REViVe:开发一种反应性环境警觉车载系统以减轻自动驾驶中困倦导致的注意力不集中

REViVe: Development of a reactive environmental vigilance in-vehicle system to mitigate drowsiness-induced inattention during automated driving

Accident Analysis & Prevention · 2025
被引 5
ABS 3

中文导读

研究开发了一种利用嗅觉和气候等替代感官通道的车载警觉系统(REViVe),在单调或条件自动驾驶中帮助困倦驾驶员快速恢复清醒并维持警觉,同时避免传统视听警报的疲劳问题。

Abstract

With monotonous or conditionally automated driving conditions that may lead to the degradation of driver vigilance and increase the risk of drowsy driving, it is crucial to implement advanced systems that assist drivers in returning to a state of optimal driving readiness. While these systems have shown significant effects in reducing the risks related to drowsy driving, most warning systems heavily rely on auditory and visual sensory channels. These modalities are susceptible to "alarm fatigue" due to frequent and annoying alarms, which may lead drivers to ignore or deactivate the systems entirely, thus rendering them less suitable for preemptive cautionary warnings. To address these limitations, a Reactive Environmental Vigilance in-Vehicle (REViVe) system was developed to counteract driver drowsiness by utilizing alternative sensory modalities. A total of 35 drivers were divided into three condition groups: olfactory, climate, and control. To evaluate the effectiveness of the system, five dependent measurements were analyzed: time taken for PERCLOS to return to baseline and engagement index to measure salient effect; time interval between drowsiness events and peripheral detection task score difference to measure sustained arousal effect; and satisfaction rating to measure driver acceptability. Both the olfactory and climate REViVe systems showed potential as effective preemptive warnings compared to control. With REViVe, drowsy drivers quickly returned to an awake state and sustained vigilance significantly longer than control, while driver satisfaction was positive. Thus, the REViVe system provides a balanced solution for alert functionality and driving experience, suggesting a novel approach to designing preemptive warning systems.

自动驾驶驾驶员警觉性困倦驾驶车载预警系统人机交互