Model Discrepancy in Robotic Calibration: Its Influence on the Experimental Parameter Identification of a Parallel Space Telescope

Thibault Gayral 1 David Daney 1 Marc Bernot 2
1 COPRIN - Constraints solving, optimization and robust interval analysis
CRISAM - Inria Sophia Antipolis - Méditerranée , ENPC - École des Ponts ParisTech
Abstract : The model of a robot may not be able to consider all the physical phenomena influencing the manipulator performances since they are too numerous and/or difficult to measure: this is model discrepancy. For a highly-accurate active space telescope, an important source of inaccuracy was measured using photogrammetry: the deformation of its mobile platform. This deformation cannot be directly measured in space and needs to be properly modeled in order to enable the telescope calibration with the available measurements. Two incremental models are proposed and the parameter observability is discussed. After experimental calibration, a micrometer accuracy can be reached. The influence of model discrepancy on the experimental parameter identification is finally discussed.
Type de document :
Communication dans un congrès
IROS - IEEE/RSJ Int. Conf. on Intelligent Robots and Systems - 2013, Nov 2013, Tokyo, Japan. 2013
Liste complète des métadonnées

https://hal.inria.fr/hal-00903848
Contributeur : Thibault Gayral <>
Soumis le : mercredi 13 novembre 2013 - 11:31:38
Dernière modification le : jeudi 11 janvier 2018 - 16:24:40

Identifiants

  • HAL Id : hal-00903848, version 1

Collections

Citation

Thibault Gayral, David Daney, Marc Bernot. Model Discrepancy in Robotic Calibration: Its Influence on the Experimental Parameter Identification of a Parallel Space Telescope. IROS - IEEE/RSJ Int. Conf. on Intelligent Robots and Systems - 2013, Nov 2013, Tokyo, Japan. 2013. 〈hal-00903848〉

Partager

Métriques

Consultations de la notice

187