| On line adaptive control is essential for Large Space Structure (LSS) where the model parameters are poorly known, due to modeling error and changing or variable configuration. Yet most of the adaptive control methods need some precise a priori structural information of the plant (at least the order). This is unsatisfactory for LSS adaptive control. In this paper, the Model Reference Adaptive Control (MRAC) for LSS using direct and indirect schemes base on a Reduced-Order Model (ROM) are obtained. In the direct MRAC scheme, the hyperstability theory is used for stability analysis, this approach provides greater flexibility in choosing adaptive laws and avoids the shortcomings of Lyapunov's method. For improving the sensors and actuators noncollocated problem, the almost strict positive realness condition is established. The indirect MRAC scheme, are based on an Auto Regressive Moving Average (ARMA) model, the spillover from the unmodel dynamics is compensated by Residual Mode Filter (RMF). The results of both schemes have been used to design an ROM-Based controller and tested successfully via computer simulation; the direct adaptive controller has been tested successfully via a flexible robot manipulator experiment. |
