Projeto de controle robusto utilizando desigualdades matriciais lineares.

Abstract

The representation of systems by mathematical models is not perfect because uncertainties during the modeling or identification process, which can be caused by simplifications or dynamics not taken into account. A controller is said to be robust when it is capable of ensuring closed-loop stability and meeting system performance specifications even under the influence of external uncertainties and disturbances. The main technique used in the design of robust controllers is to minimize the ℋ1 norm, which can be formulated as a convex optimization problem using linear matrix inequalities. In this work, advanced techniques for the design of optimal and robust controllers are studied, using the linear matrix inequality approach. The procedures are used in the robust control design of a multivariable temperature system.