|本期目录/Table of Contents|

[1]朱兴林,王诺,刘泓君,等.基于可解释集成学习的电动自行车骑行风险识别及致因解析*[J].中国安全生产科学技术,2026,22(4):88-97.[doi:10.11731/j.issn.1673-193x.2026.04.011]
 ZHU Xinglin,WANG Nuo,LIU Hongjun,et al.Risk identification and causal analysis of electric bike riding using explainable ensemble learning[J].Journal of Safety Science and Technology,2026,22(4):88-97.[doi:10.11731/j.issn.1673-193x.2026.04.011]
点击复制

基于可解释集成学习的电动自行车骑行风险识别及致因解析*

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
22
期数:
2026年4期
页码:
88-97
栏目:
安全工程技术
出版日期:
2026-04-30

文章信息/Info

Title:
Risk identification and causal analysis of electric bike riding using explainable ensemble learning
文章编号:
1673-193X(2026)-04-0088-10
作者:
朱兴林王诺刘泓君吴隽朝王光东郭瑞
(1.新疆农业大学 交通与物流工程学院,新疆 乌鲁木齐 830052;
2.新疆农业大学 新疆交通运输与物流工程重点实验室,新疆 乌鲁木齐 830052;
3.新疆道路交通安全工程技术研究中心,新疆 乌鲁木齐 830094)
Author(s):
ZHU Xinglin WANG Nuo LIU Hongjun WU Junchao WANG Guangdong GUO Rui
(1.School of Transportation & Logistics Engineering,Xinjiang Agricultural University,Urumqi Xinjiang 830052,China;
2.Key Laboratory of Transportation and Logistics Engineering in Xinjiang,Xinjiang Agricultural University,Urumqi Xinjiang 830052,China;
3.Xinjiang Engineering Technology Research Center for Road Traffic Safety,Urumqi Xinjiang 830094,China)
关键词:
交通安全电动自行车骑行风险识别Stacking集成学习部分依赖图算法SHAP算法
Keywords:
traffic safety electric bicycle risk identification stacking ensemble learning partial dependence plot SHAP algorithm
分类号:
X951
DOI:
10.11731/j.issn.1673-193x.2026.04.011
文献标志码:
A
摘要:
为实现对城市电动自行车骑行风险的识别与预防,基于自然骑行实验构建了包含骑行者违规行为、生理特征、车辆运行特征、道路条件、土地利用特征的多源异构数据库,依据车辆运行数据将骑行风险划分为低、中、高3个等级用于风险标签,采用Stacking集成学习算法构建骑行风险识别模型,结合部分依赖图(partial dependence plot,PDP)算法与SHAP值分析(SHapley Additive exPlanations,SHAP)等解释方法探讨各特征对骑行风险的影响规律。研究结果表明:构建的骑行风险识别模型准确率为94.26%;影响骑行风险的关键因素依次为闯红灯率、车速均值、逆行发生率、加速度标准差、航向角标准差、心率平均值;车辆运行特征与道路条件、土地利用特征的交互作用对骑行风险具有差异化影响;合理设置非机动车道宽度与机非分隔形式,均能有效提升骑行安全性。研究结果可为实现城市电动自行车骑行安全风险防控提供参考。
Abstract:
In order to achieve the identification and prevention of urban electric bicycle riding risks,a multi-source heterogeneous database is constructed based on naturalistic riding experiments,containing rider violation behaviors,physiological characteristics,vehicle operational parameters,road conditions,and land-use characteristics.Riding risk is classified into three levels (low,medium,and high) based on vehicle operational data to create risk labels.A Stacking ensemble learning algorithm is employed to construct a riding risk identification model.Interpretability methods such as the partial dependence plot (PDP) and SHapley Additive exPlanations (SHAP) are integrated to investigate the influence patterns of various features on riding risk.Results indicate that the proposed model achieves an accuracy of 94.26%.The key factors affecting riding risk,in descending order of importance,are red-light running rate,average speed,incidence of riding against traffic,standard deviation of acceleration,standard deviation of heading angle,and average heart rate.The interaction between vehicle operational characteristics and road or land-use features exerts a differentiated impact on riding risk.Appropriately setting the width of non-motorized lanes and the form of motorized-non-motorized separation can effectively enhance riding safety.The findings provide actionable insights for the prevention and control of safety risks in urban electric bicycle riding.

参考文献/References:

[1]新浪财经.中国电动自行车保有量已破4亿辆,大城市求解全链路闭环治理[EB/OL].(2024-12-13)[2025-02-29].https://finance.sina.cn/2024-12-13/detail-incziqvp8941082.d.html.
[2]寇敏,张萌萌,赵军学,等.道路交通安全风险辨识与分析方法综述[J].交通信息与安全,2022,40(6):22-32. KOU Min,ZHANG Mengmeng,ZHAO Junxue,et al.A review of identification and analysis methods for road safety risk[J].Journal of Transport Information and Safety,2022,40(6):22-32.
[3]PATEL H,GORE N,EASA S,et al.Novel traffic conflict-based framework for real-time traffic safety evaluation under heterogeneous and weak lane-discipline traffic[J].Transportation Research Record,2024,2678(2):118-134.
[4]覃国峰,王涛,赵军,等.基于SEM的电动自行车驾驶员危险性驾驶行为研究[J].黑龙江科技信息,2016(16):140-141. QIN Guofeng,WANG Tao,ZHAO Jun,et al.Study on dangerous driving behaviors of electric bicycle riders based on SEM[J].Heilongjiang Science and Technology Information,2016(16):140-141.
[5]张晓龙,边扬,赵晓华,等.共享电动自行车路段超速风险影响因素分析[J].东南大学学报(自然科学版),2024,54(1):214-223. ZHANG Xiaolong,BIAN Yang,ZHAO Xiaohua,et al.Analysis of influencing factors of speeding risk on road segments for shared electric bikes[J].Journal of Southeast University(Natural Science Edition),2024,54(1):214-223.
[6]边扬,杨家夏,赵晓华,等.基于轨迹数据的共享电动自行车逆行风险行为影响因素研究[J].中国公路学报,2021,34(12):262-275. BIAN Yang,YANG Jiaxia,ZHAO Xiaohua,et al.Research on influencing factors of reverse riding risk behavior of shared e-bike based on trajectory data[J].China Journal of Highway and Transport,2021,34(12):262-275.
[7]曹士强,边扬,荣建,等.基于道路交通环境的非机动车骑行行为特征研究[J].交通工程,2015,15(4):44-48. CAO Shiqiang,BIAN Yang,RONG Jian,et al.Study on riding behavior characteristics of non-motor vehicles based on road traffic environment[J].Journal of Transportation Engineering,2015,15(4):44-48.
[8]邓院昌,蒋昀轩,陶胜芹.基于可解释集成学习的异常驾驶行为风险识别方法[J].交通运输系统工程与信息,2025,25(2):180-189. DENG Yuanchang,JIANG Yunxuan,TAO Shengqin.Risk identification method for abnormal driving behavior based on interpretable ensemble learning model[J].Journal of Transportation Systems Engineering and Information Technology,2025,25(2):180-189.
[9]谢厅,刘星良,刘唐志,等.基于DE-EL的城市快速路合流区危险驾驶行为识别方法[J].交通信息与安全,2024,42(6):23-30. XIE Ting,LIU Xingliang,LIU Tangzhi,et al.Identification method of dangerous driving behaviors in weaving areas of urban expressways based on DE-EL[J].Journal of Transportation Information and Safety,2024,42(6):23-30.
[10]KENPANKHO P,CHAICHANA A,TRACHU K,et al.Real-time GPS receiver bias estimation[J].Advances in Space Research,2021,68(5):2152-2159.
[11]中华人民共和国工业和信息化部.电动自行车安全技术规范:GB 17761—2024[S].北京:中国标准出版社,2025.
[12]陈必俊.基于交通轨迹大数据的驾驶行为安全性评价方法[D].福州:福建工程学院,2021.
[13]交通运输部公路局.公路安全生命防护工程实施技术指南[M].北京:人民交通出版社,2015.
[14]ZHAO X,YANG H,YAO Y,et al.Factors affecting traffic risks on bridge sections of freeways based on partial dependence plots[J].Physica A:Statistical Mechanics and its Applications,2022,598:127343.
[15]DONG S,KHATTAK A,ULLAH I,et al.Predicting and analyzing road traffic injury severity using boosting-based ensemble learning models with SHapley Additive exPlanations[J].International Journal of Environmental Research and Public Health,2022,19(5):2925.
[16]WILDE G J S.The theory of risk homeostasis:implications for safety and health[J].Risk Analysis,1982,2(4):209-225.
[17]PELTZMAN S.The effects of automobile safety regulation[J].Journal of Political Economy,1975,83(4):677-725.

相似文献/References:

[1]陈 春.道路交通事故的影响因素研究——基于结构方程模型的实证研究[J].中国安全生产科学技术,2014,10(5):103.[doi:10.11731/j.issn.1673-193x.2014.05.018]
 CHEN Chun.Research on influencing factors of road traffic accidents——empirical study based on structural equation model[J].Journal of Safety Science and Technology,2014,10(4):103.[doi:10.11731/j.issn.1673-193x.2014.05.018]
[2]赵亮,刘浩学.驾驶行为特性与人格特征关系研究[J].中国安全生产科学技术,2016,12(8):171.[doi:10.11731/j.issn.1673-193x.2016.08.029]
 ZHAO Liang,LIU Haoxue.Study on relationship between driving behavior characteristics and personality traits[J].Journal of Safety Science and Technology,2016,12(4):171.[doi:10.11731/j.issn.1673-193x.2016.08.029]
[3]张芳燕,牛玺荣.2000~2013年我国公路交通安全研究进展及其分析[J].中国安全生产科学技术,2015,11(4):173.[doi:10.11731/j.issn.1673-193X.2015.04.028]
 ZHANG Fang-yan,NIU Xi-rong.Research progress and analysis on highway traffic Safety in China during 2000 to 2013[J].Journal of Safety Science and Technology,2015,11(4):173.[doi:10.11731/j.issn.1673-193X.2015.04.028]
[4]高岩,董宪元,田飞.高速公路交通安全现状分析及管理对策[J].中国安全生产科学技术,2015,11(10):110.[doi:10.11731/j.issn.1673-193x.2015.10.019]
 GAO Yan,DONG Xian-yuan,TIAN Fei.Present situation analysis on highway traffic safety and management countermeasures[J].Journal of Safety Science and Technology,2015,11(4):110.[doi:10.11731/j.issn.1673-193x.2015.10.019]
[5]吴玲,路巧珍,朱彤,等.视觉记忆型次任务对驾驶绩效及安全的影响[J].中国安全生产科学技术,2016,12(5):158.[doi:10.11731/j.issn.1673-193x.2016.05.027]
 WU Ling,LU Qiaozhen,ZHU Tong,et al.Influence of visual memory secondary task on driving performance and safety[J].Journal of Safety Science and Technology,2016,12(4):158.[doi:10.11731/j.issn.1673-193x.2016.05.027]
[6]刘君,胡伟超,孙广林.公路突发事件应急预案自动生成系统开发及应用[J].中国安全生产科学技术,2017,13(10):53.[doi:10.11731/j.issn.1673-193x.2017.10.009]
 LIU Jun,HU Weichao,SUN Guanglin.Development and application of automatic generation system for emergency plan of highway emergency[J].Journal of Safety Science and Technology,2017,13(4):53.[doi:10.11731/j.issn.1673-193x.2017.10.009]
[7]王宗直,路堃,陈涛,等.基于格子气模型的北京地铁站内人流移动分析[J].中国安全生产科学技术,2018,14(5):31.[doi:10.11731/j.issn.1673-193x.2018.05.005]
 WANG Zongzhi,LU Kun,CHEN Tao,et al.Analysis on pedestrian flow movement in a subway station of Beijing based on lattice gas model[J].Journal of Safety Science and Technology,2018,14(4):31.[doi:10.11731/j.issn.1673-193x.2018.05.005]
[8]赵俊玮,华珺,刘永涛,等.不同警告方式下驾驶人应激反应能力模拟试验研究[J].中国安全生产科学技术,2019,15(6):162.[doi:10.11731/j.issn.1673-193x.2019.06.026]
 ZHAO Junwei,HUA Jun,LIU Yongtao,et al.Simulation test study on stress response ability of drivers under different warning modes[J].Journal of Safety Science and Technology,2019,15(4):162.[doi:10.11731/j.issn.1673-193x.2019.06.026]
[9]王羽尘,马健霄,陆涛,等.公路隧道火灾发生位置与人群疏散通道仿真研究[J].中国安全生产科学技术,2019,15(10):38.[doi:10.11731/j.issn.1673-193x.2019.10.006]
 WANG Yuchen,MA Jianxiao,LU Tao,et al.Simulation study on fire location and crowd evacuation route in highway tunnel[J].Journal of Safety Science and Technology,2019,15(4):38.[doi:10.11731/j.issn.1673-193x.2019.10.006]
[10]胡立伟,赵雪亭,杨锦青,等.城市快速过境通道衔接节点交通风险耦合致因模型研究[J].中国安全生产科学技术,2019,15(12):150.[doi:10.11731/j.issn.1673-193x.2019.12.024]
 HU Liwei,ZHAO Xueting,YANG Jinqing,et al.Research on coupling cause model of traffic risk in connecting nodes of urban rapid transit channels[J].Journal of Safety Science and Technology,2019,15(4):150.[doi:10.11731/j.issn.1673-193x.2019.12.024]

备注/Memo

备注/Memo:
收稿日期: 2025-10-17
* 基金项目: 新疆维吾尔自治区自然科学基金项目(2024D01A64);新疆维吾尔自治区高校基本科研业务费科研项目(XJEDU2024P040)
作者简介: 朱兴林,博士,教授,主要研究方向为交通安全与环境。
更新日期/Last Update: 2026-04-29