Filtering of discrete-time Markov jump linear systems Markov jump linear systems

Author(s)

André Ricardo Fioravanti

• Document type: Master's Dissertation
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação
• Defense date: 2007-07-10

Examining board members

José Cláudio Geromel; Alexandre Trofino Neto; João Bosco Ribeiro do Val; Renato da Rocha Lopes

• Advisor: José Cláudio Geromel

Abstract

This thesis addresses the H2 and Hoo filtering design problem of discrete-time Markov jump linear systems. First, under the assumption that the Markov parameter is measurable, we provide the characterization of all filters such that the estimation errar remains bounded by a given narm leveI, yielding the complete solution of the mode-dependent filtering design problem. Based on this result, a robust filter design to deal with convex bounded parameter uncertainty is considered. In the sequeI, a design procedure for modeindependent filtering design is proposed. All filters are designed by solving linear matrix inequalities. The theory is illustrated by means of a practical example, consisting the data communication through a markovian channel (AU)