• Physics-based and empirical modeling of the hydraulic propulsion system. • A practical, model-independent fault diagnosis strategy and a dual-adaptive fault-tolerant control framework. • Comprehensive validation via multi-scenario simulation tests and pool experiments. Deep-sea mining operations subject marine mining vehicles, such as Remotely Operated Vehicles (ROVs), to unknown external disturbances and potential actuator failures, thereby compromising operational safety. To ensure reliable trajectory tracking under such conditions, this study presents a comprehensive control scheme for a hydraulic ROV. First, a physics-based model of the throttle-controlled, speed-regulated propulsion system is established, explicitly incorporating load–pressure dynamic coupling. Building on this model, a practical threshold-based Fault Diagnosis (FD) strategy is developed to isolate faulty thrusters via motion-deviation analysis, without requiring a precise model. Subsequently, a backstepping Fault-Tolerant Control (FTC) law is designed. Distinct from conventional adaptive schemes that often lump faults with system uncertainties, the proposed method employs dual adaptive estimators to reconstruct actuator faults and compensate for model uncertainties. Validated via simulations of linear, segmented, and figure-eight trajectories, as well as pool experiments, the proposed scheme demonstrates superior tracking performance compared with a conventional Fuzzy Proportional-Integral-Derivative (FPID) controller.
Liang et al. (Tue,) studied this question.