Error Bounds for FIR Models in Conditional Set-Membership Identification

Authors:	Casini Marco, University of Siena, Italy Garulli Andrea, University of Siena, Italy Vicino Antonio, University of Siena, Italy
Topic:	1.1 Modelling, Identification & Signal Processing
Session:	Set-Membership Techniques for Identification
Keywords:	Set-membership identification, bounded error, FIR models

Abstract

The present paper deals with the problem of reduced complexity model estimation in theframework of conditional set-membership identification. The measurement noise is assumed to beunknown but bounded, while the estimated model quality is evaluated according to a worst-casecriterion. Since optimal conditional estimators are generally hard to compute, projectionestimators are often used in view of their better tractability from a complexity viewpoint.Tight bounds on the suboptimality level of central projection estimators as compared to optimalones are derived for the case when FIR models are employed for approximation. These boundsimprove over known bounds holding for the general class of linearly parameterized models.