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

Journal of Chongqing Jiaotong University(Natural Science) ›› 2019, Vol. 38 ›› Issue (05): 1-7.DOI: 10.3969/j.issn.1674-0696.2019.05.01

• Transport+Big Data and Artificial Intelligence •     Next Articles

Highway Abnormal State Detection Based on Support Vector Machine (SVM)

SUN Jingyi1, MOU Ruojin2, SU Xiaobo3   

  1. (1. Faculty of Transportation Engineering, Kunming University of Science and Technology, Kunming 650500, Yunnan P. R. China; 2. Hefei Urban Rail Transit Co. Ltd., Hefei 230041, Anhui, P. R. China; 3. Anhui Huadian Engineering Consultating & Design Co. Ltd., Hefei 230022, Anhui, P. R. China)
  • Received:2017-11-19 Revised:2017-12-20 Online:2019-05-15 Published:2019-05-15

基于支持向量机的高速公路异常状态检测

孙静怡1,牟若瑾2,苏晓波3   

  1. (1.昆明理工大学 交通工程学院,云南 昆明 650500; 2. 合肥城市轨道交通有限公司,安徽 合肥 230041; 3. 安徽华电工程咨询设计有限公司,安徽 合肥 230022)
  • 作者简介:孙静怡(1968—),女,湖南益阳人,副教授,博士,主要从事综合交通系统优化与管理方面的研究。E-mail:jysun@kmust.edu.cn。 通信作者:牟若瑾(1993—),女,辽宁丹东人,硕士,主要从事高速公路管理方面的研究。E-mail:568427938@qq.com。

Abstract: Aiming at the abnormal state caused by expressway accidents, the detection program of expressway abnormal states on the basis of support vector machine (SVM) was proposed. Firstly, the vehicle breakdown accident model at one-way three-lane basic section of Kun-Yu Expressway was established by VISSIM simulation software and the abnormal events under different conditions were simulated; furthermore, the influence of abnormal states on the traffic capacity of highway section was analyzed. Then, based on SVM model, the abnormal state detection model of expressway was established, and the difference between unilateral and bilateral input was analyzed. The results of the simulation demonstrate that: the established SVM model for detection of abnormal states has good comprehensive performance; with unilateral input, it can also provide satisfactory detection results of the occurrence of abnormal states.

Key words: traffic engineering, highway, event detection, VISSIM simulation, support vector machine

摘要: 针对意外事件所引发的高速公路异常状态,提出了基于支持向量机模型的高速公路异常状态检测方案。首先应用VISSIM仿真软件建立了昆玉高速单向三车道基本路段的车辆抛锚事故模型;对不同条件下的异常事件进行了仿真,并分析了异常状态对高速断面通行能力的影响;然后基于支持向量机模型建立了高速公路异常状态检测模型,分析单侧与双侧输入差异。仿真结果表明:所建立的支持向量机异常状态检测模型综合性能较好,单侧输入也可以较好地检测异常状态的发生。

关键词: 交通工程, 高速公路, 事件检测, VISSIM仿真, 支持向量机

CLC Number: