Robust Control And Filtering Of T-S Nonlinear Fuzzy Networked Control Systems

With the increasing requirement of technical performance such as communication,control and computer,more and more control systems can implement closed-loop control through the network to carry out remote control and resource sharing among different regional components.We call it a Networked Control Systems.NCSs have been widely used in the field of industrial control because of their many advantages.However,due to the limited network bandwidth and the irregular changes of data traffic and other factors,there will be the phenomenon of signal transmission delay,data packet dropout and quantization error and so on,at the same time,the requirement of system security and reliability is increased.Therefore,it is very important to study the fault-tolerant control problem of NCSs.In the actual engineering site,the controlled object often has a certain degree of nonlinearity,and the T-S fuzzy identification method can be widely applied to the modeling of nonlinear system,because it has the ability of approximating the nonlinear system with arbitrary precision.In addition,considering the wide existence of persistent and amplitude-bounded external interference signals in the actual systems,this paper studies the nonlinear NCSs from the traditional robust H_? and L₂-L_? performance constraints to make a challenge to robust L₁ constraints with more practical applications,and the main research results are as follows:First of all,for the non-ideal network environment with quantization error,the robust H_?and L₂-L_? fault-tolerant control problems are studied when the outputs of uncertain nonlinear NCSs are energy bounded and persistent and amplitude-bounded respectively.The quantization error is expressed in the form of parameter uncertainty by sector bounded method,and by using Parallel Distributed Compensation algorithm and establishing the fault model matrix,the global fuzzy model of closed-loop systems are obtained based on T-S modeling.On this basis,the existence conditions of robust H_? and L₂-L_? fault-tolerant controllers that guarantee the systems to be asymptotically stable are given according to LMI technique.Secondly,taking into account the influence of factors such as network communication limitation and non-linearity,and considering that the external interference of the actual system is not always square integrable,therefore,for the nonlinear NCSs approximated by T-S fuzzy model,the constraint conditions that guarantee the the closed-loop systems to have the H_? and L₁ noise attenuation level with respect to both the input and output signals are energy bounded and all persistent and amplitude-bounded disturbances are given,respectively.And then the algorithm of the corresponding filter matrix is also given.The numerical simulation results show that the proposed L₁ filtering method has a significant advantage over the H_? filtering method to reduce the peak value of the interference signal,and the L₁ filtering method is applied to the tunnel diode model,which shows the theoretical significance and practical application value of the proposed method.Finally,considering the dynamic output feedback control strategy can more effectively meet the performance index of the closed-loop system,and it has the advantages of improving the reliability of system and reducing the implementation of control cost,accordingly,for the T-S fuzzy NCSs with transmission delay and data packet dropout,a delay-dependent and fuzzy basis-dependent Lyapunov-Krasovskii functional is adopted,which can make the system asymptotically stable and have the worst-case performance criterion.Because of the existence of coupling between the Lyapunov-Krasovskii functional matrix and the system parameter matrix,we need to introduce the additional matrix and apply the projection lemma to obtain the L₁ performance criterion which is easy to process decoupling,furthermore,a method for solving the parameters of dynamic output feedback controller with peak to peak performance gain is obtained.