CAIMAN: Causal Action Influence Detection for Sample-efficient Loco-manipulation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912935561920512 |
|---|---|
| author | Yuan, Yuanchen Cheng, Jin Urpí, Núria Armengol Coros, Stelian |
| author_facet | Yuan, Yuanchen Cheng, Jin Urpí, Núria Armengol Coros, Stelian |
| contents | Enabling legged robots to perform non-prehensile loco-manipulation is crucial for enhancing their versatility. Learning behaviors such as whole-body object pushing often requires sophisticated planning strategies or extensive task-specific reward shaping, especially in unstructured environments. In this work, we present CAIMAN, a practical reinforcement learning framework that encourages the agent to gain control over other entities in the environment. CAIMAN leverages causal action influence as an intrinsic motivation objective, allowing legged robots to efficiently acquire object pushing skills even under sparse task rewards. We employ a hierarchical control strategy, combining a low-level locomotion module with a high-level policy that generates task-relevant velocity commands and is trained to maximize the intrinsic reward. To estimate causal action influence, we learn the dynamics of the environment by integrating a kinematic prior with data collected during training. We empirically demonstrate CAIMAN's superior sample efficiency and adaptability to diverse scenarios in simulation, as well as its successful transfer to real-world systems without further fine-tuning. A video demo is available at https://www.youtube.com/watch?v=dNyvT04Cqaw. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_00835 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CAIMAN: Causal Action Influence Detection for Sample-efficient Loco-manipulation Yuan, Yuanchen Cheng, Jin Urpí, Núria Armengol Coros, Stelian Robotics Machine Learning Enabling legged robots to perform non-prehensile loco-manipulation is crucial for enhancing their versatility. Learning behaviors such as whole-body object pushing often requires sophisticated planning strategies or extensive task-specific reward shaping, especially in unstructured environments. In this work, we present CAIMAN, a practical reinforcement learning framework that encourages the agent to gain control over other entities in the environment. CAIMAN leverages causal action influence as an intrinsic motivation objective, allowing legged robots to efficiently acquire object pushing skills even under sparse task rewards. We employ a hierarchical control strategy, combining a low-level locomotion module with a high-level policy that generates task-relevant velocity commands and is trained to maximize the intrinsic reward. To estimate causal action influence, we learn the dynamics of the environment by integrating a kinematic prior with data collected during training. We empirically demonstrate CAIMAN's superior sample efficiency and adaptability to diverse scenarios in simulation, as well as its successful transfer to real-world systems without further fine-tuning. A video demo is available at https://www.youtube.com/watch?v=dNyvT04Cqaw. |
| title | CAIMAN: Causal Action Influence Detection for Sample-efficient Loco-manipulation |
| topic | Robotics Machine Learning |
| url | https://arxiv.org/abs/2502.00835 |