Evaluation of A Semi-Autonomous Lane Departure Correction System Using Naturalistic Driving Data

摘要

Evaluating the effectiveness and benefits of driver assistance systems isessential for improving the system performance. In this paper, we propose anefficient evaluation method for a semi-autonomous lane departure correctionsystem. To achieve this, we apply a bounded Gaussian mixture model to describedrivers' stochastic lane departure behavior learned from naturalistic drivingdata, which can regenerate departure behaviors to evaluate the lane departurecorrection system. In the stochastic lane departure model, we conduct adimension reduction to reduce the computation cost. Finally, to show theadvantages of our proposed evaluation approach, we compare steering systemswith and without lane departure assistance based on the stochastic lanedeparture model. The simulation results show that the proposed method caneffectively evaluate the lane departure correction system.