Decentralized Observer-Based Robust Model Predictive Control for a Class of Distributed Networked Systems

This paper presents a decentralized observer-based robust model predictive control strategy for a class of distributed networked systems. The overall system is composed of a number of interconnected nonlinear subsystems with time-varying transmission delays. Time delays are appeared in distributed subsystems due to the information transmission in the communication networks. A decentralized control strategy is implemented for each subsystem based on model predictive controllers. Each MPC problem is formulated with a memory-less observer-based feedback control law to minimize the upper bound of the infinite horizon cost that satisfies the sufficient stability conditions. Novel delay dependent sufficient conditions are proposed for the existence of such controllers in the form of a linear matrix inequality optimization problem. The simulation results for a chemical reactor plant are exploited to illustrate the applicability of the proposed method. Key-Words: Model Predictive ControlDecentralizedtransmission delaysobserverlinear matrix inequality