Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Identification method for time-varying ARX Quentin Rentmeesters, P.-A. Absil, and Paul Van Dooren
 

Summary: Identification method for time-varying ARX
models
Quentin Rentmeesters, P.-A. Absil, and Paul Van Dooren
D´epartement d'ing´enierie math´ematique
Universit´e catholique de Louvain
B-1348 Louvain-la-Neuve, Belgium
{Quentin.Rentmeesters,PA.Absil,Paul.Vandooren}@uclouvain.be
Summary. This paper presents a new approach to identify time-varying ARX mod-
els by imposing a penalty on the coefficient variation. Two different coefficient nor-
malizations are compared and a method to solve the two corresponding optimization
problems is proposed.
1 Introduction
Time-varying processes appear in many applications such as speech process-
ing, time-varying behavior detection (fault detection or wear detection) or
more generally when some parameters of a linear system vary over time. In
this paper, we are interested in time-varying systems identification using an
ARX model of order N - 1:
N-1
i=0
y(t - i)i(t) =

  

Source: Absil, Pierre-Antoine - Département d'ingénierie Mathématique, Université Catholique de Louvain

 

Collections: Mathematics