A recurrent neural network (RNN)-based attitude control method for a VSCMG-actuated satellite

A recurrent neural network (RNN)-based adaptive attitude controller is developed, which achieves attitude tracking in the presence of uncertain actuator inertia in addition to time-varying, uncertain satellite inertia. The satellite control torques are produced by means of a cluster of variable speed control moment gyroscopes (VSCMGs). The adaptive attitude controller results from a RNN structure while simultaneously acting as a composite VSCMG steering law. A null motion strategy is exploited to simultaneously perform the gimbal reconfiguration for singularity avoidance and wheel speed reg-ularization for internal momentum management. In designing the adaptive attitude controller, one of the challenges confronted is that the control input is premultiplied by a non-square, time-varying, nonlinear, uncertain matrix.