Effects of adopting voice assistants countermeasures on alerting drowsy drivers across age.

Abstract

Drowsy driving is one of the leading contributing factors to traffic accidents. As drivers continue driving despite being aware of their drowsy states, it is necessary to adopt active countermeasures to avoid them falling asleep behind the wheel. Based on the multiple resources model, this study proposed five countermeasure types using voice assistants (no-countermeasure vs. meaninglessly listening vs. meaningfully listening vs. repeating vs. answering) and investigated their effects on countering drowsiness for young and middle-aged drivers. A five-week simulated manual driving experiment was conducted to evaluate the effectiveness of countermeasures based on driver performance, physiological indicators, eye movement indicators, and subjective ratings. Results indicated that repeating and answering were the two most effective countermeasures. The self-reported sleepiness of these two countermeasures (6.328 and 5.276 for repeating and answering, respectively) were significantly (p < 0.05) lower than that of the no-countermeasure (7.983). The physiological arousal state in term of skin conductance level for the two (1.798 and 1.990 for repeating and answering, respectively) were significantly (p < 0.05) higher than that of the no-countermeasure (1.022). Consistent patterns were also found in driving performance, eye movement indicators, psychomotor vigilance task accuracy, and subjective acceptance ratings. These findings suggested that countermeasures involving more resource stages can enhance drivers' alertness. The trend of higher alertness in the answering condition compared to the repeating condition highlighted the advantage of increasing cognitive load in countering drowsiness. Although young and middle-aged drowsy drivers did not differ on their average responses to different countermeasures, they showed differences in their physiological activation in reacting to the countermeasures. Drowsiness countermeasures should be tailored to fulfill individual needs, particularly for middle-aged drivers. This study not only clarifies how drowsiness countermeasures work but also offers practical guidance for vehicle designers on using in-vehicle voice assistants effectively.