Experimental evaluation of assistance system to predict driving behavior considering pre-preceding vehicle information

Abstract

This study examines a driver assistance system to predict driving behavior considering the pre-preceding vehicle behavior. The authors have defined an anticipation index called PRE3 and proposed a driving assistance system that indicates the PRE3 to the driver in real-time during a car-following. The proposed system indicates the evaluation index considering the relations not only between the preceding and following vehicles, but also between the pre-preceding and following vehicles. This paper aims to evaluate the performance of PRE3 by a test track field experiment that employs three passenger cars as the test vehicles. The test cars are equipped with various sensors and communication system including the camera and the target for headway distance measurement and the CAN data of the vehicles. Five drivers which participated in the experiment are instructed to follow the preceding and the visible pre-preceding vehicles with and without the driving assistance system. The field test showed that the proposed system is succeeded in reducing the relative velocity with the pre-preceding vehicle. Also, the positive high acceleration rate is infrequent when installing the PRE3 assistant system. It was concluded that the proposed assistance system is helpful to reduce the relative velocity with the pre-preceding vehicle, unnecessary positive acceleration and the variance of collision risk to the preceding vehicle.