Driving Features Extraction between Skilled and Unskilled Drivers’ Groups

Abstract

    One way to improve drivers’ skill is to provide effective advice for the inexperienced driver on maneuvers. So it is important to understand the difference between skilled and inexperienced drivers beforehand. Such critical difference is expected to appear in accelerator, brake, and steering operations accompanying the state of vehicle. This paper proposes a statistical method to extract these differences from large amounts of candidate features derived from driving operation data. And then, a basic experiment with a driving simulator is carried out to verify the method.