Using Linear Matrix Inequalities to Optimize Anti-Windup Gain for Convex Systems

Abstract

This research sought out to determine a method for finding an optimal anti-windup gain for a convex system. The combination of two past works led to this method. The first publication, Static Anti-Windup Design, used a linear matrix inequality, or LMI, to determine an anti-windup gain for a nonlinear system. This LMI guaranteed regional stability paired with performance. The second paper, Saddle Point Convergence of Constrained Primal-Dual Dynamics, showed how a convex primal dual system can be implemented with control. This research bridges the gap between these works.