过刊检索
年份
《城市交通》杂志
2018年 第3期
基于主成分贡献度的道路事故热点成因分析
点击量:1328

文章编号: 1672-5328(2018)03-0046-06

曾令秋1,王瑞梅1,韩庆文2,曾孜3,朱颖祥2,张程1
(1.重庆大学计算机学院,重庆400044;2.重庆大学通信工程学院,重庆400044;3.重庆市第七中学校,重庆 400030)

摘要: 事故热点成因分析是解决交通安全问题的关键和难点,不同成因的事故热点直接决定驾驶人 或自动驾驶的通过策略。针对DTH3N算法聚类得到的事故热点展开研究,提出一种基于主成分贡 献度的道路事故热点成因分析方法。通过对5 类事故热点成因因素:道路、行人、车辆、环境和管 制因素贡献度的分析,提取事故热点道路物理成因和社会成因的影响权值。基于英国STATs19 数据 库选取5 个测试区域进行实验分析,结果表明,由事故热点成因分析方法获取的输出参数能够合理 解释事故热点成因,可有效指导驾驶行为决策和优化交通管制。

关键词: 交通安全;道路事故热点;主成分分析;内特性分析;道路物理成因;社会成因

中图分类号: U491.3

文献标识码:A

Traffic Accident Causation Analysis at Hotspots

Zeng Lingqiu1,Wang Ruimei1, Han Qingwen2, Zeng Zi3, Zhu Yingxiang2, Zhang Cheng1
(1.College of Computer Science, Chongqing University, Chongqing 400044, China; 2.College of Communication Engineering, Chongqing University, Chongqing 400044, China; 3.Chongqing No.7 High School, Chongqing 400030, China)

Abstract: The key to and difficulty in solving traffic safety problems lie in the causes of accident at hotspots. The accident hotspots caused by different factors determine the operation strategies for drivers and automatic vehicles. Through researching the accident hotspots with the clustering DTH3N algorithm, this paper proposes an analytical method to identify the causes of road accident hotspots based on main accident contributing factors. By analyzing the contributions of five types of accident causation factors at hotspots, namely road, pedestrian, vehicle, environment, and regulation, the paper calculates the influencing weights of both road physical and social factors for accident hotspots. The paper conducts an experimental analysis on five selected testing areas from the British STATs19 database. The results show that the output parameters obtained from the analytical method on accident hotspot serve as a reasonable explanation for the causes of accident hotspots, and can provide effective guidance in driving behavior decision-making as well as in optimizing traffic control.

Keywords: traffic safety; road accident hotspot; principal component analysis; internal characteristic analysis; road physical factors; social factors