긴급 충돌 회피를 위한 비선형 모델 예측 제어 기반 조향 및 브레이크 통합 제어

Abstract

As the demand for active safety systems for vehicles increases, various driving assistance technologies have been developed for driving safety. In particular, collision avoidance systems that use braking and active steering to prevent vehicles from collapsing in emergencies are studied intensively. However, previous studies have a limitation in that the vehicle is controlled to follow a pre-determined collision-free path where vehicle dynamics are not considered, which stops the vehicle from making the most of the collision avoidance ability. Thus, this study presents an integrated steer and brake controller that optimizes the control input considering the vehicle dynamics for collision avoidance situations without a pre-determined collision-free path. A nonlinear model predictive control is utilized to handle the nonlinear vehicle model and optimize the control input for collision avoidance, satisfying constraints that guarantee vehicle stability. The proposed controller was verified through experiments and showed improved results in avoiding collision compared to other methods.