Predictive Braking Assistance System for Intersection Safety Based on Risk Potential

This paper describes the assessment of a predictive braking assistance system by using a driving simulator which can reconstruct near-miss incident scenario relevant to pedestrians. A braking assistance system algorithm for collision avoidance is designed based on pedestrian motion prediction and risk potential. A virtual spring connected between the vehicle and the pedestrian is used to determine the potential field. The feasibility of the proposed algorithm is examined by experiments using the driving simulator and a comparison with actual expert driving data. Near-miss incident data relevant to pedestrians in intersections collected by image-captured drive recorders are analyzed to get the basic parameters of pedestrian relevant crash scenario model. The effectiveness of the proposed system is verified by driving simulator experiments.