An Adaptive Driver Support System: User Experiences and Driving Performance in a Simulator
An Adaptive Driver Support System: User Experiences and Driving Performance in a Simulator
复制标题
DOI：
10.1177/0018720811430502
复制
发表时间：
2012-10-01
期刊：
影响因子：
3.3
通讯作者：
de Waard, Dick
中科院分区：
文献类型：
--
作者：
Dijksterhuis, Chris;Stuiver, Arjan;de Waard, Dick
目的：本研究的目的是测试自适应驾驶员支持系统的实施。背景：从安全角度来看，提供支持可能并不总是可取的，因为支持可能会导致与操作人员脱离相关的问题。相比之下，自适应支持系统旨在通过仅在必要时提供支持来尽可能让操作员了解情况。方法：共有 31 名经验丰富的驾驶员接受了三种车道保持支持模式：非自适应、自适应和无支持。支持涉及平视显示器上显示的不断更新的横向位置反馈。当适应性时，支持是由基于绩效的努力投资指示触发的。缩小车道宽度和增加迎面交通密度会增加转向需求，并且在所有情况下速度都是固定的，以防止任何补偿性速度反应。结果：参与者更喜欢自适应支持模式主要作为警告信号，并倾向于忽略非自适应反馈。此外，与无支持模式和非自适应支持模式相比，自适应支持改善了驾驶行为，因为当支持处于自适应模式时，参与者驾驶更加集中，显示出更少的横向变化，并且驾驶更少超出车道轮廓。结论：无论启动什么支持，人类操作员都可能使用机器触发的适应作为已经超过阈值的指示。因此，仅支持具有自适应自动化的人类信息处理系统的感觉处理阶段可能是不可行的。应用：这些结论与设计自适应驾驶员支持系统相关。
Objective: The aim of this study was to test the implementation of an adaptive driver support system.Background: Providing support might not always be desirable from a safety perspective, as support may lead to problems related to a human operator being out of the loop. In contrast, adaptive support systems are designed to keep the operator in the loop as much as possible by providing support only when necessary.Method: A total of 31 experienced drivers were exposed to three modes of lane-keeping support: nonadaptive, adaptive, and no support. Support involved continuously updated lateral position feedback shown on a head-up display. When adaptive, support was triggered by performance-based indications of effort investment. Narrowing lane width and increasing density of oncoming traffic served to increase steering demand, and speed was fixed in all conditions to prevent any compensatory speed reactions.Results: Participants preferred the adaptive support mode mainly as a warning signal and tended to ignore nonadaptive feedback. Furthermore, driving behavior was improved by adaptive support in that participants drove more centrally, displayed less lateral variation and drove less outside the lane's delineation when support was in the adaptive mode compared with both the no-support mode and the nonadaptive support mode.Conclusion: A human operator is likely to use machine-triggered adaptations as an indication that thresholds have been passed, regardless of the support that is initiated. Therefore supporting only the sensory processing stage of the human information processing system with adaptive automation may not feasible.Application: These conclusions are relevant for designing adaptive driver support systems.