Design of output feedback controllers for Takagi-Sugeno fuzzy systems (Q5947598): Difference between revisions

The study is concerned with the design of output feedback controllers for Takagi-Sugeno models governed by the following ``\(R\)'' rules of the form \[ \begin{aligned} & \text{-if }z_1 (t)\text{ is }M_{i1}\text{ and }z_2 (t) \text{ is }M_{i2} \text{ and\dots and }z_p(t) \text{ is }M_{ip}\\ & \text{then }dx/dt= A_ix(t)+ B_iu(t);\;y(t)= C_ix(t), \end{aligned} \] \(i=1,2, \dots,R\), with \(x(t)\) and \(u(t)\) being ``\(n\)'' and ``\(m\)'' dimensional vectors of state and control and \(y(t)\) denoting a \(q\)-dimensional output. The matrices \(A_i\), \(B_i\), and \(C_i\) form a description of a linear system (governed by the \(i\)-th rule). Concerning the above model, two general classes of problems are studied where (a) the premise parts of the rules include output variables of the system, and (b) the premise parts comprise state variables. In the first case, the authors derive sufficient conditions for the asymptotic convergence of observers, and they design an output feedback controller that stabilizes the system. In the second class, the paper includes a design method for observers and output feedback controllers for an approximation of the original fuzzy system. Both continuous and discrete-time cases are considered. A number of illustrative examples are included.