Decentralized Robust Control of Infinite-dimensional Systems

Abstract

Decentralized robust control of infinite-dimensional systems is considered. A decentralized robust controller design approach for such systems is proposed. In this approach, overlapping decompositions is first used to obtain local models. An error function, which depends on frequency, is then derived. The proposed approach then uses the local models and the derived function to design decentralized stabilizing controllers. The advantage of the proposed approach is that, apart from the derivation of the error function, the design process is totally based on the local nominal models. Satisfying a simple local condition, which involves the aforementioned error function, guarantees the robust stability of the actual overall system.