•  
  •  
 

Abstract

In order to identify highway accident-prone sections, this paper presented a method based on the K-Means clustering algorithm to identify accident-prone sections. According to the accident severity, road production loss and average casualty compensation were introduced as the evaluation indexes of equivalent accident number, and the traditional equivalent accident number was improved. According to the statistical distribution characteristics of the improved equivalent accident number, the section length was determined, and the accident-prone section was initially identified by the cumulative frequency method. The K-Means clustering algorithm was used to cluster the initially selected accident-prone sections, and the final accident-prone sections were obtained. In order to verify the correctness of the proposed method, the accident-prone sections from Hekou to Pingtai of the Guangzhou–Wuzhou Expressway were identified. The results show that compared with the traditional equivalent accident number, the improved equivalent accident number can better reflect the accident severity. The improved equivalent accident number follows the negative binomial distribution, and the objective section length can be obtained according to its statistical distribution characteristics. The selection result of the K-Means clustering algorithm is superior to the DBSCAN algorithm, and the total length of accident-prone sections screened by the K-means clustering algorithm accounts for 66.7% of the initial results. This method can provide a strong theoretical basis for the management of accident-prone sections of expressways.

Publication Date

1-18-2024

DOI

10.14048/j.issn.1671-2579.2022.06.049

First Page

260

Last Page

264

Submission Date

May 2025

Reference

[1] 孙丽璐, 陈甜, 赵娟, 等. 我国交通事故损失影响因素及地区特征分析: 基于全国31个省市自治区2004—2015年面板数据[J]. 西南大学学报(自然科学版), 2019, 41(8): 114-123. SUN Lilu, CHEN Tian, ZHAO Juan, et al. Analysis of the influencing factors of traffic accident losses and their regional features in China: a study based on the panel data of 31 provinces from 2004 to 2015[J]. Journal of Southwest University (Natural Science Edition), 2019, 41(8): 114-123. [2] 王颖志, 王立君. 基于网络时空核密度的交通事故多发点鉴别方法[J]. 地理科学, 2019, 39(8): 1238-1245.WANG Yingzhi, WANG Lijun. An identification method of traffic accident black point based on street-network spatial-temporal kernel density estimation[J]. Scientia Geographica Sinica, 2019, 39(8): 1238-1245. [3] 金霞, 雷桂荣, 刘峰, 等. 交通事故黑点鉴别与改善排序研究[J]. 公路与汽运, 2018(2): 45-48.JIN Xia, LEI Guirong, LIU Feng, et al. Study on identification and improvement of traffic accident black spots[J]. Highways & Automotive Applications, 2018(2): 45-48. [4] 孟祥海, 覃薇, 霍晓艳. 基于统计与假设检验的高速公路交通事故数据分布特性[J]. 交通运输工程学报, 2018, 18(1): 139-149. MENG Xianghai, QIN Wei, HUO Xiaoyan. Distribution characteristics of traffic crash data of freeway based on statistics and hypothesis test[J]. Journal of Traffic and Transportation Engineering, 2018, 18(1): 139-149. [5] 瞿庆亮, 曲国庆, 徐工. 车载GPS在公路交通事故多发路段判别中的应用[J]. 公路, 2016, 61(7): 224-228.QU Qingliang, QU Guoqing, XU Gong. Application of kinematic GPS in identification of road accident-prone sections[J]. Highway, 2016, 61(7): 224-228. [6] 张长生, 马荣国. 山区高速公路交通事故分析及多发路段鉴别[J]. 长安大学学报(自然科学版), 2010, 30(6): 76-80. ZHANG Changsheng, MA Rongguo. Identifying method of traffic accident black spots on mountain freeway[J]. Journal of Chang’an University (Natural Science Edition), 2010, 30(6): 76-80. [7] ELVIK R. Evaluations of road accident blackspot treatment: a case of the iron law of evaluation studies?[J]. Accident Analysis & Prevention, 1997, 29(2): 191-199. DOI:10.1016/S0001-4575(96)00070-X. [8] LORD D, WASHINGTON S P, IVAN J N. Poisson, Poisson-gamma and zero-inflated regression models of motor vehicle crashes: balancing statistical fit and theory[J]. Accident Analysis & Prevention, 2005, 37(1): 35-46. [9] WRIGHT C C, ABBESS C R, JARRETT D F. Estimating the regression-to-mean effect associated with road accident black spot treatment: Towards a more realistic approach[J]. Accident Analysis & Prevention, 1988, 20(3): 199-214. [10] 彦伟 ,宇仁德 ,孙连超.鉴别道路交通事故黑点的聚类分析方法研究[J] .交通标准化 ,2011(11) :110-113. [10] Yan Wei, Yu Rende, Sun Lianchao. Research on Clustering Analysis Methods for Identifying Road Traffic Accident Black Spots [J]. Transportation Standardization, 2011(11): 110–113. [11] 邵祖峰.交通事故黑点鉴别方法研究综述[J] .道路交通与安全 ,2008(2) :44-49. [11] Shao Zufeng. Review of Identification Methods for Traffic Accident Black Spots [J]. Road Traffic & Safety, 2008(2): 44–49. [12] 王钱. 基于改进累计频率曲线法的高速公路事故多发点鉴别研究[D]. 西安: 长安大学, 2018. WANG Qian. Study on identification of expressway accident-prone points based on improved cumulative frequency curve method[D]. Xi’an: Changan University, 2018. [13] 孟祥海, 刘振博. 基于Tobit回归的山区高速公路事故率分析模型[J]. 中外公路, 2020, 40(2): 294-299.MENG Xianghai, LIU Zhenbo. Crash rate analysis model of mountain freeway based on Tobit regression[J]. Journal of China & Foreign Highway, 2020, 40(2): 294-299. [14] 王龙健, 成嘉琪. 累计频率法鉴别道路事故多发点中单位取样长度的限定[J]. 公路与汽运, 2015(5): 36-39. WANG Longjian, CHENG Jiaqi. Limitation of unit sampling length in identification of road accident-prone points by cumulative frequency method [J]. Highways & Automotive Applications, 2015(5): 36-39. [15] 高曼,韩勇,陈戈,等.基于K-means聚类算法的公交行程速度计算模型[J]. 计算机科学, 2016, 43(S1): 422-424, 439.GAO Man, HAN Yong, CHEN Ge, et al. Bus travel speed calculation model based on K-means clustering algorithm [J]. Computer Science, 2016, 43(S1): 422-424, 439. [16] 邓国忠, 曹帆, 吴勇, 等. 互通式立交与隧道出口小间距路段事故影响因素分析[J]. 中外公路, 2019, 39(4): 283-287. DENG Guozhong, CAO Fan, WU Yong, et al. Analysis on accidents influence factors of the small spacing section between intersection and tunnel exit[J]. Journal of China & Foreign Highway, 2019, 39(4): 283-287. [17] CHANG L Y. Analysis of freeway accident frequencies: negative binomial regression versus artificial neural network[J]. Safety Science, 2005, 43(8): 541-557. [18] 耿超, 彭余华. 基于动态分段和DBSCAN算法的交通事故黑点路段鉴别方法[J]. 长安大学学报(自然科学版), 2018, 38(5): 131-138. GENG Chao, PENG Yuhua. Identification method of traffic accident black spots based on dynamic segmentation and DBSCAN algorithm[J]. Journal of Chang’an University (Natural Science Edition), 2018, 38(5): 131-138.

Download

Share

COinS