Highly automated driving: the role of visuo-attentional and executive abilities in take-over success

Previous studies have sought to better understand exogenous factors explaining drivers' ability to regain control or not from highly automated driving. However, few studies have examined the role of individual factors in drivers' take-over behavior and performance. The present study sought to examine the extent to which take-over performance can be predicted by individual differences in visuo-attentional and executive abilities. After a period of automated driving on a simulator, participants aged 20 to 60 (<i>N</i> = 118) had to regain control of their vehicle in a critical take-over situation (i.e., avoiding an obstacle by changing lanes while other vehicles are approaching at higher speeds). The take-over manoeuvre was considered successful if the participants managed to complete it without colliding with the obstacle or another vehicle. All participants completed a battery of cognitive (working memory, inhibitory control, attentional flexibility, and planning) and visuo-attentional tests (visuomanual coordination, multiple object avoidance, and multiple object tracking), and reported their age and driving experience. Several partial least-squares models predicting the success of the manoeuvre from individual abilities were compared. The most accurate model had an accuracy of 70.79% and identified spatial working memory (measured with the CORSI task), visuomanual coordination (one's ability to manually track a moving target) and driving experience (annual mileage) as key factors for the success of the take-over. Conversely, higher inhibitory control ability, as measured by the Flanker and Stop-Signal tasks, was negatively related to take-over success, possibly because these participants exerted strong cognitive control over the non-driving task during automated driving, which came at the cost of flexibility.