摘要
为了分析城市道路环境下高度自动驾驶中非驾驶相关任务和接管紧迫度对接管绩效的影响,基于驾驶模拟器设计了自动驾驶紧急接管场景并开展驾驶模拟试验,接管请求时间分别设定为3,4,5s,非驾驶相关任务为读新闻、看视频、玩游戏,自动驾驶车速为50km·h^1,试验中共招募了49名被试(男性30名,女性19名),被试的平均年龄为31.06岁(标准差为7.1岁),驾驶人在自动驾驶阶段始终执行非驾驶相关任务,听到接管请求提示后需要接管车辆的控制权,并实施紧急避让操作。研究结果表明:在紧急接管情况下,接管紧迫度对合成加速度和最小TTC有影响,而对接管时间无影响,与5s的接管请求时间条件相比,3,4s的接管请求时间条件下的合成加速度明显增加,而最小TTC则随接管请求时间的减少而降低;非驾驶相关任务对接管时间和最小TTC有影响,而对合成加速度无影响,与无非驾驶相关任务相比,非驾驶相关任务会显著增加接管时间和降低最小TTC;碰撞几乎都发生在3s和4s的接管请求时间下,5s的接管请求时间能够基本保证接管的安全性。
To examine the effect of non-driving-related tasks ( NDRTs) and the urgency of the take-over during highly automated driving in urban road environment conditions,critical takeover scenarios during automated driving were designed using a driving simulator,and an experimental study was conducted.Take-over request times were set at 3 s,4 s,and 5 s,and NDRTs included reading news,watching videos,and playing games.The automated driving speed was 50 km·h-^1.49 participants (30 males,19 females) were recruited for the experiments,and the mean age of the participants was 31.06 years (SD 7.10 years).The participants were asked to engage in NDRTs during the automated driving mode.They were then asked to take back control of the vehicle when a take-over request was given and perform avoidance maneuvers.The results show that,in critical take-over situations,the urgency of the take-over affects the resultant acceleration and minimum time to collision (TTC) but has no effect on take-over time.Moreover,compared to the take-over request time of 5 s,the resultant accelerations with the take-over request times of 3 s and 4 s are significantly higher,whereas the minimum TTC decreases with a shorter take-over request time.NDRTs have an impact on the take-over time and minimum TTC but do not influence the resultant acceleration.In addition,NDRTs can significantly increase the take-over time and decrease the minimum TTC compared to cases when no NDRTs are engaged.Crashes nearly occur when the take-over request times are 3 s and 4 s,and safety can be ensured when the take-over request time is 5 s.
作者
林庆峰
王兆杰
鲁光泉
LIN Qing-feng;WANG Zhao-jie;LU Guang-quan(Beijing Advanced Innovation Center for Big Data and Brain Computing,Beihang University,Beijing 100191,China;School of Transportation Science and Engineering,BeihangUniversity,Beijing 100191,China)
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2019年第6期240-247,共8页
China Journal of Highway and Transport
基金
国家自然科学基金项目(U1664262)
国家重点研发计划项目(2017YFC0804802)
关键词
交通工程
自动驾驶
驾驶模拟
接管
驾驶行为
traffic engineering
automated driving
driving simulation
take-over
driving behavior
