Learning to Close the Gap: Combining Task Frame Formalism and Reinforcement Learning for Compliant Vegetable Cutting

Abhishek Padalkar, Matthias Nieuwenhuisen, Sven Schneider, Dirk Schulz

2020

Abstract

Compliant manipulation is a crucial skill for robots when they are supposed to act as helping hands in everyday household tasks. Still, nowadays, those skills are hand-crafted by experts which frequently requires labor-intensive, manual parameter tuning. Moreover, some tasks are too complex to be specified fully using a task specification. Learning these skills, by contrast, requires a high number of costly and potentially unsafe interactions with the environment. We present a compliant manipulation approach using reinforcement learning guided by the Task Frame Formalism, a task specification method. This allows us to specify the easy to model knowledge about a task while the robot learns the unmodeled components by reinforcement learning. We evaluate the approach by performing a compliant manipulation task with a KUKA LWR 4+ manipulator. The robot was able to learn force control policies directly on the robot without using any simulation.

Paper Citation

in Harvard Style

Padalkar A., Nieuwenhuisen M., Schneider S. and Schulz D. (2020). Learning to Close the Gap: Combining Task Frame Formalism and Reinforcement Learning for Compliant Vegetable Cutting.In Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO, ISBN 978-989-758-442-8, pages 221-231. DOI: 10.5220/0009590602210231

in Bibtex Style

@conference{icinco20,
author={Abhishek Padalkar and Matthias Nieuwenhuisen and Sven Schneider and Dirk Schulz},
title={Learning to Close the Gap: Combining Task Frame Formalism and Reinforcement Learning for Compliant Vegetable Cutting},
booktitle={Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,},
year={2020},
pages={221-231},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0009590602210231},
isbn={978-989-758-442-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,
TI - Learning to Close the Gap: Combining Task Frame Formalism and Reinforcement Learning for Compliant Vegetable Cutting
SN - 978-989-758-442-8
AU - Padalkar A.
AU - Nieuwenhuisen M.
AU - Schneider S.
AU - Schulz D.
PY - 2020
SP - 221
EP - 231
DO - 10.5220/0009590602210231