| The attitude control and momentum management of spacecraft actuated by momentum exchange devices was investigated in this study. The spacecraft were assumed to be rigid and nonlinear, subject to uncertain spacecraft mass parameters and disturbances with their known bounds. The control objective was to achieve the attitude tracking maneuvers of the spacecraft to a reference trajectory while keeping the angular momentum of the momentum exchange devices bounded. In the design of feedback control system, we proposed a control law consisting of an artificial neural network and linear control logic. The Lyapunov stability theory was used to obtain closed-loop stability conditions. In the neural network, on-line network tunings were designed to minimize tracking errors while satisfying the stability condition on the neural network. To demonstrate the applicability of the neural-network-based control law, computer simulations were performed. In the simulations, the spacecraft tracking maneuvers were achieved in the vicinity of the reference trajectory while the device momentums were bounded during spacecraft inertia changes and under the disturbances. Therefore, the proposed neural-network-based attitude control and momentum manager is robust to mass parameter uncertainties and disturbances. Moreover, the learning capability of the neural network has been verified from the bounded network weights during the simulations. |
