首页|基于智能驾驶场景的人机信任影响因素及作用机制研究

基于智能驾驶场景的人机信任影响因素及作用机制研究

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
目的 研究智能驾驶场景下人机环系统中的三个重要因素(系统预测准确率、环境可认知状态和驾驶经验)对驾驶员智能驾驶辅助系统人机信任的影响.方法 采用三因素组内组间混合设计的人因学实验,招募了24 名不同驾驶经验的被试在智能驾驶辅助系统的三种预测准确率(90%、76%、60%)及两种环境可认知状态(清晰认知和模糊认知环境)下进行模拟驾驶任务,并收集驾驶过程中被试的主观人机信任、驾驶行为绩效和生理数据等多维指标,采用重复测量的方差分析处理实验数据.结果 不同驾驶经验下,被试的主观信任存在显著差异.系统预测准确率和驾驶经验对主观信任存在交互作用.不同环境可认知状态下,被试的瞳孔直径、眼跳次数和皮电波峰幅值均存在显著差异.环境可认知状态和驾驶经验对眼跳次数存在交互作用.人机环三因素对被试的驾驶行为绩效无明显影响.结论 系统预测准确率、环境可认知状态和驾驶经验均可在不同程度上影响主观人机信任及其相关生理指标.环境可认知状态较差时驾驶员会增加对驾驶辅助系统的信任.随着驾驶员驾驶经验的增加,其对驾驶辅助系统的信任也会增加.
The Mechanism of Influencing Factors of Human-Machine Trust in Intelligent Driving Scenes
Objective The effects of three factors from the human-machine-environment system(i.e.,system prediction accuracy,environmental cognitive status,and driving experience)on drivers'trust towards an intelligent driving assistance system are examined.Methods A three-factor mixed design was employed in this human factors experiment.A total of 24 participants with different driving experience were recruited to perform simulated driving tasks under three accuracy levels of an intelligent driving assistance system(i.e.,90%,76%,and 60%)and two en-vironmental cognitive states(i.e.,clear and ambiguous states).Multi-dimensional indicators such as subjective hu-man-machine trust,driving behavior performance and physiological indicators during the driving tasks were collect-ed.Repeated measures analysis of variance was used to examine the effects of the three independent factors on de-pendent variables.Results There were significant differences in subjective trust for participants with different driving experience.There was a significant interaction effect between system prediction accuracy and driving experience on subjective trust.There were significant differences in pupil diameter,number of saccades and skin conductance level in different environmental cognitive states.Environmental cognitive status and driving experience had significant interac-tion effects on mean number of saccades.The three factors had no significant impact on driving behavior performance.Conclusion The findings indicate that system prediction accuracy,environmental cognitive status,and driving expe-rience can affect subjective human-machine trust and its related physiological indicators to varying degrees.It ap-pears that,in poor environmental cognitive status or for more experienced drivers,drivers would trust more in the intel-ligent driving assistance system.

traffic engineeringhuman-vehide co-drivingintelligent drivingsystem prediction accuracyenvi-ronmental cognitive statusdriving experiencehuman-machine trusttraffrc safety

何文浩、王铁雁、张婷茹、陶达

展开 >

深圳大学 人因工程研究所,广东 深圳 518060

厦门市美亚柏科信息股份有限公司,厦门 361008

交通工程 人机共驾 智能驾驶 系统预测准确率 环境可认知状态 驾驶经验 人机信任 交通安全

国家自然科学基金项目广东省自然科学基金项目广东省自然科学基金项目深圳市基础研究面上项目深圳市基础研究面上项目

32271130240500006942023A1515012843JCYJ20230808105219038JCYJ20210324100014040

2024

10.13837/j.issn.1006-8309.2024.02.0002

人类工效学
中国人类工效学学会 安徽三联事故预防研究所

人类工效学

CSTPCD
影响因子：0.651
ISSN：1006－8309
年,卷(期)：2024.30(2)
  • 20