OATAO - Open Archive Toulouse Archive Ouverte Open Access Week

An improved instrumental variable method for industrial robot model identification

Brunot, Mathieu and Janot, Alexandre and Young, Peter C. and Carrillo, Francisco Javier An improved instrumental variable method for industrial robot model identification. (2018) Control Engineering Practice, 74. 107-177. ISSN 0967-0661

[img]
Preview
(Document in English)

PDF (Author's version) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
1MB

Official URL: http://doi.org/10.1016/j.conengprac.2018.02.010

Abstract

Industrial robots are electro-mechanical systems with double integrator behaviour, necessitating operation and model identification under closed-loop control conditions. The Inverse Dynamic Identification Model (IDIM) is a mechanical model based on Newton's laws that has the advantage of being linear with respect to the parameters. Existing Instrumental Variable (IDIMIV) estimation provides a robust solution to this estimation problem and the paper introduces an improved IDIM-PIV method that takes account of the additive noise characteristics by adding prefilters that provide lower variance estimates of the IDIM parameters. Inspired by the prefiltering approach used in optimal Refined Instrumental Variable (RIV) estimation, the IDIM-PIV method identifies the nonlinear physical model of the robot, as well as the noise model resulting from the feedback control system. It also has the advantage of providing a systematic prefiltering process, in contrast to that required for the previous IDIM-IV method. The issue of an unknown controller is also considered and resolved using existing parametric identification. The evaluation of the new estimation algorithms on a six degrees-of-freedom rigid robot shows that they improve statistical eficiency, with the controller either known or identified as an intrinsic part of the IDIM-PIV algorithm.

Item Type:Article
Audience (journal):International peer-reviewed journal
Uncontrolled Keywords:
Institution:Other partners > Australian National University - ANU (AUSTRALIA)
Université de Toulouse > Institut National Polytechnique de Toulouse - Toulouse INP (FRANCE)
French research institutions > Office National d'Etudes et Recherches Aérospatiales - ONERA (FRANCE)
Laboratory name:
Statistics:download
Deposited On:11 Feb 2019 14:14

Repository Staff Only: item control page