An online learning procedure for feedback linearization control without torque measurements

M. Capotondi; G. Turrisi; C. Gaz; V. Modugno; G. Oriolo; A. De Luca

An online learning procedure for feedback linearization control without torque measurements

M. Capotondi, G. Turrisi, C. Gaz, V. Modugno, G. Oriolo, A. De Luca

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

By exploiting an a-priori estimate of the dynamic model of a manipulator, it is possible to command joint torques which ideally realize a Feedback Linearization (FL) controller. The exact cancellation may nevertheless not be achieved due to model uncertainties and possible errors in the estimation of the dynamic coefficients. In this work, an online learning scheme for control based on FL is presented. By reading joint positions and joint velocities information only (without the use of any torque measurement), we are able to learn those model uncertainties and thus achieve perfect FL control. Simulations results on the popular KUKA LWR iiwa robot are reported to show the quality of the proposed approach.

Original language	English
Title of host publication	Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),
Publication status	Published - 30 Oct 2019

Bibliographical note

Note: Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498

Organising Body: International Foundation of Robotics Research (IFRR),

Organising Body: International Foundation of Robotics Research (IFRR),

Keywords

Computer science and informatics

Access to Document

https://proceedings.mlr.press/v100/capotondi20a.html

1 Paper

An online learning procedure for feedback linearization control without torque measurements
Capotondi, M., Turrisi, G., Gaz, C., Modugno, V., Oriolo, G. & De Luca, A., 30 Oct 2019.
Research output: Contribution to conference › Paper › peer-review

Cite this

Capotondi, M., Turrisi, G., Gaz, C., Modugno, V., Oriolo, G., & De Luca, A. (2019). An online learning procedure for feedback linearization control without torque measurements. In Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR), https://proceedings.mlr.press/v100/capotondi20a.html

Capotondi, M. ; Turrisi, G. ; Gaz, C. et al. / An online learning procedure for feedback linearization control without torque measurements. Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),. 2019.

@inproceedings{6aa9509691694a278717d8230939c2e3,

title = "An online learning procedure for feedback linearization control without torque measurements",

abstract = "By exploiting an a-priori estimate of the dynamic model of a manipulator, it is possible to command joint torques which ideally realize a Feedback Linearization (FL) controller. The exact cancellation may nevertheless not be achieved due to model uncertainties and possible errors in the estimation of the dynamic coefficients. In this work, an online learning scheme for control based on FL is presented. By reading joint positions and joint velocities information only (without the use of any torque measurement), we are able to learn those model uncertainties and thus achieve perfect FL control. Simulations results on the popular KUKA LWR iiwa robot are reported to show the quality of the proposed approach.",

keywords = "Computer science and informatics",

author = "M. Capotondi and G. Turrisi and C. Gaz and V. Modugno and G. Oriolo and \{De Luca\}, A.",

note = "Note: Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),",

year = "2019",

month = oct,

day = "30",

language = "English",

booktitle = "Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),",

}

Capotondi, M, Turrisi, G, Gaz, C, Modugno, V, Oriolo, G & De Luca, A 2019, An online learning procedure for feedback linearization control without torque measurements. in Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),. <https://proceedings.mlr.press/v100/capotondi20a.html>

An online learning procedure for feedback linearization control without torque measurements. / Capotondi, M.; Turrisi, G.; Gaz, C. et al.
Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),. 2019.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - An online learning procedure for feedback linearization control without torque measurements

AU - Capotondi, M.

AU - Turrisi, G.

AU - Gaz, C.

AU - Modugno, V.

AU - Oriolo, G.

AU - De Luca, A.

N1 - Note: Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),

PY - 2019/10/30

Y1 - 2019/10/30

N2 - By exploiting an a-priori estimate of the dynamic model of a manipulator, it is possible to command joint torques which ideally realize a Feedback Linearization (FL) controller. The exact cancellation may nevertheless not be achieved due to model uncertainties and possible errors in the estimation of the dynamic coefficients. In this work, an online learning scheme for control based on FL is presented. By reading joint positions and joint velocities information only (without the use of any torque measurement), we are able to learn those model uncertainties and thus achieve perfect FL control. Simulations results on the popular KUKA LWR iiwa robot are reported to show the quality of the proposed approach.

AB - By exploiting an a-priori estimate of the dynamic model of a manipulator, it is possible to command joint torques which ideally realize a Feedback Linearization (FL) controller. The exact cancellation may nevertheless not be achieved due to model uncertainties and possible errors in the estimation of the dynamic coefficients. In this work, an online learning scheme for control based on FL is presented. By reading joint positions and joint velocities information only (without the use of any torque measurement), we are able to learn those model uncertainties and thus achieve perfect FL control. Simulations results on the popular KUKA LWR iiwa robot are reported to show the quality of the proposed approach.

KW - Computer science and informatics

M3 - Conference contribution

BT - Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),

ER -

Capotondi M, Turrisi G, Gaz C, Modugno V, Oriolo G, De Luca A. An online learning procedure for feedback linearization control without torque measurements. In Published in: Kaelbling, L.P., Kragic, D. and Sugiura, K. (eds.) (2020) PMLR Proceedings of Machine Learning Research: Vol. 100: Conference on Robot Learning: ML Research Press. pp. 1359-1368, ISSN (online) 2640-3498 Organising Body: International Foundation of Robotics Research (IFRR), Organising Body: International Foundation of Robotics Research (IFRR),. 2019

An online learning procedure for feedback linearization control without torque measurements

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Research output

An online learning procedure for feedback linearization control without torque measurements

Cite this