Velocity-Free Adaptive Neural-Fuzzy Predefined-Time Attitude Control for Spacecraft

The high-performance control of the spacecraft attitude is significant for successfully executing diverse tasks. To realize this goal, a velocity-free adaptive neural-fuzzy predefined-time attitude controller is presented for the spacecraft with uncertain inertia, exogenous disturbances, and input saturation. Firstly, an improved predefined-time stable system is established, featuring an adjustable convergence time to enhance the flexibility of the controller design. Utilizing the robust approximation ability of the neural-fuzzy network, a state observer and a nonsingular sliding mode controller are developed to achieve accurate state measurements, improve strong robustness, and eliminate singularity issues. Subsequently, a modified anti-saturation method is designed via the Gaussian function and auxiliary compensation system to resolve the input saturation problem. Based on the Lyapunov theorem, the predefined-time stability of the whole system is confirmed. Finally, through comparative simulations and numerical analysis, it can be concluded that: 1) the system state converges within a predefined time related to only a single parameter, and the actual convergence time is adjustable; and 2) compared to existing control schemes, the proposed control scheme demonstrates superior anti-disturbance ability, avoids potential singularities, achieves faster convergence, and eliminates input saturation.