A bayesian approach to closed-loop system identification

Stéphane Thil (1) , Marion Gilson (1)

1 CRAN - Centre de Recherche en Automatique de Nancy

The challenging issue of identifying a closed-loop system from short and/or non-informative data records is addressed. A bayesian approach is developed within this framework. It is shown that accurate estimates and realistic confidence intervals can be obtained by taking into account prior knowledge on the system. The performances of the proposed method are illustrated with a simulation example.

Mots clés: Closed-loop identification; Bayesian technique; MCMC methods; non-informative data records

Domaines: Automatique / Robotique

Stéphane Thil, Marion Gilson. A bayesian approach to closed-loop system identification. 16th IFAC World Congress, Jul 2005, Prague, Czech Republic. pp.CD-ROM. ⟨hal-00152697⟩