This is the authors' abstract. We don't add key points for this paper.
Model predictive control (MPC) has been widely applied to depth tracking of autonomous underwater vehicles (AUVs). However, actuator stuck failures compromise the system’s ability to track the target reference, which may become inadmissible under the system constraints. As a consequence, actuator saturation may occur, potentially leading to the loss of stability and recursive feasibility. Simply modifying the associated input constraint in MPC to cope with the stuck actuator is insufficient to resolve the above issue. To address this challenge, we investigate a fault-tolerant model predictive control method with reconfigurable reference (FTMPC-RR) for AUV depth tracking under actuator stuck failures. Specifically, we design an invariant set for tracking to ensure closed-loop stability, whose computation requires only the identification of the stuck actuator, without relying on the stuck position. In the event that the target reference is inadmissible, the reconfigurable reference takes effect. The FTMPC-RR then generates a reconfigured reference that is admissible under the fault condition, ensuring the recursive feasibility. The proposed method is evaluated using a linear longitudinal model of the ARIES AUV, and its performance is improved compared to that of conventional FTMPC in the presence of actuator stuck failures.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Chen et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: