Design Of Non-fragile Tracking Controller For T-S Fuzzy System

The missions of stabilization and tracking are two typical control theroy problems.As one of the areas of active research, tracking control has very important value inengineering applications. The purpose of tracking control is to make the control systemoutput variables or state variables tracking a given reference signal model and meet theH tracking control performance index. For the non-linear system, the trackingproblem is more complicated than the stabilization problem.The design of tracking controller base on T-S fuzzy system will have a widersignificance in practical projects and theoretical researches because it have powerfuladvantage in nonlinear system analysis and design. According to continuous T-S fuzzysystem, continuous T-S fuzzy system with uncertainties and discrete T-S fuzzy systemwith delay, the design and research of non-fragile tracking controller is proposed in thisthesis. At the same time, system state variables are estimated via the fuzzy observer.The designed controller with some controller additive gain variations can still make theclosed-loop system meet the tracking control performance index.The sufficient conditions of non-fragile controller and observer are obtained vialinear matrix inequality (LMI) method. In order to avoid the traditional process of usinga two-step iterative algorithm, the observer-based controller design conditions aredecomposed into independent existence conditions in the form of linear matrixinequalities for continuous T-S fuzzy system. The sufficient conditions of controller canensure the tracking performance with less conservative. In order to eliminate thecoupling between the system uncertainties and the controller uncertainties, a descriptorrepresentation approach is used for T-S fuzzy system with uncertainties. The fuzzyLyapunov function approach and effective matrix inequalities transformation techniquesare employed for discrete T-S fuzzy system with delay. Additionally, the effectivenessand superiority of the proposed design methods are further proved via numerical oractual simulation examples respectively.Finally, the results of the thesis are summarized and the topics of the furtherresearch are pointed out.