A Study on Driving Safety Evaluation Methods for Minimal Risk Maneuver of Automated Driving System; 자율주행시스템의 위험 최소화 기동에 대한 운행 안전성 평가 방법에 관한 연구

초록

This paper proposes a framework for evaluating operational safety in DDT fallback situations that may occur during autonomous driving. A safety assessment algorithm for MRM is also presented. The framework outlines the evaluation requirements for DDT fallback, scenario design, experimental methods, and analysis techniques. The safety assessment algorithm for MRM validates vehicle safety rates and proposes evaluation metrics for collision risks in both front and rear of the vehicle, based on predicted collision time and intensity. Simulations are conducted to differentiate use when stopping within the lane and stopping in a safe zone under unified DDT fallback conditions, with each MRM timing evaluated. The difference in front collision risk for stopping within the lane at various MRM timings was found to be approximately 17 %, while the rear collision safety rate was approximately 99.2 %. When stopping in a safe zone, the difference in front vehicle collision risk due to lane change attempts was approximately 65 %, with a rear collision safety rate of about 98 %. This confirms the validity of the proposed algorithm. Future research will focus on enhancing operational safety evaluation methods by establishing the criteria for assessing failures in autonomous driving systems and combinations of conditions leading to deviations from ODD. © © 2024 KSAE.

키워드