Robust feedback design for stabilization of nonlinear systems with sampled-data and quantized output: atractive ellipsoid method

For a wide class of nonlinear systems, containing internal uncertainties as well as external bounded perturbations, the technique for robust feedback designing is suggested. The class of stabilizing dynamic feedbacks with a given linear structure is characterized by the corresponding BMI's, which is shown to be reduced to the system of LMI's. The optimal parameters of the feedback controllers realize the matrix solution to the conditional optimization problems under a set of specific linear matrix constraints. This optimization problem is solved by the standard MATLAB packages. Numerical examples illustrate the workability of the suggested approach.