行车风险量化新方法及其防控策略仿真研究  

Novel Quantitative Approach for Assessing Driving Risks and Simulation Study of Its Prevention and Control Strategies

作　　者：郑讯佳 蒋骏皓 黄荷叶 王建强[3] 许庆[3] 张强 ZHENG Xunjia;JIANG Junhao;HUANG Heye;WANG Jianqiang;XU Qing;ZHANG Qiang(School of Intelligent Manufacturing Engineering,Chongqing University of Arts and Sciences,Chongqing 402160;China Automotive Engineering Research Institute Co.,Ltd.,Chongqing 401122;School of Vehicle and Mobility,Tsinghua University,Beijing 100084)

机构地区：[1]重庆文理学院智能制造工程学院,重庆402160 [2]中国汽车工程研究院股份有限公司,重庆401122 [3]清华大学车辆与运载工程学院,北京100084

出　　处：《机械工程学报》2024年第10期207-221,共15页Journal of Mechanical Engineering

基　　金：国家自然科学基金(52102454);中国博士后科学基金面上(2021M700169);重庆市自然科学基金面上(cstc2021jcyj-msxm X0395);重庆市博士后研究(2021XM3069);重庆市教委科学技术研究(KJQN202001302);智能绿色车辆与交通全国重点实验室开放基金课题(KFY2412)资助项目。

摘　　要：行车风险受人、车、路多类时变因素的耦合影响,如何对其进行准确量化一直是制约汽车智能安全技术发展的难题。提出考虑驾驶人行为特性的行车风险量化新方法并介绍对应的行车风险防控策略,将行车风险量化与等效力模型结合,设计行车风险的场模型框架。通过考虑车道线约束和车辆行驶路径的变化,在车辆行驶的纵向和横线上设置不同的风险梯度调整系数并采用Frenet坐标转换,使得行车风险在车辆行驶的纵向和横向上具有较大的差异。结合行车风险的防控指标,设计3种高速公路直行冲突场景和9种无信号灯交叉路口冲突仿真场景,并分别与纵向控制模型(Longitudinal control model,LCM)和预计时间(Time to intersection,TTI)阈值对比,仿真结果表明,所设计的车辆行驶决策模型能够识别各个方向上的风险并能主动执行风险防控,相比LCM模型在安全性上更优;另外,建立的左转风险防控算法在900次无信号交叉路口仿真中的碰撞事故发生率为0,相比TTI阈值算法具有更高的通行效率和安全性。Accurately quantifying the integrated influence of diverse,time-varying factors associated with drivers,vehicles,and road conditions on driving risk has persistently presented a formidable obstacle in advancing intelligent automotive safety technologies.The novel approach proposed quantifies driving risks by considering driving behavior characteristics and introduces corresponding strategies for preventing and controlling these risks.The method combines driving risk quantification with an equivalent force model,designing a field model framework for driving risk.It considers the variability of the vehicle’s driving path by Frenet coordinate transformation and considers lane constraints when establishing different risk gradient adjustment coefficients for the longitudinal and lateral directions of the vehicle,which is employed to create significant distinctions in driving risk between the longitudinal and lateral directions of the vehicle.As a result,three high-speed highway straight-ahead conflict scenarios and nine unsignalized intersection conflict scenarios are developed to effectively compare driving risk prevention and control indicators with the LCM model and TTI threshold,respectively.Simulation results show that the developed vehicle-driving decision-making model can identify risks in all directions and actively implement risk prevention and control measures,providing enhanced safety compared to the LCM model.Furthermore,the left-turn risk prevention and control algorithm established here achieves a collision rate of 0 in 900 unsignalized intersection simulations,offering improved traffic efficiency and safety compared to the TTI threshold algorithm.

关 键 词：行车风险 量化建模 风险防控 行车安全场 

分 类 号：U461[机械工程—车辆工程]