Dynamic Obstacle Avoidance Strategy for High-Speed Vehicles Via Constrained Model Predictive Control and Improved Artificial Potential Field

Abstract

The paper proposes an automatic planning and control strategy to keep the automatic driving high-speed vehicles collision-free based on the improved artificial potential field (APF) and constrained model predictive control (MPC). First, the road potential field satisfying the constraints of vehicle safe driving is constructed according to the driving characteristics of vehicles and road boundary conditions. Focusing on the influence of the speed and direction of the obstacle vehicle on the potential field, the dynamic obstacle vehicle potential field is established. Secondly, the road potential field and the dynamic obstacle vehicle potential field are incorporated into the objective function of the path planning module to establish a real-time path planner. Thirdly, the yaw stability constraints of the vehicle are established, which are added to the QP solver and updated in real-time according to the current vehicle states, so as to establish the constrained MPC controller. In the end, the safety of this planning and control strategy for obstacle avoidance overtaking and the effectiveness of constrained MPC in improving vehicle stability are verified by comparative simulation analysis in multi-obstacle vehicles scenarios.