基于生存分析的自动驾驶接管时间分析

doi:10.16265/j.cnki.issn1003-3033.2023.08.1474

摘要/Abstract

摘要：

为探究有条件自动驾驶接管过程中非驾驶相关任务、接管请求时间、接管事件类型及交通流量对接管时间的影响,利用正交试验原理设计并搭建接管场景,招募驾驶人进行自动驾驶接管试验,采集试验过程中的车辆运行状态及驾驶人行为数据,分析不同因素对驾驶人接管时间的影响,并构建随机效应生存分析模型,预测接管时间。结果表明:与无次任务相比,执行非驾驶相关任务会显著增加接管时间;接管请求时间为4 s时的接管时间要显著小于5和6 s时的接管时间;与施工区事件相比,事故车辆与换道并减速事件下驾驶人的接管时间较长;交通流量越大,接管时间越短。相比于固定效应生存分析模型,随机效应生存分析模型能揭示数据中未观测到的异质性,具有良好的拟合优度及预测效果。

关键词: 自动驾驶, 接管时间, 驾驶模拟, 生存分析, 随机效应

Abstract:

In order to gain a comprehensive understanding of the impacts of factors including non-driving related tasks, takeover request lead time, type of takeover events, and traffic volume on takeover time during the automated driving takeover process, an orthogonal experimental design was applied to develop the takeover scenarios. Drivers were recruited to participate in the takeover experiments based on a driving simulator, with driver behavior and vehicle trajectory data collected. Factors' impacts on drivers' takeover time were analyzed, and a random-effect survival duration model was developed to predict takeover time. Results reveal that compared with no secondary tasks, performing the non-driving related tasks increased takeover time significantly. When the takeover request lead time was 4 s, takeover time was shorter than that with 5 and 6 s. Takeover time in the accident car and cut-in events was longer than that for the work zone event. Moreover, larger traffic volume led to shorter takeover time. Compared with the fixed-effects survival analysis model, the random-effects survival analysis model accounted for the unobserved heterogeneity in the data and provided good goodness-of-fitness and prediction accuracy.

Key words: automated driving, takeover time, driving simulation, survival analysis, random effect

王长帅, 徐铖铖, 邵永成, 佟昊, 彭畅. 基于生存分析的自动驾驶接管时间分析[J]. 中国安全科学学报, 2023, 33(8): 142-148.

WANG Changshuai, XU Chengcheng, SHAO Yongcheng, TONG Hao, PENG Chang. Analysis of takeover time of automated driving based on survival analysis[J]. China Safety Science Journal, 2023, 33(8): 142-148.

图/表 9

图1

图2

表1

图3

图4

表2

图5

表3

模型参数估计结果

变量	参数值	标准差	P值
新手驾驶人	-0.025	0.036	0.499
女性	0.038	0.039	0.333
看视频	0.172	0.018	<0.001
玩游戏	0.193	0.018	<0.001
事故车辆	0.070	0.017	<0.001
换道并减速	0.089	0.017	<0.001
接管请求时间	0.023	0.009	0.009
交通流量	-0.089	0.012	<0.001
截距	0.316	0.057	<0.001
σ	0.110	0.011	<0.001
随机效应	0.010	0.008	<0.001
韦布尔分布参数 $μ$ 的值	2.111	0.045	<0.001
LL(β)	-2.553	χ²(9)	200.64

表3

图6

参考文献 16

[1]	DU Na, ZHOU Feng, PULVER E M, et al. Examining the effects of emotional valence and arousal on takeover performance in conditionally automated driving[J]. Transportation Research Part C: Emerging Technologies, 2020, 112: 78-87. doi: 10.1016/j.trc.2020.01.006
[2]	J3016-2018, Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicle[S].
[3]	郭子慧, 郭伟伟, 谭墍元. 驾驶员接管自动驾驶车辆的眼动特性和行为分析[J]. 中国安全科学学报, 2022, 32(1): 65-71. doi: 10.16265/j.cnki.issn1003-3033.2022.01.009
	GUO Zihui, GUO Weiwei, TAN Jiyuan. Analysis on eye movement characteristics and behavior of drivers taking over automated vehicles[J]. China Safety Science Journal, 2022, 32(1): 65-71. doi: 10.16265/j.cnki.issn1003-3033.2022.01.009
[4]	GOLD C, DAMBÖCK D, LORENZ L, et al. "Take over!" how long does it take to get the driver back into the loop?[C]. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 2013: 1938-1942.
[5]	ZHANG Bo, DE WINTER J, VAROTTO S, et al. Determinants of take-over time from automated driving: a meta-analysis of 129 studies[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2019, 64:285-307. doi: 10.1016/j.trf.2019.04.020
[6]	LI Shuo, BLYTHE P, GUO Weihong, et al. Investigation of older driver's takeover performance in highly automated vehicles in adverse weather conditions[J]. IET Intelligent Transport Systems, 2018, 12(9): 1157-1165. doi: 10.1049/itr2.v12.9
[7]	GOLD C, HAPPEE R, BENGLER K. Modeling take-over performance in level 3 conditionally automated vehicles[J]. Accident Analysis & Prevention, 2018, 116: 3-13. doi: 10.1016/j.aap.2017.11.009
[8]	KÖRBER M, GOLD C, LECHNER D, et al. The influence of age on the take-over of vehicle control in highly automated driving[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2016, 39: 19-32. doi: 10.1016/j.trf.2016.03.002
[9]	WAN Jingyan, WU Changxu. The effects of lead time of take-over request and nondriving tasks on taking-over control of automated vehicles[J]. IEEE Transactions on Human-Machine Systems, 2018, 48(6): 582-591. doi: 10.1109/THMS.2018.2844251
[10]	鲁光泉, 赵鹏云, 王兆杰, 等. 自动驾驶中视觉次任务对年轻驾驶人接管时间的影响[J]. 中国公路学报, 2018, 31(4): 165-171.
	LU Guangquan, ZHAO Pengyun, WANG Zhaojie, et al. Impact of visual secondary task on young drivers' take-over time in automated driving[J]. China Journal of Highway and Transport, 2018, 31(4): 165-171.
[11]	王晓勇, 罗珅, 任杰, 等. 公交驾驶员违规间隔时间及影响因素研究[J]. 中国安全科学学报, 2019, 29(6): 128-133. doi: 10.16265/j.cnki.issn1003-3033.2019.06.021
	WANG Xiaoyong, LUO Shen, REN Jie, et al. Research on interval time of traffic violations for bus drivers and its influencing factors[J]. China Safety Science Journal, 29(6): 128-133.
[12]	HENSHER D A, MANNERING F L. Hazard-based duration models and their application to transport analysis[J]. Transport Reviews, 1993, 14(1): 63-82. doi: 10.1080/01441649408716866
[13]	WASHINGTON S, KARLAFTIS M, MANNERING F, et al. Statistical and econometric methods for transportation data analysis[M]. New York: Chapman and Hall/CRC, 2020: 1-496.
[14]	ZHU Tong, WANG Changshuai, YANG Chenxuan, et al. Evaluation of effectiveness of speed reduction markings on driving speed in highway tunnel entrance and exit areas[J]. Promet-Traffic & Transportation, 2020, 32(1): 141-152.
[15]	高岩, 尤志栋, 罗毅, 等. 基于驾驶模拟的自动驾驶接管行为研究[J]. 中国安全生产科学技术, 2021, 17(2): 140-146.
	GAO Yan, YOU Zhidong, LUO Yi, et al. Study on take-over behavior of automated driving based on driving simulation[J]. Journal of Safety Science and Technology, 2021, 17(2): 140-146.
[16]	RADLMAYR J, GOLD C, LORENZ L, et al. How Traffic situations and non-driving related tasks affect the take-over quality in highly automated driving[C]. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 2014(1): 2063-2067.

变量名称	类别	接管时间/s
驾驶人类型	经验驾驶人	1.59±0.347
驾驶人类型	新手驾驶人	1.58±0.302
性别	女性	1.62±0.336
性别	男性	1.57±0.315
非驾驶相关任务	看视频	1.67±0.245
	玩游戏	1.72±0.255
	无次任务	1.37±0.344
接管请求时间/s	4	1.51±0.396
	5	1.64±0.272
	6	1.61±0.269
接管事件	事故车辆	1.61±0.277
	施工区	1.50±0.391
	换道并减速	1.65±0.268
交通流量/ (辆·h^-1)	500	1.65±0.285
	1 000	1.63±0.301
	2 000	1.47±0.348