Optimal Control And Scheduling Of Networked System With Limited Channels

Networked control systems (NCSs) are closed-loop feedback control systems whose components such as sensors, actuators, and controllers are connected through a real-time communication network. In the context of NCSs, since communication capacity and bandwidth of the network are limited, only parts of sensors and actuators are allowed to exchange information with a remote controller, which is named medium access constraint or channels limitation of NCSs. Unfortunately, medium access constraint usually cause that measured information and control input signals can not be updated timely, which has potential detrimental effects on closed loop performance or even causes instability. Therefore, it is of importance and significance to further investigate the co-design of control and scheduling for NCSs with medium access constraints.First of all, the research background and significance of NCSs are stated, and a survey on the progress of NCSs issues in the existing literature is given. Then, in order to address the issues on limited channels and other communication constraints of NCSs, some integrated design methods for optimal control and communication scheduling strategies are proposed, which are summarized as follows:Based on the analysis techniques including optimal control theory and switching system method, an approach to co-design static periodic communication sequences and optimal controller is prsesented for NCSs with medium access constraints and network-induced delays.A switched system approach is proposed to solve the modeling, analysis and synthesis problem of NCSs with limited channels. According to the access characteristics of communication medium, the networked system is equivalently transformed into a switched system. Then, under a dynamic communication scheduling, optimal controllers are designed with the zero input control and control maintained policy.A novel method for the joint design of optimal H-infinity controller and dynamic scheduling policy is presented for constrained NCSs with parameter uncertainties. Sequently, the co-design of quantized feedback controller and dynamic communication sequences for NCSs with access constraints and quantization errors is derived. The issues on the optimal controller design of NCSs with access constraints are discussed by taking into account the effect of data packet dropouts. In the light of the medium access mechanism of actuators, the closed-loop networked system is converted as a markovian jump system. By employing markovian jump system theory and dynamic programming technique, the state feedback controller is obtained to guarantee the stability of the closed-loop system.Finally, the results of the dissertation are summarized, and further discussion on future work is also provided.