Deep adversarial reinforcement learning for object disentangling

Melvin Laux, Oleg Arenz, Jan Peters, Joni Pajarinen

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

Abstract

Deep learning in combination with improved training techniques and high computational power has led to recent advances in the field of reinforcement learning (RL) and to successful robotic RL applications such as in-hand manipulation. However, most robotic RL relies on a well known initial state distribution. In real-world tasks, this information is however often not available. For example, when disentangling waste objects the actual position of the robot w.r.t. the objects may not match the positions the RL policy was trained for. To solve this problem, we present a novel adversarial reinforcement learning (ARL) framework. The ARL framework utilizes an adversary, which is trained to steer the original agent, the protagonist, to challenging states. We train the protagonist and the adversary jointly to allow them to adapt to the changing policy of their opponent. We show that our method can generalize from training to test scenarios by training an end-to-end system for robot control to solve a challenging object disentangling task. Experiments with a KUKA LBR+ 7-DOF robot arm show that our approach outperforms the baseline method in disentangling when starting from different initial states than provided during training.

Original languageEnglish
Title of host publication2020 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2020
PublisherIEEE
Pages5504-5510
Number of pages7
ISBN (Electronic)9781728162126
DOIs
Publication statusPublished - 24 Oct 2020
Publication typeA4 Article in a conference publication
EventIEEE/RSJ International Conference on Intelligent Robots and Systems -
Duration: 24 Oct 202024 Jan 2021

Publication series

NameIEEE International Conference on Intelligent Robots and Systems
ISSN (Print)2153-0858
ISSN (Electronic)2153-0866

Conference

ConferenceIEEE/RSJ International Conference on Intelligent Robots and Systems
Period24/10/2024/01/21

Publication forum classification

  • Publication forum level 1

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Software
  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'Deep adversarial reinforcement learning for object disentangling'. Together they form a unique fingerprint.

Cite this