Optimal Decentralized Output-feedback Control with Communication Delay: A System Augmentation Approach

This paper studies the optimal output-feedback control problem for a large-scale system defined over a strongly connected directed graph. The system information travels across an edge in the graph with one time-step delay. To solve the optimal control problem with this delay model, an augmented system is introduced. Based on the augmented system, the optimal controller is designed via computing the estimated augmented state and constructing two gain matrices. The estimated augmented state is computed by a modified Kalman filter. Furthermore, explicit linear matrix equations are established, whose solutions are the optimal gain matrices. Finally, a numerical example is used to demonstrate the theory proposed in this paper.