Eigenstructure Assignment and Robustness Improvement Using a Gradient-Based Method

Abstract

This paper presents a gradient-based method for increasing the robustness of multivariable systems, measured by the stability margins of the broken SISO loops. The method uses iterative perturbations of the closed loop eigenvalues to minimize a certain cost function containing weighted gain and phase margins and the variation of the eigenvalues. The initial and perturbed closed loop dynamics are specified using eigenstructure assignment. The algorithm presented is used in the gain design process of the lateral part of the flight control system for a generic transport aircraft. Results obtained with this novel approach are analyzed regarding the SISO and MIMO stability margins.