Predefined-time sliding mode based fault tolerant path-following control for underactuated UUV with actuator faults and input saturation

A fault-tolerant control (FTC) scheme combining predefined-time sliding mode and quadratic programing algorithm is proposed to achieve path-following of unmanned underwater vehicle (UUV) subject to actuator faults and input saturation. Initially, line of sight method based on fixed-time predictors is used to construct path-following guidance law, incorporating a predefined-time sliding mode surface to design the path-following controller, which effectively accelerates the system convergence speed. Subsequently, a neural network is trained for real-time actuator fault diagnosis, enabling prompt acquisition of fault parameters. These parameters are then utilized in a quadratic programing thrust redistribution system to compensate for the failed thrust while also addressing input saturation constraint. By combining predefined-time sliding mode control with quadratic programing, this approach balances error convergence time, control accuracy, and fault tolerance within the UUV fault tolerant control system. Through the application of Lyapunov functions, the study establishes the system's uniform global fixed-time stability. Numerical simulations demonstrate that the devised controller adeptly accomplishes horizontal path tracking tasks within a stipulated timeframe, effectively compensating for thruster failures, unknown ocean currents, and model parameter perturbations.