Computational Techniques for Robust and Fixed-Structure Controller Design

摘要

In practice discrepancies between real physical systems and their mathematical models are unavoidable. These uncertainties in the mathematical model often result in severe degradation in control system performance and sometimes even instability. One of the main objectives of feedback control theory is to design controllers that are stable and guarantee certain performance objectives, in the face of these uncertainties. While robust control theory has reached a certain maturity in recent years, much remains to be done as far as numerical algorithms for practical robust controller synthesis is concerned. In practice, because of throughput limitations on control processors, the order (and sometimes additionally the structure) of the controller may have to be constrained a priori. Traditional controller reduction schemes do not guarantee robustness or optimality of the resulting controllers. The goal of this research is to develop algorithms for the analysis of closed-loop robust stability and the synthesis of fixed-structure and robust controllers.