留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于交通流稳定性系数的高速公路交通事故实时风险预测

刘星良 单珏 刘唐志 饶畅 刘通

刘星良, 单珏, 刘唐志, 饶畅, 刘通. 基于交通流稳定性系数的高速公路交通事故实时风险预测[J]. 交通信息与安全, 2022, 40(4): 71-81. doi: 10.3963/j.jssn.1674-4861.2022.04.008
引用本文: 刘星良, 单珏, 刘唐志, 饶畅, 刘通. 基于交通流稳定性系数的高速公路交通事故实时风险预测[J]. 交通信息与安全, 2022, 40(4): 71-81. doi: 10.3963/j.jssn.1674-4861.2022.04.008
LIU Xingliang, SHAN Jue, LIU Tangzhi, RAO Chang, LIU Tong. Real-time Forecast Models for Traffic Accidents on Expressways Using Stability Coefficients of Traffic Flow[J]. Journal of Transport Information and Safety, 2022, 40(4): 71-81. doi: 10.3963/j.jssn.1674-4861.2022.04.008
Citation: LIU Xingliang, SHAN Jue, LIU Tangzhi, RAO Chang, LIU Tong. Real-time Forecast Models for Traffic Accidents on Expressways Using Stability Coefficients of Traffic Flow[J]. Journal of Transport Information and Safety, 2022, 40(4): 71-81. doi: 10.3963/j.jssn.1674-4861.2022.04.008

基于交通流稳定性系数的高速公路交通事故实时风险预测

doi: 10.3963/j.jssn.1674-4861.2022.04.008
基金项目: 

国家自然科学基金项目 52172341

重庆市教育委员会项目 KJQN202100718

重庆市高校创新研究群体项目 CXQT21022

详细信息
    作者简介:

    刘星良(1989—),博士,讲师. 研究方向:交通流理论、交通安全等. E-mail:xingliang1125@outlook.com

    通讯作者:

    刘唐志(1976—),博士,教授. 研究方向:山区道路交通安全、智慧交通、应急救援等. E-mail:391873717@qq.com

  • 中图分类号: U491.3

Real-time Forecast Models for Traffic Accidents on Expressways Using Stability Coefficients of Traffic Flow

  • 摘要: 预测交通事故实时风险时,存在大量指标变量,导致数据难以采集,不仅不利于构建预测模型,且带来的过拟合问题会降低模型预测可靠性。为了减少预测指标数量,提升预测模型可用性,降低预测模型过拟合影响,构建具有可解释性的2种交通流稳定性系数以简化指标集,分别为纵向交通流稳定系数和横向交通流稳定系数。采集西安市G3001高速公路交通事故与交通流历史数据,选用支持向量机、随机森林、Logistic回归模型,分别构建高速公路交通事故实时风险预测模型。通过改进的GI指数评估交通流稳定性系数的显著性,以检验其有效性;通过指标集在训练与测试数据中的预测精度、AUC值差异评估交通流稳定性系数对降低预测模型过拟合的作用,并通过训练耗时评估模型的计算效率,以检验新方法的可靠性。研究结果表明:2种交通流稳定性系数对应的改进GI指数分别为0.952和0.922,显著大于其他受试指标,与交通事故实时风险显著相关。在3种预测模型中,包含2种交通流稳定性系数的简化指标集在训练和测试数据中的预测精度分别为91.1%和90.5%,与完整指标集相近。2种指标集在训练与测试数据中的平均预测精度差异分别为0.69%和4.87%;平均AUC值差异分别为1.61%和5.87%;平均训练时间下降了15.2%。交通流稳定性系数大幅提高了预测模型的可靠性,同时显著提升了模型的计算效率。

     

  • 图  1  G3001道路基本线形与交通流监控系统布置

    Figure  1.  Layout of G3001 basic alignment and traffic flow monitoring system

    图  2  2015—2019年G3001交通事故数

    Figure  2.  Number of G3001 traffic incidents in 2015—2019

    图  3  各测试组预测精度

    Figure  3.  Forecast Accuracy of Each Tested Group

    图  4  各测试组AUC

    Figure  4.  AUC in Each Tested Group

    表  1  高速公路交通事故实时风险预测指标集

    Table  1.   Predictor set of expressway traffic accidents real-time risk forecast

    主要类别 次级类别 基础指标 基础指标代码
    交通状态 交通量(VC 上游平均交通量/(veh/h) VCup
    上游交通量标准差/(veh/h) Std. VCup
    上游相邻车道间平均交通量差值/(veh/h) Dif. VCup
    下游平均交通量/(veh/h) VCdo
    下游交通量标准差/(veh/h) Std. VCdo
    下游相邻车道间平均交通量差值/(veh/h) Dif. VCdo
    上下游平均交通量差值/(veh/h) Dif. VCup - do
    占有率(OCC 上游平均占有率/% OCCup
    上游占有率标准差/% Std. OCCup
    上游相邻车道间平均占用率差值/% Dif. OCCup
    下游平均占有率/% OCCdo
    下游占有率标准差/% Std. OCCdo
    下游相邻车道间平均占用率差值/% Dif. OCCdo
    上下游平均占有率差值/% Dif. OCCup - do
    速度(S 上游平均速度/(km/h) Sup
    上游速度标准差/(km/h) Std. Sup
    上游相邻车道间平均速度差值/(km/h) Dif. Sup
    下游平均速度/(km/h) Sdo
    下游速度标准差/(km/h) Std. Sdo
    下游相邻车道间平均速度差值/(km/h) Dif. Sdo
    上下游平均速度差值/(km/h) Dif. Sup - do
    道路几何线形 主线 路段长度/m SL
    车道数/条 NL
    路面宽度/m RSW
    车道宽度/m LW
    内侧路肩宽度/m ISW
    外侧路肩宽度/m OSW
    分隔带宽度/ m MW
    匝道 合流区占路段总长的比例/% MA
    分流区占路段总长的比例/% DA
    分合流区间距/m DMR
    环境 天气情况 WC
    时间 TD
    咼峰时间段 PP
    限速(km/h) VL
    下载: 导出CSV

    表  2  基于交通流稳定性系数的高速公路交通事故实时风险预测简化指标集

    Table  2.   Traffic flow stability coefficients based simplified predictor set of expressway traffic accidents real-time prediction

    基础指标 指标代码
    交通流纵向稳定性系数 Dif.DEup - do
    交通流横向稳定性系数 Dif. DEdo
    重车混人率/% PT
    合流区占路段总长的比例/% MA
    分流区占路段总长的比例/% DA
    天气情况 WC
    下载: 导出CSV

    表  3  各路段中受试指标的改进GI指数

    Table  3.   Improved Gini index of tested predictors in each road section

    改进GI指数指标代码 路段编码
    1-2# 2-3# 3-4# 4-5# 5-6# 6-7# 7-8# 8-9# 9-10# 10-11# 11-12# 12-13# 13-14# 14-1# 均值 标准差
    Dif.DEup-do 0.968 0.934 0.967 0.948 0.976 0.978 0.947 0.941 0.956 0.934 0.945 0.970 0.938 0.932 0.952 0.016
    Dif. DEdo 0.932 0.887 0.956 0.939 0.919 0.931 0.883 0.935 0.938 0.921 0.885 0.957 0.895 0.930 0.922 0.025
    Sdo 0.886 0.838 0.907 0.890 0.894 0.845 0.877 0.867 0.905 0.899 0.872 0.846 0.834 0.887 0.875 0.025
    OCCdo 0.791 0.756 0.805 0.820 0.825 0.770 0.761 0.781 0.775 0.821 0.791 0.798 0.823 0.786 0.793 0.023
    Sup 0.724 0.678 0.749 0.708 0.752 0.688 0.683 0.672 0.703 0.685 0.680 0.770 0.761 0.690 0.710 0.034
    DA 0.685 0.638 0.741 0.640 0.642 0.737 0.683 0.668 0.725 0.688 0.679 0.640 0.723 0.730 0.687 0.039
    MA 0.667 0.625 0.621 0.642 0.691 0.638 0.652 0.668 0.651 0.653 0.683 0.633 0.680 0.674 0.655 0.022
    WC 0.653 0.614 0.640 0.655 0.652 0.667 0.643 0.629 0.655 0.635 0.639 0.626 0.661 0.625 0.643 0.016
    PT 0.619 0.585 0.602 0.588 0.631 0.589 0.622 0.584 0.635 0.598 0.643 0.611 0.635 0.599 0.610 0.021
    VCdo 0.598 0.565 0.563 0.620 0.615 0.571 0.580 0.579 0.591 0.612 0.588 0.597 0.613 0.593 0.592 0.019
    OCCup 0.553 0.523 0.558 0.549 0.578 0.543 0.536 0.549 0.571 0.547 0.564 0.561 0.555 0.557 0.553 0.014
    SL 0.514 0.455 0.502 0.518 0.509 0.530 0.497 0.519 0.492 0.480 0.474 0.537 0.504 0.510 0.503 0.022
    VCup 0.419 0.383 0.381 0.429 0.418 0.435 0.391 0.417 0.456 0.390 0.402 0.413 0.443 0.428 0.415 0.023
    下载: 导出CSV

    表  4  各测试组训练耗时情况

    Table  4.   Train time in each tested group

    路段编号 SVM RF LR
    简化 完整 简化 完整 简化 完整
    1-2# 2.45 3.05 2.35 2.83 2.27 2.58
    2-3# 2.51 3.03 2.40 2.85 2.25 2.62
    3-4# 2.62 3.05 2.37 2.91 2.25 2.60
    4-5# 2.70 3.15 2.44 2.97 2.35 2.71
    5-6# 2.68 3.16 2.46 2.97 2.40 2.69
    6-7# 2.73 3.15 2.46 2.99 2.39 2.70
    7-8# 2.73 3.15 2.44 2.97 2.39 2.70
    8-9# 2.70 3.19 2.48 3.05 2.37 2.68
    9-10# 2.67 3.19 2.44 3.01 2.41 2.71
    10-11# 2.73 3.21 2.43 2.97 2.35 2.75
    11-12# 2.65 3.07 2.38 2.92 2.17 2.58
    12-13# 2.63 3.05 2.38 2.93 2.20 2.60
    13-14# 2.63 3.05 2.40 2.94 2.25 2.55
    14-1# 2.64 3.08 2.41 2.94 2.29 2.59
    平均值训练耗时/s 2.65 3.11 2.42 2.95 2.31 2.65
    耗时差异/% 14.79 17.97 12.83
    下载: 导出CSV
  • [1] MITCHELL T M. Machine learning: A guide to current research[M]. Boston: Springer, 2011.
    [2] 游锦明, 方守恩, 张兰芳, 等. 高速公路实时事故风险研判模型及可移植性[J]. 同济大学学报(自然科学版), 2019, 47(3): 347-352. https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ201903007.htm

    YOU J M, FANG S E, ZHANG L F, et al. Real-time crash prediction models and transferability analysis on freeways[J]. Journal of Tongji University (Natural Science), 2019, 47 (3): 347-352. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-TJDZ201903007.htm
    [3] CHEN F, CHEN S, MA X. Analysis of hourly crash likelihood using unbalanced panel data mixed logit model and real-time driving environmental big data[J]. Journal of Safety Research, 2018, 65: 153-159. doi: 10.1016/j.jsr.2018.02.010
    [4] 曾强, 苏绮琪, 郑嘉仪, 等. 基于贝叶斯时空建模的高速公路事故黑点判别[J]. 交通信息与安全, 2020, 38(6): 87-94. doi: 10.3963/j.jssn.1674-4861.2020.06.012

    ZENG Q, SU Q Q, ZHENG J Y, et al. Identification of freeway crash hotspots based on bayesian spacetime modeling[J]. Journal of Transport Information and Safety, 2020, 38(6): 87-94. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2020.06.012
    [5] XU C, LIU P, WAND W, et al. Evaluation of the impacts of traffic states on crash risks on freeways[J]. Accident Analysis & Prevention, 2012, 47(1): 162-171. http://search.ebscohost.com/login.aspx?direct=true&db=buh&AN=73569892&site=ehost-live
    [6] WANG J, ZHENG Y, LI X, et al. Driving risk assessment using near-crash database through data mining of tree-based model[J]. Accident Analysis & Prevention, 2015, 84: 54-64. http://www.sciencedirect.com/science?_ob=ShoppingCartURL&_method=add&_eid=1-s2.0-S0001457515300129&originContentFamily=serial&_origin=article&_ts=1492997092&md5=3c80968da0a43c3d2404eab6f16f51ec
    [7] OH C, PARK S, RITCHIE S G. A method for identifying rear-end collision risks using inductive loop detectors[J]. Accident Analysis & Prevention, 2006, 38(2): 295-301. http://www.onacademic.com/detail/journal_1000034577385310_f6a9.html
    [8] THOEFILATOS A, CHEN C, ANTONIOU C. Comparing machine learning and deep learning methods for real-time crash prediction[J]. Transportation Research Record, 2019, 2673(8): 169-178. doi: 10.1177/0361198119841571
    [9] 赵海涛, 程慧玲, 丁仪, 等. 基于深度学习的车联边缘网络交通事故风险预测算法研究[J]. 电子与信息学报, 2020, 42(1): 50-57. https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX202001006.htm

    ZHAO H T, CHENG H L, DING Y, et al. Research on traffic accident risk prediction algorithm based on deep learning in car link edge network[J]. Acta Electronica Sinica, 2020, 42(1): 50-57. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-DZYX202001006.htm
    [10] HOSSAIN M, ABDEL-ATY M, QUDDUS M A, et al. Real-time crash prediction models: State-of-the-art, design pathways and ubiquitous requirements[J]. Accident Analysis & Prevention, 2019(124): 66-84. http://www.xueshufan.com/publication/2910624182
    [11] SUN J, SUN J. A dynamic Bayesian network model for real-time crash prediction using traffic speed conditions data[J]. Transportation Research Part C: Emerging Technologies, 2015(54): 176-186. http://www.researchgate.net/profile/Jie_Sun50/publication/274405799_A_dynamic_Bayesian_network_model_for_real-time_crash_prediction_using_traffic_speed_conditions_data/links/566666e808ae418a786f445e.pdf
    [12] YASMIN S, ELURU N, WANG L, et al. A joint framework for static and real-time crash risk analysis[J]. Analytic Methods in Accident Research, 2018, 18: 45-56. doi: 10.1016/j.amar.2018.04.001
    [13] PANDE A, ABDEL-ATY M. Comprehensive analysis of the relationship between real-time traffic surveillance data and rear-end crashes on freeways[J]. Transportation Research Record, 2006, 1953(1): 31-40. doi: 10.1177/0361198106195300104
    [14] XU C C, WANG W, LIU P. A genetic programming model for real-time crash prediction on freeways[J]. IEEE Transactions on Intelligent Transportation Systems, 2013, 14(2): 574-586. doi: 10.1109/TITS.2012.2226240
    [15] AHMED M M, ABDEL-ATY M. The viability of using automatic vehicle identification data for real-time crash prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2012, 13(2): 459-468. doi: 10.1109/TITS.2011.2171052
    [16] WANG L, ABDEL-ATY M, SHI Q, et al. Real-time crash prediction for expressway weaving segments[J]. Transportation Research Part C: Emerging Technologies, 2015(61): 1-10. http://www.researchgate.net/profile/Ling_Wang43/publication/284068999_Real-time_crash_prediction_for_expressway_weaving_segments/links/5733520d08ae9f741b2610e1.pdf
    [17] 程国柱, 刚杰, 程瑞, 等. 公路货运通道路侧事故多发路段判别与线形设计[J]. 哈尔滨工业大学学报, 2022, 54 (3): 131-138. https://www.cnki.com.cn/Article/CJFDTOTAL-HEBX202203015.htm

    CHEN G Z, GANG J, CHEN R, et al. Identification of roadside accident blackspot and geometric design of dedicated freight corridor on highways[J]. Journal of Harbin Institute of Technology, 2022, 54(3): 131-138. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-HEBX202203015.htm
    [18] 高珍, 高屹, 余荣杰, 等. 连续数据环境下的道路交通事故风险预测模型[J]. 中国公路学报, 2018, 31(4): 280-287. doi: 10.3969/j.issn.1001-7372.2018.04.032

    GAO Z, GAO Y, YU R J, et al. Road crash risk prediction model for continuous streaming data environment[J]. China Journal of Highway and Transport, 2018, 31(4): 280-287. (in Chinese) doi: 10.3969/j.issn.1001-7372.2018.04.032
    [19] 沈静. 高速公路事故风险实时预测及事后时空影响分析[D]. 南京: 东南大学, 2017.

    SHEN J. Real-time risk prediction and spatiotemporal impact analysis for freeway accident[D]. Nanjing: Southeast University, 2017. (in Chinese)
    [20] 姜正申, 刘宏志, 付彬, 等. 集成学习的泛化误差和AUC分解理论及其在权重优化中的应用[J]. 计算机学报, 2019, 42(1): 1-15. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJX201901001.htm

    JIANG Z S, LIU H Z, FU B, et al. Decomposition theories of generalization error and AUC in ensemble learning with application in weight optimization[J]. Chinese Journal of Computers, 2019, 42(1): 1-15. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSJX201901001.htm
    [21] 王文宪, 况瑢, 郭经纬, 等. 铁路专用线危险货物运输安全指标属性约简研究[J]. 中国安全生产科学技术, 2017, 13 (11): 59-65. https://www.cnki.com.cn/Article/CJFDTOTAL-LDBK201711011.htm

    WANG W X, KUANG R, GUO J W, et al. Research on attribute reduction for safety indexes of dangerous goods transportation in special railway[J]. Journal of Safety Science and Technology, 2017, 13(11): 59-65. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-LDBK201711011.htm
    [22] 石宁宁. 驾驶员事故频次分布及其影响因素分析[D]. 北京: 北京交通大学, 2018.

    SHI N N. Drivers' accident frequency distribution and its influencing factors[D]. Beijing: Beijing Jiaotong University, 2018. (in Chinese)
    [23] THEOFILATOS A. Incorporating real-time traffic and weather data to explore road accident likelihood and severity in urban arterials[J]. Journal of Safety Research, 2017(61): 9-21. http://www.onacademic.com/detail/journal_1000039838751910_a13a.html
    [24] ABDEL-ATY M, HASSAN H, AHMED M, et al. Real-time prediction of visibility related crashes[J]. Transportation Research Part C: Emerging Technologies, 2012(24): 288-298. http://www.sciencedirect.com/science/article/pii/S0968090X12000514
    [25] LIU X L, XU J L, DONG Y P, et al. Defining highway node acceptance capacity (HNAC): Theoretical analysis and data simulation[J]. Journal of Advanced Transportation, 2020, 2020: 8939621. http://www.researchgate.net/publication/338613933_defining_highway_node_acceptance_capacity_hnac_theoretical_analysis_and_data_simulation
    [26] SHANG W Q, HUANG H K, ZHU H B, et al. A novel feature selection algorithm for text categorization[J]. Expert Systems with Applications, 2007, 33(1): 1-5. doi: 10.1016/j.eswa.2006.04.001
    [27] 杨杰明. 文本分类中文本表示模型和特征选择算法研究[D]. 长春: 吉林大学, 2013.

    YANG J M. The research of text representation and feature selection in text categorization[D]. Changchun: Jilin University, 2013. (in Chinese)
  • 加载中
图(4) / 表(4)
计量
  • 文章访问数:  936
  • HTML全文浏览量:  287
  • PDF下载量:  679
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-04-22
  • 网络出版日期:  2022-09-17

目录

    /

    返回文章
    返回