Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression

Ali Al-Yacoub, Sara Sharifzadeh, Niels Lohse, Zahid Usman, Yee Mey Goh, Mike Jackson

Research output: Chapter in Book/Report/Conference proceedingConference proceeding

Abstract

Haptics, as well as force and torque measurements, are increasingly gaining attention in the fields of kinesthetic learning and robot Learning from demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6-axis force and torque sensor mounted at the end effector of an industrial robot are known to be corrupted due to the robots internal forces, gravity, un-modelled dynamics and nonlinear effects. This paper presents an evaluation of two techniques, SVR and Random Forests, to recover the external forces and accurately selected possible contact situations by estimating a robots internal forces. The performance of the learned models have been evaluated using different performance metrics and comparing them with respect to the features contained in the input space. Both SVR and Random Forests require low computational complexity without intensive training over the operational space under the given ssumptions. In addition, these methods do not need data to be available online. The SVR and Random Forests models are experimentally validated using Motoman SDA10D dual-arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on Random Forests has outperformed Support Vector Regression.
Original languageEnglish
Title of host publication(846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016
PublisherActa Press
ISBN (Electronic)9780889869868
DOIs
Publication statusPublished - 2016
EventIntelligent Systems and Robotics - Zurich, Switzerland
Duration: 6 Oct 20168 Oct 2016
Conference number: 1
http://www.iasted.org/conferences/pastinfo-847.html

Conference

ConferenceIntelligent Systems and Robotics
Abbreviated titleISAR 2016
CountrySwitzerland
CityZurich
Period6/10/168/10/16
Internet address

Fingerprint

Robot learning
Torque measurement
Force measurement
Torque
Robots
Robotic arms
Industrial robots
End effectors
Computational complexity
Gravitation
Demonstrations
Sensors
Compensation and Redress
Experiments

Cite this

Al-Yacoub, A., Sharifzadeh, S., Lohse, N., Usman, Z., Goh, Y. M., & Jackson, M. (2016). Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression. In (846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016 Acta Press. https://doi.org/10.2316/P.2016.847-002

Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression. / Al-Yacoub, Ali; Sharifzadeh, Sara; Lohse, Niels; Usman, Zahid; Goh, Yee Mey; Jackson, Mike.

(846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016. Acta Press, 2016.

Research output: Chapter in Book/Report/Conference proceedingConference proceeding

Al-Yacoub, A, Sharifzadeh, S, Lohse, N, Usman, Z, Goh, YM & Jackson, M 2016, Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression. in (846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016. Acta Press, Intelligent Systems and Robotics, Zurich, Switzerland, 6/10/16. https://doi.org/10.2316/P.2016.847-002
Al-Yacoub A, Sharifzadeh S, Lohse N, Usman Z, Goh YM, Jackson M. Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression. In (846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016. Acta Press. 2016 https://doi.org/10.2316/P.2016.847-002
Al-Yacoub, Ali ; Sharifzadeh, Sara ; Lohse, Niels ; Usman, Zahid ; Goh, Yee Mey ; Jackson, Mike. / Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression. (846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016. Acta Press, 2016.
@inproceedings{c1c851e2b1ad43169b7cdc821d663342,
title = "Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression",
abstract = "Haptics, as well as force and torque measurements, are increasingly gaining attention in the fields of kinesthetic learning and robot Learning from demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6-axis force and torque sensor mounted at the end effector of an industrial robot are known to be corrupted due to the robots internal forces, gravity, un-modelled dynamics and nonlinear effects. This paper presents an evaluation of two techniques, SVR and Random Forests, to recover the external forces and accurately selected possible contact situations by estimating a robots internal forces. The performance of the learned models have been evaluated using different performance metrics and comparing them with respect to the features contained in the input space. Both SVR and Random Forests require low computational complexity without intensive training over the operational space under the given ssumptions. In addition, these methods do not need data to be available online. The SVR and Random Forests models are experimentally validated using Motoman SDA10D dual-arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on Random Forests has outperformed Support Vector Regression.",
author = "Ali Al-Yacoub and Sara Sharifzadeh and Niels Lohse and Zahid Usman and Goh, {Yee Mey} and Mike Jackson",
year = "2016",
doi = "10.2316/P.2016.847-002",
language = "English",
booktitle = "(846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016",
publisher = "Acta Press",
address = "Canada",

}

TY - GEN

T1 - Learning Industrial Robot Force/torque Compensation: A Comparison of Support Vector and Random Forests Regression

AU - Al-Yacoub, Ali

AU - Sharifzadeh, Sara

AU - Lohse, Niels

AU - Usman, Zahid

AU - Goh, Yee Mey

AU - Jackson, Mike

PY - 2016

Y1 - 2016

N2 - Haptics, as well as force and torque measurements, are increasingly gaining attention in the fields of kinesthetic learning and robot Learning from demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6-axis force and torque sensor mounted at the end effector of an industrial robot are known to be corrupted due to the robots internal forces, gravity, un-modelled dynamics and nonlinear effects. This paper presents an evaluation of two techniques, SVR and Random Forests, to recover the external forces and accurately selected possible contact situations by estimating a robots internal forces. The performance of the learned models have been evaluated using different performance metrics and comparing them with respect to the features contained in the input space. Both SVR and Random Forests require low computational complexity without intensive training over the operational space under the given ssumptions. In addition, these methods do not need data to be available online. The SVR and Random Forests models are experimentally validated using Motoman SDA10D dual-arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on Random Forests has outperformed Support Vector Regression.

AB - Haptics, as well as force and torque measurements, are increasingly gaining attention in the fields of kinesthetic learning and robot Learning from demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6-axis force and torque sensor mounted at the end effector of an industrial robot are known to be corrupted due to the robots internal forces, gravity, un-modelled dynamics and nonlinear effects. This paper presents an evaluation of two techniques, SVR and Random Forests, to recover the external forces and accurately selected possible contact situations by estimating a robots internal forces. The performance of the learned models have been evaluated using different performance metrics and comparing them with respect to the features contained in the input space. Both SVR and Random Forests require low computational complexity without intensive training over the operational space under the given ssumptions. In addition, these methods do not need data to be available online. The SVR and Random Forests models are experimentally validated using Motoman SDA10D dual-arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on Random Forests has outperformed Support Vector Regression.

U2 - 10.2316/P.2016.847-002

DO - 10.2316/P.2016.847-002

M3 - Conference proceeding

BT - (846) Telehealth and Assistive Technology / 847: Intelligent Systems and Robotics - 2016

PB - Acta Press

ER -