Nonlinear Model Predictive Control for Path Following of AUVs: A Multiobjective Approach

This is the authors' abstract. We don't add key points for this paper.

This paper addresses the path following problem of autonomous underwater vehicles (AUVs) under input and state constraints, propose a novel multiobjective model predictive optimization control scheme. First, by treating path parameters as decision variables, it introduces an innovative real-time optimization of reference states, ensuring both accuracy and efficiency in path tracking. Second, to address varying task priorities, a constraint weighted sum (CWS) method is developed. This approach effectively resolves priority conflicts within the multiobjective model predictive control (MOMPC) framework, significantly improving control performance and flexibility for AUVs in complex environments. These advancements offer new technical support for both theoretical research and practical applications in AUVs path tracking. Then, through rigorous theoretical analysis, the feasibility and stability of the MOMPC path following control algorithm are demonstrated. Finally, simulation studies on the AUVs validate the effectiveness of the proposed MOMPC path following control method.