Adaptive neural network-based satellite attitude control in the presence of CMG uncertainty

Abstract

An attitude tracking controller is developed for control moment gyroscope (CMG)-actuated satellites, which is shown to achieve accurate attitude tracking in the presence of unmodeled external disturbance torques, parametric uncertainty, and nonlinear CMG disturbances. Since the disturbances/uncertainties do not all satisfy the typical linear-in-the-parameters (LP) assumption, a neural network (NN) is included in the control development. The innovation of the result is the development of a Lyapunov-based design/analysis that indicates exponential convergence to an arbitrarily small domain. The result is obtained despite the characteristics of the uncertainty; the nonvanishing disturbance terms; and the fact that the control input is premultiplied by a non-square, time-varying, nonlinear, uncertain matrix. In addition to the Lyapunov-based analysis, experimental results demonstrate the performance of the developed controller.