| ABSTRACT Due to possessing both non-linearity and impulsive property, it is achallenge to investigate descriptor nonlinear systems. Fuzzy descriptor systemcan be used to approximate a nonlinear descriptor system very well. Thus,fuzzy descriptor control systems have attracted attention in control community.So far, the research results on this topic are few. In this thesis, some new resultresults are developed as follows: 1) A new modeling method of the fuzzy descriptor system is presented to solve the control problem for systems with a time-varying matrix E , which is more general than the case when E is constant. The stability of the fuzzy descriptor system is discussed and a sufficient condition of the system stabilization is proposed. Furthermore, the existence of the common matrix X is discussed and a design method of matrix X is obtained, which can reduce the computation burden. 2) The design problems of fuzzy controllers and state observers are discussed in detail. The design approaches of state feedback controllers, output feedback controllers, full-order and reduced-order observers are developed. A sufficient condition is given to ensure the closed-loop system to be stable by using state observer-fuzzy controllers. Furthermore, the property of the separation for designing state-feedback gain and observer gain is obtained. 3) H∞ technique is used to investigate fuzzy descriptor control systems. Two approached are presented to design H∞ controllers. The definition of the tracking control for the fuzzy descriptor system is proposed, which are new and can avoid some difficulty due to the dependence of state vectors. The tracking controller design law is obtained by using the H∞ technique and the robustness of tracking control is also discussed. 4) Finally, some examples are given to illustrate the design procedures and IIABSTRACT the simulation results have shown the nice control or observation performance.
|
PDF Full Text Request