H₂/H_∞Filtering For Networked Systems With Multi-channel Communication Constraints

There usually exists noise interference in networked control systems (NCSs),which is known to be the main cause for performance deterioration or eveninstability of NCSs. So it is necessary to use the filtering techniques to eliminate orreduce noise interference to improve the system performance. Meanwhile, somenetwork-related non-deterministic issues (e.g. network-induced delay, data packetloss and packet disorder) are inevitably introduced, which make NCSs morecomplex. Therefore, multi-channel network-related non-deterministic issues must betaken into account in analyzing and designing H2/H∞filtering with single network orhybrid network or based on the event-driven mechanism. The main work issummarized as follows:(1) H2/H∞filtering of the networked discrete-time systems with multi-channeldata packet losses is studied. Considering multi-channel data packet losses in thesensor-to-filter, the random data packet losses of each channel was firstly describedby a two-state Markov chain. Then the filtering error dynamic system was modeledas a Markov jump system with multiple modes. Based on the Lyapunov method andlinear matrix inequality (LMI) technique, a sufficient condition for the existence of afilter is presented, and the designed filter enables the error dynamic system to bestochastic stability and to guarantee a prescribed H2/H∞performance γ.(2) H2/H∞filtering for complex multi-channel networked systems with differentnetwork-related non-deterministic issues in each channel is studied. Using amaximum data packet out-of-order strategy, multi-channel communicationconstraints from the sensor to the filter are firstly solved within one framework. Ageneralized comprehensive filtering error dynamic system model which containsmulti-channel communication constraints is then established. Under the knownmaximum network-induced delays and the known number of data packet dropout,the designed filter enables multi-channel error dynamic system to be stochasticstability and to guarantee a prescribed H2/H∞performance. The results are furtherextended to multi-channel networked system with parameter uncertainty.(3) H2/H∞filtering for complex multi-channel hybrid networked systems withdifferent network delays and packet dropout in each channel is investigated. Thecharacters of hybrid network are firstly analyzed, which are then described byanalytical formulations. A comprehensive filtering error dynamic system modelwhich contains multi-channel hybrid network communication constraints is thenestablished. The designed filter enables multi-channel error dynamic system to be stochastic stability and to guarantee a prescribed H2/H∞performance, under theknown maximum network delays of wire and wireless network.(4) Due to the communication bandwidth limitation and multi-channelcommunication constraints, H2/H∞filtering based on the event-triggered mechanismis investigated. The event-triggered mechanism is analyzed firstly andnetwork-related non-deterministic parameters are described by analyticalformulations. A comprehensive dynamic filtering error system model which containsevent-triggered parameters and multi-channel communication parameters isestablished. The designed filter enables multi-channel error dynamic system to bestochastic stability and to guarantee a prescribed H2/H∞performance, under theknown maximum network-induced delay. The relationship among event-drivenparameters, control parameters and network performance parameters is alsoestablish.