基于信任理论的自动驾驶接管场景下HMI设计方法

doi:10.3969/j.issn.1674-8484.2026.02.013

摘要/Abstract

摘要：

为改善自动驾驶接管安全性，提出一种基于动态信任理论的人机交互界面（HMI）设计方法。将信任水平化分为信任不足、信任校准与过度信任3类，设计了分级情境感知透明度（SAT）界面，使事件锁时与关键时间窗口对齐，来记录分析眼动数据；通过调查问卷的形式开展主观评价。结果显示：经信任机理优化的界面，比原界面使驾驶员接管反应时间缩短26.71%提升关键时刻安全裕度；并使迟接管发生率由23.7%降至9.2%，降低潜在碰撞风险。研究结论可为智能座舱HMI的工程化应用提供方法。

关键词: 自动驾驶汽车, 人机交互界面（HMI）, 眼动追踪, 信任理论, 情境感知透明度（SAT）

Abstract:

A human-machine interface (HMI) interface design method was proposed based on the dynamic Trust Theory to improve the safety at autonomous driving takeovers. The trust levels were classified into three categories: the insufficient trust, the calibrated trust, and the over-trust. A hierarchical situation-aware transparency (SAT) interface was designed, and an empirical investigation was conducted in high-speed lane-changing scenarios. The eye-tracking data were recorded and analyzed by aligning event lock timing (ELT) with critical time windows (CTW), and the subjective evaluations were carried out through questionnaires. The results indicate that the interface optimized through the trust mechanism reduced drivers' takeover response time (RT) by 26.71% compared to the original interface, thereby enhancing safety margins during critical moments; it also decreased the incidence of delayed takeovers from 23.7% to 9.2%, reducing potential collision risks. Therefore, these results would provide a methodological reference for the engineering application of intelligent cockpit HMIs.

Key words: autonomous vehicles, human-machine interface (HMI), eye tracking, Trust Theory, situation awareness transparency (SAT)

中图分类号:

薛清元, 瞿珏, 王崴, 牛天林, 李幸. 基于信任理论的自动驾驶接管场景下HMI设计方法[J]. 汽车安全与节能学报, 2026, 17(2): 270-277.

XUE Qingyuan, QU Jue, WANG Wei, NIU Tianlin, LI Xing. Human-machine interface design methods in autonomous driving takeover scenarios based on the Trust Theory[J]. Journal of Automotive Safety and Energy, 2026, 17(2): 270-277.

图/表 15

图1 自动驾驶车辆界面示意图

图2 信任水平与系统能力的匹配关系

图3 数据采集设备

图4 并道场景

图5 实验流程图

信任组别	RT / ms
信任组别	均值	标准差	中位数	四分位数
1	393.0	227.6	405.0	279.6~430.8
2	894.4	126.7	862.8	813.6~968.4
3	5 004	1 448	3 972	1 735~6 036

表1 接管反应时间

图6 信任组接管时间对比

图7 情景感知透明度(SAT)分级界面

图8 接管时间窗口下的注视轨迹图

图9 注视时长的热点图

图10 3类信任组中控界面AOI注视比例对比

图11 脑力需求与绩效评分箱线图

图12 主观信任度评分分布图

图13 三类信任组注视平均持续时间对比结果

AOI	注视频次，F		注视时间比例，γ
AOI	(M±SD)	95%CI	(M±SD)	95%CI
前方车道	92.3±118.1	[70.9, 113.7]	62.3±11.5	[59.0, 65.6]
目标车道	42.0±64.8	[21.4, 62.6]	18.7±8.2	[16.1, 21.3]
接管提示窗口	31.1±41.6	[17.9, 44.3]	11.4±6.9	[9.3, 13.5]
信任度弹窗	12.4±11.8	[8.6, 16.2]	7.6±4.8	[6.1, 9.1]

表2 不同兴趣区(AOI)下的注视频次与注视时间比例(N = 38)

参考文献 21

[1]	谭征宇, 张瑞佛, 刘芝孜, 等. 智能网联汽车人机交互信任研究现状与展望[J]. 机械工程学报, 2024, 60(10): 366-383.
	TAN Zhengyu, ZHANG Ruifo, LIU Zhizi, et al. Human-machine interaction trust in intelligent and connected vehicles: overview and perspectives[J]. *J Mech Engi*, 2024, 60(10): 366-383. (in Chinese)
[2]	SAE International. Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles: J3016[S]. Warrendale (PA): SAE International, 2018.
[3]	李洋, 杨智元, 陈晓龙. 智能汽车人机协同决策关键技术综述[J]. 中国公路学报, 2024, 37(3): 98-116.
	LI Yang, YANG Zhiyuan, CHEN Xiaolong, et al. A survey on the human-machine collaborative decision making in intelligent vehicles[J]. *Chin J Highw Transport*, 2024, 37(3): 98-116. (in Chinese)
[4]	Lee J D, See K A. Trust calibration for adaptive automation[J]. *Huma Fact*, 2019, 61(5): 728-743.
[5]	CHEN Jiexi, Barnes M J. Human-agent teaming for multirobot control: A review of human factors issues[J]. *IEEE Trans Human-Mach Syst*, 2014, 44(1): 13-29.
[6]	Hergeth S, Lorenz L, Vilimek R, et al. Keep your scanners peeled: gaze behavior as a measure of automation trust during highly automated driving[J]. *Huma Fact*, 2016, 58(3): 509-519.
[7]	高在峰, 李文敏, 梁佳文, 等. 自动驾驶车中的人机信任[J]. 心理科学进展, 2021, 29(12): 2172-2183. doi: 10.3724/SP.J.1042.2021.02172
	GAO Zaifeng, LI Wenmin, LIANG Jiawen, et al. Trust in automated vehicles[J]. *Advan Psych Sci*, 2021, 29(12): 2172-2183. (in Chinese)
[8]	解煜彬, 周荣刚. 新型人机关系下的人机双向信任[J]. 心理科学进展, 2025, 33(6): 916-932. doi: 10.3724/SP.J.1042.2025.0916
	XIE Yubin, ZHOU Ronggang. The bidirectional trust in the context of new human-machine relationships[J]. *Advan Psych Sci*, 2025, 33(6): 916-932. (in Chinese)
[9]	Muir B M. Trust between humans and machines, and the design of decision aids[J]. *Int’l J Man-Mach Stud*, 1987, 27(5-6): 527-539.
[10]	Parasuraman R, Riley V. Humans and automation: use, misuse, disuse, abuse[J]. *Huma Fact*, 1997, 39(2): 230-253.
[11]	Kim S, van Egmond R, Happee R. How manoeuvre information via auditory (spatial and beep) and visual UI can enhance trust and acceptance in automated driving[J]. Transport Res F: *Traf Psych Behav, 2024, 100*: 22-36. doi: 10.1016/j.trf.2023.11.007 URL
[12]	Beggiato M, Krems J F. The evolution of mental model, trust, and acceptance of adaptive cruise control in relation to initial information[J]. Transport Res F: *Traf Psych Behav*, 2013, 18: 47-57. doi: 10.1016/j.trf.2012.12.006 URL
[13]	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]. *J Transport Syst Engi Info Tech*, 2022, 22(2): 72-90. (in Chinese)
[14]	曹莉凌, 刘君丽, 金升烨, 等. 面向自动驾驶的远程多维信息实时交互系统设计[J]. 汽车安全与节能学报, 2024, 15(6): 934-942.
	CAO Liling, LIU Junli, JIN Shengye, et al. Design of a remote multidimensional information real-time interactive system for automated driving[J]. *J Autom Safe Energ*, 2024, 15(6): 934-942. (in Chinese)
[15]	陈发城, 鲁光泉, 林庆峰, 等. 有条件自动驾驶下驾驶人接管行为综述[J]. 吉林大学学报(工学版), 2025, 55(2): 419-433.
	CHEN Facheng, LU Guangquan, LIN Qingfeng, et al. Review of drivers’ takeover behavior under conditional automated driving[J]. J Jilin Univ (Engi Tech Ed), 2025, 55(2): 419-433. (in Chinese)
[16]	Hoff K A, Bashir M. Trust in automation: theoretical framework for interpersonal and human-automation trust[C]// Proc Human Factors Ergonomics Soc Annu Meeting. Chicago, IL: HFES, 2014: 268-272.
[17]	瞿广跃, 杨澜, 袁梦, 等. 面向自动驾驶汽车的信号交叉口行人多模态轨迹预测方法[J]. 汽车安全与节能学报, 2024, 15(5): 689-701.
	QU Guangyue, YANG Lan, YUAN Meng, et al. Multimodal pedestrian trajectory prediction method at signalized intersections for autonomous vehicles[J]. *J Autom Safe Energ*, 2024, 15(5): 689-701. (in Chinese)
[18]	LI Jue, LIU Jiawen, WANG Xiaoshan, et al. The impact of transparency on driver trust and reliance in highly automated driving: presenting appropriate transparency in automotive HMI[J]. *Appl Sci*, 2024, 14(8): 32-33.
[19]	Becker F, Flemisch O, Winner H. Does trust influence the driver’s response to automated vehicle failures?[J]. *IEEE Trans Human-Mach Syst*, 2022, 52(6): 1119-1128.
[20]	HU Chuan, HUANG Siwei, ZHOU Yu, et al. Dynamic and quantitative trust modeling and real-time estimation in human-machine co-driving process[J]. Transport Res F: Traf Psych Behav, 2024, 106: 306-327.
[21]	Jouhri A, Sharkawy A, Paksoy H, et al. The influence of a color themed HMI on trust and take-over performance in automated vehicles[J]. *Front Psychol*, 2023, 14: 25-28.