基于主成分贡献度的道路事故热点成因分析
|
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。
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.
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.
引文
[1]Yu Hao,Liu Pan,Chen Jun,et al.Comparative Analysis of the Spatial Analysis Methods for Hotspot Identification[J].Accident Analysis&Prevention,2014,66:80-88.
[2]薛大维,纪峻岭,白竹.基于主成分分析法的高速公路交通安全评价[J].黑龙江工程学院学报(自然科学版),2014(2):46-49.Xue Dawei,Ji Junling,Bai Zhu.Study on Traffic Safety Evaluation for Freeway Based on Principal Component Analysis[J].Journal of Heilongjiang Institute of Technology,2014(2):46-49.
[3]Yu Rende,Zhang Qiang,Zhang Xiaohong,et al.Traffic Accidents Forecasting Based on Neural Network and Principal Component Analysis[J].Research Journal of Applied Sciences Engineering&Technology,2013,6(6):1065-1073.
[4]Li Qiangwei.Evaluation Model of Region Traffic Safety Based on Principal Component Analysis[R/OL].2009[2016-07-10].https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnum ber=5168451.
[5]Han Qingwen,Zhu Yingxiang,Zeng Lingqiu,et al.A Road Hotspots Identification Method Based on Natural Nearest Neighbor Clustering[R/OL].2015[2016-07-10].https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7313189.
[6]黄金龙.基于自然最近邻的无参聚类算法研究[D].重庆:重庆大学,2014.Huang Jinlong.Study on Non-Parametric Clustering Based on Natural Nearest Neighborhood[D].Chongqing:Chongqing University,2014.
[2]薛大维,纪峻岭,白竹.基于主成分分析法的高速公路交通安全评价[J].黑龙江工程学院学报(自然科学版),2014(2):46-49.Xue Dawei,Ji Junling,Bai Zhu.Study on Traffic Safety Evaluation for Freeway Based on Principal Component Analysis[J].Journal of Heilongjiang Institute of Technology,2014(2):46-49.
[3]Yu Rende,Zhang Qiang,Zhang Xiaohong,et al.Traffic Accidents Forecasting Based on Neural Network and Principal Component Analysis[J].Research Journal of Applied Sciences Engineering&Technology,2013,6(6):1065-1073.
[4]Li Qiangwei.Evaluation Model of Region Traffic Safety Based on Principal Component Analysis[R/OL].2009[2016-07-10].https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnum ber=5168451.
[5]Han Qingwen,Zhu Yingxiang,Zeng Lingqiu,et al.A Road Hotspots Identification Method Based on Natural Nearest Neighbor Clustering[R/OL].2015[2016-07-10].https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7313189.
[6]黄金龙.基于自然最近邻的无参聚类算法研究[D].重庆:重庆大学,2014.Huang Jinlong.Study on Non-Parametric Clustering Based on Natural Nearest Neighborhood[D].Chongqing:Chongqing University,2014.