A DIRECT MODEL REFERENCE ADAPTIVE APPROACH TO THE CONTROL OF SPACE STATIONS

Of all large space structural systems, space stations present a unique challenge and requirement to advanced control technology. Their operations require control system stability over an extremely broad range of parameter changes and high level of disturbances. During shuttle docking the system mass may suddenly increase by more than 100% and during station assembly the mass may vary even more drastically. These coupled with the inherent dynamic model uncertainties associated with large space structural systems require highly sophisticated control systems that can grow as the stations evolve and cope with the uncertainties and time-varying elements to maintain the stability and pointing of the space stations.;This dissertation first deals with the aspects of space station operational properties including configurations, dynamic models, shuttle docking contact dynamics, solar panel interaction and load reduction to yield a set of system models and conditions. A model reference adaptive control algorithm along with the inner-loop plant augmentation design for controlling the space stations under severe operational conditions of shuttle docking, excessive model parameter errors, and model truncation are then investigated. The instability problem caused by the zero frequency rigid body modes and a proposed solution using plant augmentation are addressed. Two sets of sufficient conditions which guarantee the globally asymptotic stability for the space station systems are obtained.;The performance of this adaptive control system on space stations is analyzed through extensive simulations. Asymptotic stability, high rate of convergence and robustness of the system are observed even under the above-mentioned severe conditions and constraints induced by control hardware saturation. It is also found that further actuation level reductions can be achieved by using model switching and disturbance modeling techniques.