Obstacle avoidance trajectory planning based on iterative optimization and integrated tracking control for intelligent vehicles

To address the challenges of low computational efficiency, poor trajectory smoothness, and delayed adjustments in intelligent vehicle obstacle avoidance, an improved trajectory planning algorithm is proposed which integrates iterative optimization of path and speed. In addition, an adaptive robust model predictive control (RMPC) controller accounting for the lateral and longitudinal coupling dynamics of the vehicle is designed to enhance the adaptability and robustness of the trajectory tracking system. For trajectory planning, the evaluation system is built to provide an initial path and speed profile across the entire space. The rough solution serves as a reference for refining a smooth trajectory that is constrained by vehicle dynamics within a convex space. Iterative optimization of path and speed is performed in each cycle to enable obstacle avoidance in time. In the aspect of trajectory tracking, the RMPC controller aims to minimize path deviations and control increments within the prediction horizon. The objective function is framed as a min-max problem and solved optimally using the linear matrix inequality (LMI) method. To further enhance tracking accuracy and system adaptability, a fuzzy-based strategy is used to adjust the weight coefficient matrix. Combination simulations for both trajectory planning and tracking are carried out under static and dynamic obstacle avoidance conditions. The results demonstrate that the vehicle successfully follows safe and comfortable obstacle avoidance paths, maintaining a small tracking error and high driving stability throughout the process.