Laux, Melvin ; Arenz, Oleg ; Peters, Jan ; Pajarinen, Joni (2022)
Deep Adversarial Reinforcement Learning for Object Disentangling.
2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Las Vegas, USA (Virtual) (24.01.2021-24.01.2021)
doi: 10.26083/tuprints-00022926
Konferenzveröffentlichung, Zweitveröffentlichung, Postprint
 | Es ist eine neuere Version dieses Eintrags verfügbar. |
Kurzbeschreibung (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.
| Typ des Eintrags: | Konferenzveröffentlichung |
|---|---|
| Erschienen: | 2022 |
| Autor(en): | Laux, Melvin ; Arenz, Oleg ; Peters, Jan ; Pajarinen, Joni |
| Art des Eintrags: | Zweitveröffentlichung |
| Titel: | Deep Adversarial Reinforcement Learning for Object Disentangling |
| Sprache: | Englisch |
| Publikationsjahr: | 2022 |
| Ort: | Darmstadt |
| Publikationsdatum der Erstveröffentlichung: | 2021 |
| Verlag: | IEEE |
| Buchtitel: | 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) |
| Kollation: | 7 ungezählte Seiten |
| Veranstaltungstitel: | 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) |
| Veranstaltungsort: | Las Vegas, USA (Virtual) |
| Veranstaltungsdatum: | 24.01.2021-24.01.2021 |
| DOI: | 10.26083/tuprints-00022926 |
| URL / URN: | https://tuprints.ulb.tu-darmstadt.de/22926 |
| Zugehörige Links: | |
| Herkunft: | Zweitveröffentlichungsservice |
| Kurzbeschreibung (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. |
| Freie Schlagworte: | Training, Visualization, Sensitivity, Shape, Reinforcement learning, Task analysis, Robots |
| Status: | Postprint |
| URN: | urn:nbn:de:tuda-tuprints-229264 |
| Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik |
| Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik 20 Fachbereich Informatik > Intelligente Autonome Systeme |
| Hinterlegungsdatum: | 25 Nov 2022 12:45 |
| Letzte Änderung: | 17 Mai 2023 12:48 |
| PPN: | 503350842 |
| Export: | |
| Suche nach Titel in: | TUfind oder in Google |
Verfügbare Versionen dieses Eintrags
- Deep Adversarial Reinforcement Learning for Object Disentangling. (deposited 25 Nov 2022 12:45) [Gegenwärtig angezeigt]
 |
Frage zum Eintrag |
Optionen (nur für Redakteure)
 |
Redaktionelle Details anzeigen |
Drucken
Impressum