H1 and H2 control design for polytopic continuous-time Markov jump linear systems with uncertain transition rates.

This paper investigates the problems of H1 and H2 state feedback control design for continuous-time Markov jump linear systems. The matrices of each operation mode are supposed to be uncertain, belonging to a polytope, and the transition rate matrix is considered partly known. By appropriately model...

Full description

Bibliographic Details
Authors: Morais, Cecilia de Freitas, Braga, Marcio Feliciano, Oliveira, Ricardo Coração de Leão Fontoura de, Peres, Pedro Luis Dias
Format: article
Status:Published version
Publication Date:2015
Country:Brasil
Institution:Universidade Federal de Ouro Preto (UFOP)
Repository:Repositório Institucional da UFOP
Language:English
OAI Identifier:oai:repositorio.ufop.br:123456789/9277
Online Access:http://www.repositorio.ufop.br/handle/123456789/9277
https://doi.org/10.1002/rnc.3329
Access Level:Open access
Keyword:Markov jump linear systems
State feedback control
Continuous-time systems
Description
Summary:This paper investigates the problems of H1 and H2 state feedback control design for continuous-time Markov jump linear systems. The matrices of each operation mode are supposed to be uncertain, belonging to a polytope, and the transition rate matrix is considered partly known. By appropriately modeling all the uncertain parameters in terms of a multi-simplex domain, new design conditions are proposed, whose main advantage with respect to the existing ones is to allow the use of polynomially parameter-dependent Lyapunov matrices to certify the mean square closed-loop stability. Synthesis conditions are derived in terms of matrix inequalities with a scalar parameter. The conditions, which become LMIs for fixed values of the scalar, can cope with H1 and H2 state feedback control in both mode-independent and modedependent cases. Using polynomial Lyapunov matrices of larger degrees and performing a search for the scalar parameter, less conservative results in terms of guaranteed costs can be obtained through LMI relaxations. Numerical examples illustrate the advantages of the proposed conditions when compared with other techniques from the literature.