Decentralized Robust Control of Large-scale Time-delay Systems

Decentralized robust controller design problem for large-scale interconnected systems which involve uncertainties in the system matrices and uncertain time-delays is considered. An error bound, which accounts for the neglected interactions between the subsystem models and uncertainties in the interactions, in the subsystem models, and in the time delays is first derived using overlapping decompositions and expansions. A decentralized controller design approach, which uses this bound, is then proposed. The advantage of the proposed approach is that, all uncertainties and neglected dynamics are summarized in one frequency-dependent scalar function and satisfying a simple condition guarantees that the overall closed-loop system under decentralized controllers, which are designed considering local models, is robustly stable. The application of the proposed approach is demonstrated on a flow control problem in a datacommunication network.