TU Darmstadt / ULB / TUprints

High Acceleration Reinforcement Learning for Real-World Juggling with Binary Rewards

Ploeger, Kai ; Lutter, Michael ; Peters, Jan (2022):
High Acceleration Reinforcement Learning for Real-World Juggling with Binary Rewards. (Publisher's Version)
In: Proceedings of Machine Learning Research, 155, In: Proceedings of the 2020 Conference on Robot Learning, pp. 642-653,
Darmstadt, PMLR, Conference on Robot Learning (CoRL) 2020, Cambridge MA, USA, 16.-18.11.2020, DOI: 10.26083/tuprints-00020583,
[Conference or Workshop Item]

[img] Text
ploeger21a.pdf
Copyright Information: CC BY 4.0 International - Creative Commons, Attribution.

Download (2MB)
Item Type: Conference or Workshop Item
Origin: Secondary publication service
Status: Publisher's Version
Title: High Acceleration Reinforcement Learning for Real-World Juggling with Binary Rewards
Language: English
Abstract:

Robots that can learn in the physical world will be important to enable robots to escape their stiff and pre-programmed movements. For dynamic high-acceleration tasks, such as juggling, learning in the real-world is particularly challenging as one must push the limits of the robot and its actuation without harming the system, amplifying the necessity of sample efficiency and safety for robot learning algorithms. In contrast to prior work which mainly focuses on the learning algorithm, we propose a learning system, that directly incorporates these requirements in the design of the policy representation, initialization, and optimization. We demonstrate that this system enables the high-speed Barrett WAM manipulator to learn juggling two balls from 56 minutes of experience with a binary reward signal and finally juggles continuously for up to 33 minutes or about 4500 repeated catches. The videos documenting the learning process and the evaluation can be found at https://sites.google.com/view/jugglingbot

Book Title: Proceedings of the 2020 Conference on Robot Learning
Series: Proceedings of Machine Learning Research
Series Volume: 155
Place of Publication: Darmstadt
Publisher: PMLR
Collation: 12 Seiten
Uncontrolled Keywords: Reinforcement Learning, Dynamic Manipulation, Juggling
Classification DDC: 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik
Divisions: 20 Department of Computer Science > Intelligent Autonomous Systems
TU-Projects: EC/H2020|640554|SKILLS4ROBOTS
Event Title: Conference on Robot Learning (CoRL) 2020
Event Location: Cambridge MA, USA
Event Dates: 16.-18.11.2020
Date Deposited: 18 Nov 2022 14:40
Last Modified: 18 Nov 2022 14:40
DOI: 10.26083/tuprints-00020583
Corresponding Links:
URN: urn:nbn:de:tuda-tuprints-205834
URI: https://tuprints.ulb.tu-darmstadt.de/id/eprint/20583
PPN:
Export:
Actions (login required)
View Item View Item