Optimal Decentralized Output-Feedback LQG Control With Random Communication Delay

摘要

This paper is concerned with the optimal decentralized output-feedback control of the large-scale systems. A random information pattern is considered, where the information is transmitted among the subsystems with random communication delays. For the random information pattern, the optimal LQC; problems for both global estimation case and local estimation case are studied. It is difficult to derive the optimal controller under random framework, because the gains of the controller must be designed to satisfy the random sparse structure constraints. In this paper, we design the optimal controller by Hadamard product method. For global estimation case, the gains of the controller are obtained by solving linear matrix equation. For local estimation case, an iterative algorithm is exploited to compute the gains. In addition, the value of the cost function achieved by the designed controller is found and shown to monotonically increase with the increase of the delay probability for both global and local estimation cases. Finally, the theoretical results are illustrated by two numerical examples.