A pedestrian crowd classification method based on the improved ant colony clustering algorithm
Crowds are an important feature of high-dense Mass Rail Transit (MRT), assessing its crowding status is a critical step in crowd management.In this paper, a pedestrian crowd classification method based on an improved ant colony clustering algorithm (ACCA) is developed for MRT systems.Firstly, survey data from Automatic Fare Collection (AFC) regarding three statuses (check-in/check-out and sum) as well as the detailed geometric features of the metro station sites were gathered.Secondly, pedestrian crowd classification method was proposed to descript the safety level in the MRT, which based on the pedestrian crowding index (PCI).Moreover, the PCI influence factors were also considered in the method, which included average daily ridership intensity, the duration of crowd, and the scope of crowd influence.Thirdly, to classify the pedestrian crowd, an improved ant colony clustering model and its solving algorithm were presented.The results show that, for the two types of time scale, the passengers time-space characteristics present a clear image of M, the variation trend of morning and evening peak hour is obvious in the MRT.The research results can be used to provide a technical support for assessing pedestrian crowd status in the MRT systems.
Mass rail transit Pedestrian crowd characteristic Pedestrian crowding index Ant colony clustering algorithm Classification
Jibiao Zhou Pengfei Zhao Yongrui Chen Sheng Dong Uwe Plank-Wiedenbeck
School of Civil and Transportation Engineering / Ningbo University of Technology, No.201, FenghuaRao Beijing Key Laboratory of Traffic Engineering / Beijing University of Technology, No.100, Pingleyuan School of Highway / Changan University, Middle Section of NanErhuan Road, Xian, China Faculty of Civil Engineering / Bauhaus-University, Weimar Marienstrasse 13 C, Weimar, Germany
国际会议
The 8th International Conference on Pedestrian and Evacuation Dynamics (第八届行人与疏散动力学国际学术会议)
合肥
英文
597-603
2016-10-17（万方平台首次上网日期，不代表论文的发表时间）