School of Industrial Engineering, Purdue University, West Lafayette, IN, United States.
Ford Motor Company, Dearborn, MI, United States.
Accid Anal Prev. 2021 Jul;157:106143. doi: 10.1016/j.aap.2021.106143. Epub 2021 May 16.
Automated driving systems are becoming increasingly prevalent throughout society. In conditionally automated vehicles, drivers may engage in non-driving-related tasks (NDRTs), which can negatively affect their situation awareness (SA) and preparedness to resume control of the vehicle, when necessary. Previous work has investigated engagement in NDRTs, but questions remain unanswered regarding its effect on drivers' SA during a takeover event. The objective of the current study is to use eye-tracking to aid in understanding how visual engagement in NDRTs affects changes in SA of the driving environment after a takeover request (TOR) has been issued. Thirty participants rode in a simulated SAE Level 3 automated driving environment and engaged in three separate pre-TOR tasks (Surrogate Reference Task, Monitoring Task, and Peripheral Detection Task) until presented with a TOR. Situation Awareness Global Assessment Technique (SAGAT) scores and gaze behavior were recorded during the post-TOR segment. Overall, longer times spent viewing the driving scene, and more dispersed visual attention allocation, were observed to be associated with better overall SA. Also, location-based eye tracking metrics show most promise in differentiating between task conditions with significantly different SAGAT scores. Findings from this work can inform the development of real-time SA assessment techniques using eye movements and ultimately contribute to improved operator roadway awareness for next-generation automated transportation.
自动驾驶系统在社会中越来越普及。在条件自动化车辆中，驾驶员可能会从事与驾驶无关的任务（NDRTs），这可能会降低他们的情境意识（SA）和对车辆进行恢复控制的准备，在必要时。以前的工作已经研究了从事 NDRTs 的情况，但对于其对接管事件期间驾驶员 SA 的影响，仍有一些问题尚未得到解答。本研究的目的是使用眼动追踪来帮助理解在发出接管请求（TOR）后，视觉参与 NDRTs 如何影响驾驶环境的 SA 变化。三十名参与者在模拟 SAE 级别 3 自动驾驶环境中行驶，并在接到 TOR 之前进行了三个单独的预 TOR 任务（替代参考任务、监控任务和周边检测任务）。在 TOR 后段记录了情境意识全局评估技术（SAGAT）得分和注视行为。总体而言，观察到花在查看驾驶场景上的时间更长，视觉注意力分配更分散，与整体 SA 更好相关。此外，基于位置的眼动追踪指标在区分具有显著不同 SAGAT 得分的任务条件方面显示出最大的潜力。这项工作的结果可以为使用眼动追踪来开发实时 SA 评估技术提供信息，并最终有助于提高下一代自动化交通的操作人员对道路的认识。
