Inverse Neural Networks Control for Nonlinear Flexible Structure

摘要

For the attitude control of nonlinear flexible structure, a special control method is proposed that combines inverse neural networks feed-forward control with output feedback control. The conventional inverse system method requests all statesfeedback. Being aimed at immeasurable characteristics of flexible structure modes, the recurrent neural network is introduced to approach the inverse system of controlled plant The inverse neural network forms the feed-forward controller. The outputfeedback constitutes the feedback loop in order to guarantee the stability of system. The simulation results are presented based on a flexible satellite model.