Decentralized robust control via quadratically invariant model projection

Jong-Han Kim, Sanjay Lall, Walter Merrill, Alireza Behbahani

Abstract: We propose a computational approach to systematically find decentralized H ∞ suboptimal controllers for general unstructured models. Exploiting the quadratically in variant model projection and the classical robust control synthesis techniques, we show that the original nonconvex problem can be conservatively solved by a series of convex optimization problems.