A self-adjusting sliding-mode control based on RBF neural network for flexible spacecraft attitude

Aiming at meeting the needs of high precision and high stability in large angle attitude maneuver control for flexible spacecraft, a self-adjusting sliding-mode control law based on RBF neural network was proposed. A delay factor in exponential form was introduced into the reaching-law of sliding-mode control to improve the stability of maneuver process. An online self-adjusting factor was designed to adjust the amplitude of symbolic functions in order to reduce chattering. Furthermore, a real-time RBF neural network was used to estimate and compensate the influence caused by coupling, unknown boundary disturbances and uncertainties. The stability of the system was proved with Lyapunov technique. Finally, the simulation results demonstrate the effectiveness of the proposed control law. The pointing precision and stability, as well as the robustness against system parameter uncertainties and unknown boundary disturbances are improved.