基于轨迹数据的城市快速路小型车换道动态风险综合量化方法  

Trajectory data based comprehensive quantitative method for light-duty vehicle dynamic lane-changing risk on urban expressway

作　　者：余加勇[1] 胡瑶 李洁[1] 高文宇 方振 YU Jiayong;HU Yao;LI Jie;GAO Wenyu;FANG Zhen(College of Civil Engineering,Hunan University,Changsha,Hunan 410082,China;Hunan Communications Research Institute Co.,Ltd.,Changsha,Hunan 410015,China)

机构地区：[1]湖南大学土木工程学院,湖南长沙410082 [2]湖南省交通科学研究院有限公司,湖南长沙410015

出　　处：《公路交通科技》2025年第2期13-22,共10页Journal of Highway and Transportation Research and Development

基　　金：湖南省科技成果转化及产业化计划项目(2020GK2026);湖南省自然科学基金项目(2021JJ30102);湖南省水利科技项目(XSKJ2021000-46);湖南省自然资源厅科技计划项目(20230120DZ)。

摘　　要：【目标】为保证车辆安全换道并防止潜在的碰撞事故,基于自然驾驶轨迹数据,应用故障树法构建一种融合时间和空间风险指标的换道动态风险综合量化方法。【方法】该方法将风险严重水平(RSL)和改良的空间暴露水平(REL)纳入综合风险指数,反映车辆之间的潜在运动相关性。改良的REL同时考虑了换道车辆和周围不同位置车辆的碰撞时间(TTC)阈值及安全停车距离指数(SDI),并在最后应用故障树法分析车辆之间的换道风险。【数据】通过对车辆轨迹数据进行处理,提取了604组有效变道样本,并对综合风险量化方法进行应用分析。【结果】在静态交通冲突方面,不同车辆位置组合的TTC阈值不同,其中换道车辆与目标车道后车的TTC严重冲突和一般冲突阈值皆更大;由于车道、距离和视野等因素,目标车道的后方驾驶员感知碰撞风险的时间和阈值更大,导致碰撞时间更小;在动态交通冲突方面,车辆换道风险和车辆的前后位置有关但和车道位置无关,其中换道车辆与后方车辆的换道风险大于与前方车辆;使用不同TTC阈值能提高风险识别准确率,综合时空风险量化方法能细化评价结果,且计算简便。【结论】综合量化方法能识别车道变换过程中的潜在冲突,可为小型车智能换道预警系统的设计提供支持。[Objective]To ensure cars lane-changing safety and prevent potential collision accidents,based on the natural driving trajectory data,a comprehensive quantification method was proposed for evaluating the lane-changing dynamic risk.The proposed method,integrating temporal and spatial risk indicators,was constructed by using the fault tree method.[Method]The method incorporated the risk severity level(RSL)and modified spatial exposure level(REL)into the comprehensive risk indicators to reflect the potential motion correlation among vehicles.The modified REL considered the different time-to-collision(TTC)thresholds and the safe stopping distance index (SDI) of lane-changing vehicles and surrounding vehicles at differentpositions. The fault tree method was finally applied to analyze the lane-changing risk among vehicles. [Data]604 sets of effective lane-changing samples were extracted by processing the vehicle trajectory data. Thecomprehensive risk quantification method was applied for analysis. [Result] In terms of static traffic conflicts,the TTC thresholds of different vehicle position combinations are different. The TTC serious conflict thresholdsand general conflict threshold between the lane-changing vehicle and the vehicle behind target lane are large.Due to lane, distance and field of view, the driver behind target lane has more time and lager threshold toperceive the collision risk. That results in less collision time. In the aspect of dynamic traffic conflict, the lanechangingrisk is related to the front and rear positons of vehicle, but not related to the lane position. The lanechangingrisk is greater with the rear vehicle than that with the front vehicle. The risk identification accuracy isimproved by using different TTC thresholds. The comprehensive spatio-temporal risk quantification method canrefine the evaluation result with simple calculations. [Conclusion] The comprehensive quantification methodcan identify the potential conflicts during lane-changing, and support for the intelligent lane-changing w

关 键 词：智能交通 换道风险 故障树法 车辆换道 快速路 

分 类 号：U491[交通运输工程—交通运输规划与管理]