Obstacle Avoidance Method for Autonomous Vehicles Based on Quadratic Programming and Model Predictive Control

This paper proposes a hierarchical framework for path planning and tracking control in autonomous vehicles, ensuring safe navigation and obstacle avoidance. The planning module combines dynamic programming and quadratic programming to generate smooth, feasible paths. The control module introduces weight-adaptive model predictive control (WAMPC) to improve responsiveness to dynamic changes and enhance tracking precision. Compared to traditional model predictive control, WAMPC boosts trajectory accuracy in complex environments and improves ride comfort. A co-simulation with PreScan, Car-Sim, and Matlab/Simulink was used for experiments under two road conditions. The results confirm the framework’s effectiveness and demonstrate WAMPC’s superior tracking performance.