The Effect of Prior Knowledge on the Least Costly Identification Experiment - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

The Effect of Prior Knowledge on the Least Costly Identification Experiment

Résumé

Design of optimal input excitations is one of the most challenging problems in the field of system identification. The main difficulty lies in the fact that the optimization problem cannot always be formulated to be convex, therefore a globally optimal excitation for the dynamic system of interest cannot be guaranteed. In this paper, optimal input design (OID) for linear systems in the presence of prior knowledge is studied. Information related to exponential decay and smoothness is incorporated in the optimal input design problem by making use of the Bayes rule of information. Three different cases of modeling the linear dynamics are considered, namely Finite Impulse Response (FIR) model with and without prior knowledge, as well as the rational transfer function case. It is shown that the prior information affects the spectrum of the minimum power optimal input. The input with the least power is always obtained for the transfer function model case.
Fichier principal
Vignette du fichier
BIRPOUTSOUKIS_BOMBOIS_2018.pdf (364.36 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01997195 , version 1 (25-04-2019)

Identifiants

Citer

Georgios Birpoutsoukis, Xavier Bombois. The Effect of Prior Knowledge on the Least Costly Identification Experiment. 18th IFAC Symposium on System Identification, Jul 2018, Stockholm, Sweden. pp.209-214, ⟨10.1016/j.ifacol.2018.09.131⟩. ⟨hal-01997195⟩
46 Consultations
48 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More