Quantized Control And Filtering Of Networked Control Systems

With the rapid development of internet, more and more practical systems are ex-panded to the networked application system (NAS), for example, power systems, wire-less communication systems, Lan-based and Wan-based communication systems, net-worked control systems (NCS) and so forth. As a type of NAS, NCS have receivedmuch attention by scholars for the significant value of application. They focus on thefactors such as time-delay, data packet dropouts, quantization, distortion, noisy whichhave impact on the network environment. This thesis concerns with the discrete-timesystem. The main contents of this thesis are shown below:Firstly, the regional stabilization of networked control systems with saturatedquantizer is investigated. Diferent from the traditional common quadratic Lyapunovmethod, here we utilize the maximal Lyapunov functional to find a non-ellipsoidal ini-tial set as large as possible and an domain of convergence as small as possible such thatall the trajectories of the system starting from the initial set will eventually converge tothe domain of convergence.Secondly, the quantized control for networked control systems subject to bothpack dropouts and time-varying delay is addressed. By converting the quantized feed-back design problem to well-known robust control problems with sector bound uncer-tainties, some sufcient conditions are derived to ensure the stochastic stability of theclosed-loop systems.Finally, the quantized H_∞filtering problem for Markov jump systems with partialinformation on the transition probabilities is considered. A full-order filter is designedsuch that the filtering error dynamics is stochastic stable and the prescribed noise atten-uation level in the H_∞sense is also achieved.