Robust finite range H-infinity filtering using LMIs with a scalar

Abstract

The aim of this project is to study the problem of full order robust filter design for linear systems with time-invariant uncertain parameters belonging to a polytope, taking into account performance criteria such as the minimization of bounds to the H-infinity and H2 norms of the transfer function from the exogenous noise input to the filter output. Both continuous and discrete-time systems will be considered. Strategies based on parameter-dependent Lyapunov functions will be adopted for the robust filter design, providing existence conditions in terms of matrix inequalities that include scalar parameters. For fixed values of the scalars, the conditions become linear matrix inequalities (LMIs). Computational tools available in Matlab will be used to program the LMIs, together with parsers and solvers of public domain. The aim of this Research Internship Abroad (BEPE) is to extend the original project of robust filtering to cope with H-infinity performance specifications in a finite frequency interval. In this sense, extensions of the so-called Kalman-Yakubovich-Popov (KYP) lemma in terms of LMIs with scalar variables will be investigated, for both continuous and discrete-time uncertain systems. (AU)