Comparison of online-parameter estimation methods applied to a linear belt drive system

verfasst von
Daniel Beckmann, Mauro Hernán Riva, Matthias Dagen, Tobias Ortmaier
Abstract

In this paper a comparison of three methods for online parameter estimation is presented. The analyzed algorithms are a well known recursive least squares method (RLS), an Extended Kalman Filter (EKF) in joint state form, and an adaptive Extended Kalman Filter (aEKF). The methods' performances regarding accuracy, respond time and computing time are compared using a commercial industrial testbed, consisting of a linear belt drive for positioning tasks, an industrial servo inverter and a programmable logic controller. In addition, the online state sensitivity w.r.t. the parameters provided by the aEKF is analyzed to check the parameter excitation. These signals can be used to stop the parameter estimation for insufficient excitation.

Organisationseinheit(en)
Institut für Mechatronische Systeme
Typ
Aufsatz in Konferenzband
Seiten
364-369
Anzahl der Seiten
6
Publikationsdatum
2016
Publikationsstatus
Veröffentlicht
Peer-reviewed
Ja
ASJC Scopus Sachgebiete
Steuerungs- und Systemtechnik, Steuerung und Optimierung
Elektronische Version(en)
https://doi.org/10.1109/ecc.2016.7810312 (Zugang: Geschlossen)