A Survey of Driver Behavior Perception Methods for Human-Computer Hybrid Enhancement of Intelligent Driving

Yi Jiwei ¹ , Du Aimin ¹ , Zhu Zhongpan ² , Ding Hongjun ¹

1. School of Automotive Studies , Tongji University

2. School of Electronic and Information Engineering , Tongji University

Abstract: The subjective uncertainty of human-in-the-loop is a core issue for hybrid enhancement of human-computer shared driving that is the research hotspot of intelligent vehicle studies.Precise perception of driver behavior is a prerequisite and a significant method to break through the problem of human-in-the-loop uncertainties.Through literature comparison and analysis, we find that the objective evaluation method based on contact and non-contact sensors has received more attention from scholars in comparison with the subjective evaluation of driver behavior.However, there is no literature that addresses the new challenges of driver behavior perception under the condition of human-computer shared driving.Therefore the methods of driver behavior perception for hybrid enhancement of human-computer shared driving are summarized in this paper and the driver behavior perception based on visual and tactile multi-sensor fusion has been pointed out as the future research direction.

Key words: driver behavior, human-computer shared driving, perception, neural network 7IkdAox6sKKVLxG1uNZQ6pnK8+ucTLHUefipUY/Vfwuqha4SOvvGCX/fBSOJrosc