自动驾驶接管决策中驾驶员认知状态识别研究*

doi:10.16265/j.cnki.issn1003-3033.2026.03.0912

摘要/Abstract

摘要：

为解决自动驾驶接管决策中人机协作失效问题,并精准识别驾驶员认知状态,模拟夜间高速弯道、手机分心驾驶及强光暴雨复合等3类典型场景,采集并分析驾驶行为与眼动特征的动态交互数据,构建多模态感知与认知协同的动态权重分配(DWA)特征融合框架及DWA随机森林(RF)模型,探究复杂环境下驾驶员认知状态与接管决策行为的动态作用机制。结果表明:分心状态显著延长接管时间;在强光暴雨场景中,分心与环境压力叠加导致扫视幅度锐减;疲劳状态则引发车道偏离距离增加,并伴随瞳孔直径缩小及异常扫视行为。DWA-RF模型的认知状态分类准确率达到了93.6%,验证了该模型在自动驾驶接管决策中驾驶员认知状态识别的有效性。

关键词: 驾驶员, 自动驾驶, 接管决策, 认知状态, 动态权重分配随机森林(DWA-RF)

Abstract:

In order to address the issue of human-machine collaboration failure in the takeover decision-making of autonomous driving and achieve precise recognition of the driver's cognitive state, this paper simulates three typical scenarios, night high-speed curves, mobile phone distracted driving, and combined scenarios of strong light and heavy rain, collecting and analyzing the dynamic interaction data of driving behavior and eye movement features construct a dynamic weight allocation(DWA)distribution feature fusion framework for multimodal perception and cognition collaboration, constructed DWA-RF model, and explore the dynamic mechanism of the driver's cognitive state and takeover decision-making behavior in complex environments. The results show that the distracted state significantly prolongs the takeover time. In scenes of strong light and heavy rain, the superimposition of distraction and environmental pressure leads to a sharp reduction in the range of scanning. Fatigue causes an increase in lane departure distance, accompanied by a reduction in pupil diameter and abnormal scanning behavior. The cognitive state classification accuracy of DWA-RF constructed in this paper has reached 93.6%, verifying the effectiveness of this model in identifying the driver's cognitive state for autonomous driving takeover decisions.

Key words: driver, autonomous driving, takeover decision, multimodal perception, dynamic weight allocation random forest (DWA-RF)

中图分类号:

X910安全人体学

邵舒羽, 李燕萍, 韩家琪. 自动驾驶接管决策中驾驶员认知状态识别研究^*[J]. 中国安全科学学报, 2026, 36(3): 255-263.

SHAO Shuyu, LI Yanping, HAN Jiaqi. Driver cognitive state recognition in autonomous driving takeover decision making[J]. China Safety Science Journal, 2026, 36(3): 255-263.

图/表 12

图1

表1

表2

表3

表4

表5

表6

图2

表7

图3

图4

表8

参考文献 15

[1]	王长帅, 徐铖铖, 邵永成, 等. 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. doi: 10.16265/j.cnki.issn1003-3033.2023.08.1474
	WANG Changshuai, XU Chengcheng, SHAO Yongcheng, et al. Analysis of takeover time of automated driving based on survival analysis[J]. China Safety Science Journal, 2023, 33(8): 142-148. doi: 10.16265/j.cnki.issn1003-3033.2023.08.1474
[2]	郭烈, 胥林立, 秦增科, 等. 自动驾驶接管影响因素分析与研究进展[J]. 交通运输系统工程与信息, 2022, 22(2): 72-90.
	GUO Lie, XU Linli, QIN Zengke, et al. Analysis and overview of influencing factors on autonomous driving takeover[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(2): 72-90.
[3]	吴付威, 杨慧, 马勇, 等. L3级自动驾驶车辆两阶段预警对接管绩效影响研究[J]. 交通运输系统工程与信息, 2025, 25(5): 40-49.
	WU Fuwei, YANG Hui, MA Yong, et al. Impact of two-stage warning on takeover performance of L3 autonomous vehicles[J]. Journal of Transportation Systems Engineering and Information Technology, 2025, 25(5): 40-49.
[4]	CHOI D, SATO T, ANDO T, et al. Effects of cognitive and visual loads on driving performance after take-over request (TOR) in automated driving[J]. Applied Ergonomics, 2020, 85:DOI: 10.1016/j.apergo.2020.103074.
[5]	罗霜, 易鑫鑫, 邵毅明, 等. 城市道路下分心行为对驾驶员驾驶绩效的影响[J]. 中国安全科学学报, 2024, 34(1): 70-76. doi: 10.16265/j.cnki.issn1003-3033.2024.01.0175
	LUO Shuang, YI Xinxin, SHAO Yiming, et al. Influences of distracted behaviors on driving performance of drivers in city[J]. China Safety Science Journal, 2024, 34(1): 70-76. doi: 10.16265/j.cnki.issn1003-3033.2024.01.0175
[6]	张宜静, 张志, 张洁, 等. 手动驾驶与接管情境相关认知能力对接管绩效的影响研究[J]. 载人航天, 2025, 31(4): 457-464.
	ZHANG Yijing, ZHANG Zhi, ZHANG Jie, et al. Research on influence of cognitive abilities related to manual driving and takeover scenarios on takeover performance[J]. Manned Spaceflight, 2025, 31(4): 457-464.
[7]	YU Zhenhua, ZHAI Linjing, JIANG Kang, et al. Impact of information-assisted modalities and levels on driver takeover performance in conditional automated driving[J]. Accident Analysis & Prevention, 2025, 220:DOI: 10.1016/j.aap.2025.108158.
[8]	LU Guangquan, ZHAI Junda, LI Penghui, et al. Measuring drivers' takeover performance in varying levels of automation: considering the influence of cognitive secondary task[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2021, 82: 96-110. doi: 10.1016/j.trf.2021.08.005
[9]	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.
[10]	WANG Ange, SHI Wenxin, HE Dengbo, et al. Impact of multi-dimensional cognitive demands on takeover performance, physiological and eye-tracking measures in conditionally automated vehicles[J]. Transportation Research Part F: Traffic Psychology and Behaviour, 2025, 114: 461-475. doi: 10.1016/j.trf.2025.06.011
[11]	宗学军, 王润鹏, 何戡, 等. 优化随机森林模型的工控网络异常检测[J]. 沈阳工业大学学报, 2024, 46(2): 197-205. doi: 10.7688/j.issn.1000-1646.2024.02.13
	ZONG Xuejun, WANG Runpeng, HE Kan, et al. Industrial control network abnormal detection based on optimized random forest model[J]. Journal of Shenyang University of Technology, 2024, 46(2): 197-205. doi: 10.7688/j.issn.1000-1646.2024.02.13
[12]	AZEVEDO-SA H, JAYARAMAN S, ESTERWOOD C, et al. Real-time estimation of drivers' trust in automated driving systems[J]. International Journal of Social Robotics, 2021, 13(8):1 911-1 927. doi: 10.1007/s12369-020-00694-1
[13]	CHEN Jiansong, ZHANG Qixiang, CHEN Jinxin, et al. A driving risk assessment framework considering driver' s fatigue state and distraction behavior[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(12): 20 120-20 136. doi: 10.1109/TITS.2024.3446832
[14]	任立海, 聂珍龙, 于潇, 等. 基于脑电信号的不良驾驶状态识别研究综述[J]. 中国公路学报, 2024, 37(8): 216-230. doi: 10.19721/j.cnki.1001-7372.2024.08.019
	REN Lihai, NIE Zhenlong, YU Xiao, et al. Review on recognition of unsatisfactory driving status based on electroencephalogram[J]. China Journal of Highway and Transport, 2025, 37(8): 216-230.
[15]	ZHAO Xia, LI Zhao, ZHAO Chen, et al. Distraction-level recognition based on stacking ensemble learning for IVIS secondary tasks[J]. Expert Systems with Applications, 2024, 244:DOI: 10.1016/j.eswa.2023.122849.

认知状态	接管时间/s	转向角速度/((°)·s^-1)	车道偏离/m	ACC跟车距离/m	预警响应延迟/s
警觉	1.12 ± 0.25	28.4 ± 4.2	0.18 ± 0.05	1.82 ± 0.35	0.45 ± 0.12
疲劳	1.58 ± 0.31^*	22.1 ± 3.8^*	0.32 ± 0.12^*	2.45 ± 0.41^*	0.62 ± 0.18^*
分心	1.82 ± 0.37^*	19.7 ± 4.5^*	0.42 ± 0.15^*	2.89 ± 0.53^*	0.71 ± 0.21^*

认知状态	接管时间/s	转向角速度/((°)·s^-1)	车道偏离/m	ACC跟车距离/m	预警响应延迟/s
警觉	0.95 ± 0.28	32.1 ± 5.1	0.15 ± 0.04	1.67 ± 0.32	0.38 ± 0.10
疲劳	1.32 ± 0.34^*	26.8 ± 4.9^*	0.38 ± 0.14^*	2.21 ± 0.47^*	0.55 ± 0.16^*
分心	1.67 ± 0.42^*	18.4 ± 4.1^*	0.45 ± 0.18^*	2.95 ± 0.58^*	0.82 ± 0.24^*

认知状态	接管时间/s	转向角速度/((°)·s^-1)	车道偏离/m	ACC跟车距离/m	预警响应延迟/s
警觉	1.05 ± 0.22	25.7 ± 4.7	0.12 ± 0.03	1.52 ± 0.28	0.48 ± 0.15
疲劳	1.45 ± 0.35^*	21.3 ± 4.3^*	0.28 ± 0.11^*	2.15 ± 0.39^*	0.68 ± 0.20^*
分心	1.72 ± 0.40^*	17.9 ± 4.0^*	0.35 ± 0.16^*	2.78 ± 0.45^*	0.95 ± 0.27^*

认知状态	平均注视时长/ms	瞳孔直径/ mm	扫视幅度/(°)	眨眼频率/ (次· min^-1)
警觉	250 ± 30	4.2 ± 0.5	3.2 ± 0.8	8 ± 2
疲劳	320 ± 40^*	3.8 ± 0.6^*	4.5 ± 1.2^*	15 ± 3^*
分心	180 ± 25^*	3.5 ± 0.4^*	2.1 ± 0.6^*	25 ± 4^*

认知状态	平均注视时长/ms	瞳孔直径/mm	扫视幅度/(°)	眨眼频率/ (次· min^-1)
警觉	220 ± 28	4.5 ± 0.6	4.8 ± 1.0	10 ± 3
疲劳	280 ± 35^*	4.1 ± 0.7^*	6.2 ± 1.5^*	18 ± 4^*
分心	150 ± 20^*	3.2 ± 0.5^*	1.8 ± 0.4^*	35 ± 5^*