Small satellite attitude control based on variable structure control and RBF network

During the operation of small satellites, the non-linear disturbances caused by external factors are the important reasons affecting the accuracy of small satellite attitude control. It is difficult to meet the control request with the traditional PID control. Many control theories has been researched, but a satisfactory result has not got. Variable structure control is an effective way to deal with nonlinear perturbation. Due to the existence of the unknown disturbance, a sliding mode variable structure control model based on RBF neural networks is established, which combined the self-learning ability of RBF neural networks and the robustness of variable structure control technology. The RBF networks were used to estimate the nonlinear disturbances on line, which overcame the adverse influence of the unknown disturbance. The effectiveness of variable structure control and RBF neural network variable structure control were simulated and compared. The results show that variable structure control based on RBF network can restrain the influence of the unknown disturbances and inhibit the " buffeting " phenomenon, the accuracy is better than ordinary variable structure control, moreover the effectiveness and robustness of the control law was also proved.