Model selection via worst-case criterion for nonlinear bounded-error estimation

Author

Brahim-Belhouari, Sofiane ; Kieffer, Michel ; Fleury, Gilles ; Jaulin, Luc ; Walter, Éric

Author_Institution

Ecole Superieure d´´Electr., Gif-sur-Yvette, France

Volume

49

Issue

3

fYear

2000

fDate

6/1/2000 12:00:00 AM

Firstpage

653

Lastpage

658

Abstract

In this paper, the problem of model selection for measurement purpose is studied. A new selection procedure in a deterministic framework is proposed. The problem of nonlinear bounded-error estimation is viewed as a set inversion procedure. As each candidate model structure leads to a specific set of admissible values of the measurement vector, the worst-case criterion is used to select the optimal model. The selection procedure is applied to a real measurement problem, grooves dimensioning using remote field eddy current inspection

Keywords

curve fitting; deterministic algorithms; eddy current testing; error statistics; inverse problems; measurement errors; measurement uncertainty; modelling; nonlinear estimation; size measurement; admissible values; deterministic framework; grooves dimensioning; interval analysis; inverse problems; mean square fitted curves; measurement vector; model selection; nonlinear bounded-error estimation; optimal model; remote field eddy current inspection; set inversion procedure; uncertainty; worst-case criterion; Acoustic noise; Current measurement; Eddy currents; Error correction; Inspection; Inverse problems; Parameter estimation; Probability density function; Stochastic resonance; Uncertainty;

fLanguage

English

Journal_Title

Instrumentation and Measurement, IEEE Transactions on

Publisher

ieee

ISSN

0018-9456

Type

jour

DOI

10.1109/19.850410

Filename

850410