Event-Triggered Based H∞Filtering Of Networked Systems

At present, filtering problems of the networked system have got more and more attention, and much research results have been obtained. Because limited network communications bandwidth resources and some unique properties the network itself existing significantly deteriorate the system performance and even lead to system instability, the existing work also considered some adverse characteristic of the network. But from the perspective of resource allocation, the method that information datas are transmitted according to a fixed period exists insufficiency. Once the measured output or filter signal must be transmitted by the sharing network, it will cause unnecessary utilization of network bandwidth. So this issue needs to be studied further. On the other hand, with the if-then rules fuzzy system has already become one of the effective modeling methods of nonlinear system in the actual industrial process. However, when considering the output feedback control problem and filtering problem, it was assumed in many literatures that the premise variable was measurable. Otherwise, the output feedback controller and filter design based on PDC are infeasible. This article will use a kind of aperiodic transmission method to save limited transmission resource and reduce the processor computational cost, namely, whether the output transfer to the filter through the network depends on whether the measured output of the plant meeting the given event-triggered conditions. And, under the network environment, event-triggered based filtering problem of the fuzzy system with immeasurable premise variable and nonlinear systems are studied when considering the network delay and packet loss.The main research works in the thesis are as follows:In chapter1, the theory significance and the current research situation at home and abroad of researches of this paper are introduced. In chapter2, considering the limitations of network communication bandwidth, a new event-triggered communication scheme is proposed to avoid the transmission of redundant information. Based on the event trigger scheme, H∞filtering problem of the T-S fuzzy system with the network transmission time-delay is studied in the network environment, in which premise variables are immeasurable system states. Then, using the fuzzy Lyapunov functional method, sufficient conditions for asymptotically stable of the filtering error system are given in terms of linear matrix inequalities (LMIs). At the same time co-design method of filter and event trigger scheme also are given. Finally, through the Matlab/Simulink simulation the effectiveness of the proposed method is illustrated.In chapter3, based on the event trigger mechanism, H∞filtering problem of the nonlinear discrete systems with network channel data packet loss and network transmission time-delay is studied, in which the packet loss process is described by the Bernoulli distribution white noise sequence. Firstly, sufficient conditions are established by Lyapunov function approach for existence of the desired H∞filter T-S fuzzy system. Secondly, a co-design method of the corresponding filter and event trigger scheme also are given. Finally, a numerical example is given to illustrate the effectiveness of the proposed design method.In chapter4, the main research works of the thesis are concluded, and some research issues in future are prospected.