AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912550694682624 |
|---|---|
| author | Fang, Hongjie Wang, Chenxi Wang, Yiming Chen, Jingjing Xia, Shangning Lv, Jun He, Zihao Yi, Xiyan Guo, Yunhan Zhan, Xinyu Yang, Lixin Wang, Weiming Lu, Cewu Fang, Hao-Shu |
| author_facet | Fang, Hongjie Wang, Chenxi Wang, Yiming Chen, Jingjing Xia, Shangning Lv, Jun He, Zihao Yi, Xiyan Guo, Yunhan Zhan, Xinyu Yang, Lixin Wang, Weiming Lu, Cewu Fang, Hao-Shu |
| contents | Scaling up robotic imitation learning for real-world applications requires efficient and scalable demonstration collection methods. While teleoperation is effective, it depends on costly and inflexible robot platforms. In-the-wild demonstrations offer a promising alternative, but existing collection devices have key limitations: handheld setups offer limited observational coverage, and whole-body systems often require fine-tuning with robot data due to domain gaps. To address these challenges, we present AirExo-2, a low-cost exoskeleton system for large-scale in-the-wild data collection, along with several adaptors that transform collected data into pseudo-robot demonstrations suitable for policy learning. We further introduce RISE-2, a generalizable imitation learning policy that fuses 3D spatial and 2D semantic perception for robust manipulations. Experiments show that RISE-2 outperforms prior state-of-the-art methods on both in-domain and generalization evaluations. Trained solely on adapted in-the-wild data produced by AirExo-2, the RISE-2 policy achieves comparable performance to the policy trained with teleoperated data, highlighting the effectiveness and potential of AirExo-2 for scalable and generalizable imitation learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_03081 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons Fang, Hongjie Wang, Chenxi Wang, Yiming Chen, Jingjing Xia, Shangning Lv, Jun He, Zihao Yi, Xiyan Guo, Yunhan Zhan, Xinyu Yang, Lixin Wang, Weiming Lu, Cewu Fang, Hao-Shu Robotics Scaling up robotic imitation learning for real-world applications requires efficient and scalable demonstration collection methods. While teleoperation is effective, it depends on costly and inflexible robot platforms. In-the-wild demonstrations offer a promising alternative, but existing collection devices have key limitations: handheld setups offer limited observational coverage, and whole-body systems often require fine-tuning with robot data due to domain gaps. To address these challenges, we present AirExo-2, a low-cost exoskeleton system for large-scale in-the-wild data collection, along with several adaptors that transform collected data into pseudo-robot demonstrations suitable for policy learning. We further introduce RISE-2, a generalizable imitation learning policy that fuses 3D spatial and 2D semantic perception for robust manipulations. Experiments show that RISE-2 outperforms prior state-of-the-art methods on both in-domain and generalization evaluations. Trained solely on adapted in-the-wild data produced by AirExo-2, the RISE-2 policy achieves comparable performance to the policy trained with teleoperated data, highlighting the effectiveness and potential of AirExo-2 for scalable and generalizable imitation learning. |
| title | AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletons |
| topic | Robotics |
| url | https://arxiv.org/abs/2503.03081 |