Parameter sensitivity reduction in fixed-order dynamic compensation

Abstract

This paper presents a simple formulation for designing fixed-order dynamic compensators that are robust to uncertainty at the plant input and have reduced sensitivity to structured uncertainty in the plant dynamics. An approximate loop transfer recovery technique is extended to include a penalty on the trajectory sensitivity dynamics. This approach avoids introducing the sensitivity states, and thus the dimension of the optimization problem is not increased. The usefulness of the technique is illustrated by two compensator designs. The first design considers a case where the uncertain parameters occur in the output matrix. The second design is for a high bandwidth controller for a flexible spacecraft.