Cognition, Languages, Language and Ergonomics (CLLE) laboratory, University of Toulouse - Jean Jaurès, Toulouse, France.
Toulouse Computer Science Research Institute (IRIT), Paul Sabatier University, Toulouse, France.
Accid Anal Prev. 2024 May;199:107512. doi: 10.1016/j.aap.2024.107512. Epub 2024 Feb 20.
In the context of high-level driving automation (SAE levels 4-5), several studies have shown that a personalized automated driving style, i.e., mimicking that of the human behind the wheel, can improve his experience. The objective of this simulator study was to examine the potential transfer of these benefits in the context of intermediate-level driving automation (SAE levels 2-3), focusing on driving speed personalization. In the first phase of the study, the driving speed of 52 participants was recorded. In the second phase, the same participants were driven by an automated car on a highway twice, and sometimes had to takeover during the drive because of a stationary vehicle on the lane. On these two drives, the automated car drove either at the same speed as them (personalized) or 20 km/h faster. The results showed that using a personalized speed driving style led to higher comfort, and that this effect was fully mediated by automated driving perceived safety. Although driving speed predicted automated driving perceived safety, this effect was actually moderated by trust in automated cars. Regarding takeover performance, the results showed that the brake use and maximum force were lower with the personalized speed driving style, leading to lower resulting maximum negative longitudinal acceleration and speed variability. Overall, the results of this study suggest that the benefits of automated driving style personalization in terms of speed extend to SAE levels 2-3. In addition to the experience benefits, this personalization approach could also improve traffic flow and safety.
在高级驾驶自动化(SAE 级别 4-5)的背景下,多项研究表明,个性化的自动驾驶风格,即模仿人类驾驶员的驾驶风格,可以提高驾驶员的体验。本模拟器研究的目的是检验这些益处在中级驾驶自动化(SAE 级别 2-3)背景下的潜在转移,重点是驾驶速度个性化。在研究的第一阶段,记录了 52 名参与者的驾驶速度。在第二阶段,同一批参与者两次被自动驾驶汽车在高速公路上驾驶,有时因为车道上的静止车辆而需要接管。在这两次驾驶中,自动驾驶汽车要么以与他们相同的速度(个性化)行驶,要么以 20 公里/小时的速度更快行驶。结果表明,采用个性化的速度驾驶风格会提高舒适度,而这种效果完全由自动驾驶感知安全介导。尽管驾驶速度预测了自动驾驶感知安全,但这种效果实际上受到对自动驾驶汽车的信任的调节。关于接管性能,结果表明,个性化速度驾驶风格的制动使用和最大力较低,导致最大负纵向加速度和速度变化较小。总的来说,这项研究的结果表明,在速度方面,自动驾驶风格个性化的益处可以扩展到 SAE 级别 2-3。除了体验方面的好处,这种个性化方法还可以改善交通流量和安全性。