Study on psychosomatic features of driver's distraction based on pattern recognition

Abstract

Driver's distraction is said as one of major psychosomatic factors which may prone to be involved in traffic accidents. Therefore it is much expected to prevent traffic accidents by means of incorporating a function of driver state monitoring into a driving support system. In our previous study, we identified the methodology for detecting driver's cognitive distraction in higher degree of accuracy by means of using the AdaBoost as pattern recognition algorithm. We used visual features such as gaze direction and head rotation angle, pupil diameter and heart rate RRI as physiological features. In this study we aimed at clarifying the degree of influence for detecting driver's distraction by four recognition features, which are gaze angle and head rotation angle (hereinafter; vision information), pupil diameter and the interval between heart R-waves (hereinafter; heart rate RRI) in order to create a driver's state monitoring function.