A Novel Iterative Learning-Model Predictive Control Algorithm for Accurate Path Tracking of Articulated Steering Vehicles

Xuanwei Chen, Changlin Yang, Jiaqi Cheng, Huosheng Hu, Guifang Shao, Yunlong Gao, Qingyuan Zhu

Abstract: Refining the path tracking of articulated steering vehicles amidst terrain disturbances presents a formidable challenge. While model predictive control (MPC) offers promise in tackling this issue, its efficacy is often hindered by model intricacies and inaccuracies. In this communication, an innovative approach termed iterative learning-model predictive control (IL-MPC) is introduced to enhance path tracking performance on rugged terrains. Initially, an MPC controller grounded in a simplified kinematic model is established to ensure stability in path tracking. Subsequently, an iterative learning algorithm is integrated to meticulously capture and mitigate MPC controller errors. A comprehensive feedforward-feedback framework coupled with a spatial indexing method is proposed to synergize the strengths of iterative learning and MPC. Through rigorous evaluations across diverse paths and terrains, the method demonstrates robustness against terrain disturbances, affirming its efficacy in real-world scenarios.