An improved finite frequency approach to robust H∞ filter design for LTI systems with polytopic uncertainties

In this paper, the finite frequency robust H filtering problem (H-FFRFP), design of a robust filter minimizing the H norm from the disturbance input to the estimation error evaluated over a prescribed finite frequency domain, is considered for continuous-time and discrete-time linear time-invariant (LTI) systems with polytopic parameter uncertainties. By means of the generalized Kalman-Yakubovich-Popov (KYP) lemma in combination with a result known as Finsler's lemma, the H-FFRFP is cast as a linear matrix inequality (LMI) optimization. Examples are given to demonstrate that the proposed condition can achieve improvement over the previous ones in the literature. Copyright (c) 2012 John Wiley & Sons, Ltd.