口大量行人过街的控制算法和模型的研究奠定基础。

关键词: 交通工程, 信号交叉口, 行人, 群集, 群集半径

Abstract: Pedestrian groups at signalized intersection were the researched object, the achievements of living things were used for reference, and the conception of pedestrian swarming was given. At the same time the conception of pedestrian swarming circus was defined. Also a method based on forward searching was provided. K- means clustering algorithm was used to analyze 737 pedestrian swarming samplings of 60 hours video data at the signalized intersection, and the values of feature parameters of different pedestrian swarming were obtained. The research result shows that pedestrian swarming circus can evaluate the level of service of pedestrian crossing at signalized intersection in China, and can make good foundation for control algorithm and model of pedestrian crossing.

Key words: traffic engineering, signalized intersection, pedestrian, swarming, swarming circus

中图分类号: