Quality of control takeover following disengagements in semi-automated vehicles

This study examines drivers' responses to automated driving technology failures in semi-automated vehicles. The study is executed in a human-in-the-loop setting, within a high-fidelity integrated car simulator capable of handling both manual and automated driving. A population of 40 individuals was tested, with metrics for control takeover quantification given by vehicle maximum lane offset with respect to the lane centerline, and integral offset over an S-curve turn to compare performance during recovery after automation technology failure and conventional manual driving. Independent variables for the study are the age of the driver, the speed at the time of disengagement, and time at which the disengagement occurs (i.e., how long automation was engaged for). Overall results show better performance for lower speed settings with maximum drift increasing over 116% for higher speed settings. Moreover, contrary to expectations, the age factor was found to be not statistically significant for drift performance, with older participants performing better than or equal to the younger age groups. The time dependency observed was non-linear, and the results highlighted the need for additional testing for this variable. A detailed analysis of variance, as well as preliminary recommendations for regulatory operational limitations, are presented. (C) 2019 Elsevier Ltd. All rights reserved.