This paper describes an unsupervised approach to retrieve the kinematic parameters of a wheeled mobile robot. The robot chooses which action to take in order to minimize the uncertainty in the parameter estimate and to fully explore the parameter space. Our method explores the effects of a set of elementary motion on the platform to dynamically select the best action and to stop the process when the estimate can be no further improved. We tested our approach both in simulation and with real robots. Our method is reported to obtain in shorter time parameter estimates that are statistically more accurate than the ones obtained by steering the robot on predefined patterns.
Unsupervised calibration of wheeled mobile platforms / Di Cicco, Maurilio; Della Corte, Bartolomeo; Grisetti, Giorgio. - ELETTRONICO. - (2016), pp. 4328-4334. (Intervento presentato al convegno 2016 IEEE International Conference on Robotics and Automation, ICRA 2016 tenutosi a Stockholm; Sweden) [10.1109/ICRA.2016.7487631].
Unsupervised calibration of wheeled mobile platforms
Di Cicco, Maurilio
;Della Corte, Bartolomeo
;Grisetti, Giorgio
2016
Abstract
This paper describes an unsupervised approach to retrieve the kinematic parameters of a wheeled mobile robot. The robot chooses which action to take in order to minimize the uncertainty in the parameter estimate and to fully explore the parameter space. Our method explores the effects of a set of elementary motion on the platform to dynamically select the best action and to stop the process when the estimate can be no further improved. We tested our approach both in simulation and with real robots. Our method is reported to obtain in shorter time parameter estimates that are statistically more accurate than the ones obtained by steering the robot on predefined patterns.File | Dimensione | Formato | |
---|---|---|---|
DiCiccio_Unsupervised_2016.pdf
solo gestori archivio
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
4.99 MB
Formato
Adobe PDF
|
4.99 MB | Adobe PDF | Contatta l'autore |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.