Robust eigenvalue assignment in descriptor systems via output feedback

Abstract

Based on a recently proposed parametric approach for eigenstructure assignment in descriptor linear systems via output feedback, the robust eigenvalue assignment problem in descriptor linear systems via output feedback is solved. The problem aims to assign a set of finite closed-loop eigenvalues which have minimum sensitivities with respect to perturbations in the closed-loop coefficient matrices, while at the same time, guarantee the closed-loop regularity. The approach optimizes the design parameters existing in the closed-loop eigenvectors to achieve the minimum eigenvalue sensitivities, and use the extra degree of freedom existing in the solution of the gain matrix to further minimize the magnitude of the gain matrix and enhance the robustness of the closed-loop regularity. The approach allows the finite closed-loop eigenvalues to be optimized within desired regions, and is demonstrated to be simple and effective.