80 MPH and out-of-the-loop: Effects of real-world semi-automated driving on driver workload and arousal

Abstract

The introduction of semi-automated driving systems is expected to mitigate the safety consequences of human error. Observational findings suggest that relinquishing control of vehicle operational control to assistance systems might diminish driver engagement in the driving task, by reducing levels of arousal. In this study, drivers drove a Tesla Model S with Autopilot in both semi-automated and manual modes. Driver behavior was monitored using a combination of physiological and behavioral measures. Compared to manual driving, a reduction in driver physiological activation was observed during semi-automated driving. Also, performance to the peripheral detection task suffered in semi-automated mode, with slower response times recorded in this condition than during manual driving. Taken together, our data suggest that semi-automated driving might not ease safety consequences of human error. Instead, these findings suggest it might cause a drop in driver monitoring, possibly followed by a spike in automation-generated distraction.