Accelerated evaluation of automated vehicles using extracted naturalistic driving data

ABSTRACT: It is important to rigorously and comprehensively evaluate the safety performance of Automated Vehicles before they are produced and deployed. Under naturalistic driving conditions, field operational tests and simulations may take a long time to finish, because of the low exposure to safety-critical scenarios. In this paper, we propose an accelerated evaluation approach. Statistics of the motion of the primary other vehicle (POV) were built using extracted naturalistic driving data then modified to present higher risk interactions. Two surrogate automated vehicles obtained from observed production vehicle behaviors were evaluated in car-following scenarios. Results show that the proposed method can accelerate the evaluation process by 5 orders of magnitude.