Optimal Control of Decentralized Networked Control System with Input Delay and Packet Dropout

This paper considers an optimal decentralized control problem for networked control systems where the two controllers are subject to input delay, and one controller is also subject to packet dropout. A Bernoulli process is used to model the control packet loss. The two controllers have a one-side control sharing information structure, that is, one controller's information is available to the other controller, and not the verse. We provide a necessary condition in terms of the packet dropout probability and the system parameter under which the optimal decentralized controllers exist. It is shown that the two controllers are linear in the state estimates, and the controller with full information is also linear in the actual state of the system.