汽车工程 ›› 2022, Vol. 44 ›› Issue (9): 1339-1349.doi: 10.19562/j.chinasae.qcgc.2022.09.005

所属专题: 智能网联汽车技术专题-感知&HMI&测评2022年

• • 上一篇    下一篇

基于驾驶人跟驰特性的前撞预警指标与阈值确定方法

龙文民,鲁光泉(),石茜,谭海天   

  1. 北京航空航天大学交通科学与工程学院,北京  102206
  • 收稿日期:2022-04-02 修回日期:2022-04-24 出版日期:2022-09-25 发布日期:2022-09-21
  • 通讯作者: 鲁光泉 E-mail:lugq@buaa.edu.cn
  • 基金资助:
    国家重点研发计划(2018YFB1600500)

Determination Method of Forward Collision Warning Indicator and Threshold Based on Drivers’ Car-Following Characteristics

Wenmin Long,Guangquan Lu(),Xi Shi,Haitian Tan   

  1. School of Transportation Science and Technology,Beihang University,Beijing  102206
  • Received:2022-04-02 Revised:2022-04-24 Online:2022-09-25 Published:2022-09-21
  • Contact: Guangquan Lu E-mail:lugq@buaa.edu.cn

摘要:

为更合理地选取前向碰撞预警指标并针对不同驾驶人个性化地设置其预警阈值,提出了基于信号检测论与秩和比法的前向碰撞预警指标综合评价方法,实现了基于跟驰数据的驾驶人反应时间在线辨识与更新。开展自然驾驶实验采集城市交通环境下的驾驶行为数据,验证了驾驶人反应时间在线辨识的有效性。进一步通过统计分析辨识得到不同驾驶人的个性化预警指标阈值。对不同预警指标阈值结果的综合比较和分析表明:6种前向碰撞预警指标中安全裕度(SM)最优;风险感知系数(RP)次之;秩和比值按大小依次为距离碰撞时间倒数(TTCi)、车头时距(THW)和预警系数(W);时间裕度(TM)评价结果波动较大。

关键词: 交通工程, 跟驰行为, 前向碰撞预警, 信号检测论, 预警指标

Abstract:

In order to more reasonably select the forward collision warning indicator and set the personalized warning threshold for different drivers, a comprehensive evaluation method of forward collision warning indicator is proposed based on signal detection theory and rank sum ratio method, achieving the on-line identification and updating of driver’s response time based on vehicle following data. Natural driving experiments are carried out to collect driving behaviors data in urban traffic environment for verifying the effectiveness of on-line identification of driver’s response time. Furthermore, the personalized warning thresholds for different drivers are obtained through the statistical analysis and identification. The comprehensive comparison and analysis on the different warning indicator thresholds obtained show that among 6 forward collision warning indicators, safety margin is the best, risk perception factor the next, the rank sum ratio follows the order of reciprocal time to collision > time headway > warning coefficient, while the evaluation results of time margin fluctuate greatly.

Key words: traffic engineering, following behavior, forward collision warning, signal detection theory, warning indicators