Adaptive CSI Compensation for Reduced-Order-Model-Based Control of a Flexible Robot Manipulator

Abstract

It is not feasible to construct controllers for large space structures or large scale systems (LSS's) which are of the same order as the structures. Because of this, the controller for system performance of such a system is based on a much smaller reduced-order model (ROM). Unfortunately, the interaction between the LSS and the ROM-based controller can produce instabilities in the closed-loop system due to the unmodeled dynamics of the LSS. Residual mode filters (RMF's) allow the systematic removal of these instabilities in a matter which does not require a redesign of the controller. An adaptive, self-tung RMF design, which compensates for uncertainty in the frequency of the residual mode, has been simulated using the continuous-time and discrete-time models of a flexible robot manipulator. This paper will present the results of that work on adaptive, self-tuning RMF's, and will clearly show the advantages of this adaptive compensation technique for controller-structure interaction (CSI) instabilities in actively-controlled LSS's.