Robustness of a Moving-Bank Multiple Model Adaptive Controller for a Large Space Structure.

摘要

The robustness of moving-bank multiple model adaptive estimation (MMAE) and control (MMAC) algorithms is analyzed in this thesis. The mismatch of a 24-state truth model and a 6-state filter model are evaluated on the basis of MMAE/MMAC performance. A model developed using finite element analysis is used to approximate a large space structure which has a large central hub with appendages radiating out from it. The mass of the hub is considered to be much larger than the mass of the flexible structure. The model is developed in physical coordinates and then transformed into modal coordinates. To obtain a reduced order filter model, the method of singular perturbations is used. The actual positions and velocities of various physical points on the structure are used in the evaluation of the MMAE/MMAC algorithm performance. The results of this study of model mismatching indicates that the MMAE provides accurate position and velocity estimates even in the face of a 6-state to 24-state model mismatch. When a non-adaptive filter is used with a mismatched parameter location, the performance suffers slightly. The use of an adaptive estimator does provide improved performance in the face of uncertain parameter location. Stable control was obtained with the use of MMAC. For the case of nonadaptive filter and mismatched parameter location, the control algorithm behaved in a possible destructive manner. By allowing the filter to adapt to the initial parameter location, the MMAC algorithm provided stable control of the structure, even in the face of large disturbances. (AW)