Moving-Bank Multiple Model Adaptive Estimation Applied to Flexible Spacestructure Control

摘要

This investigation focused on the use of moving-bank multiple model adaptive estimation and control (MMAE). Moving-bank MMAE reduces the computational burden of MMAE by implementing only a subset of the Kalman filters (9 filters versus 100 in this research) that are necessary to describe the system to be estimated/controlled. Important to the development of the moving-bank MMAE are the decision logics governing the selection of the subset of filters. The decision logics cover three situations: initial acquisition of unknown parameter values; tracking unknown parameter values; and reacquisition of the unknown parameters following a jump change in these parameter values. This thesis applied moving-bank MMAE to a rotating two bay truss model of a flexible spacestructure. The rotating two bay truss approximated a space structure that had a hub with appendages extending from the structure. The mass of the hub was large relative to the mass of the appendage. The mathematical model was developed using finite element analysis, transformed into modal formulation, and reduced using a method referred to as singular perturbations. Multiple models were developed by assuming that variation occurred in the mass and stiffness of the structure.