Salvato in:
| Autori principali: | , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2509.13833 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915523751575552 |
|---|---|
| author | Zhang, Zhikai Guo, Jun Chen, Chao Wang, Jilong Lin, Chenghuai Lian, Yunrui Xue, Han Wang, Zhenrong Liu, Maoqi Lyu, Jiangran Liu, Huaping Wang, He Yi, Li |
| author_facet | Zhang, Zhikai Guo, Jun Chen, Chao Wang, Jilong Lin, Chenghuai Lian, Yunrui Xue, Han Wang, Zhenrong Liu, Maoqi Lyu, Jiangran Liu, Huaping Wang, He Yi, Li |
| contents | A foundational humanoid motion tracker is expected to be able to track diverse, highly dynamic, and contact-rich motions. More importantly, it needs to operate stably in real-world scenarios against various dynamics disturbances, including terrains, external forces, and physical property changes for general practical use. To achieve this goal, we propose Any2Track (Track Any motions under Any disturbances), a two-stage RL framework to track various motions under multiple disturbances in the real world. Any2Track reformulates dynamics adaptability as an additional capability on top of basic action execution and consists of two key components: AnyTracker and AnyAdapter. AnyTracker is a general motion tracker with a series of careful designs to track various motions within a single policy. AnyAdapter is a history-informed adaptation module that endows the tracker with online dynamics adaptability to overcome the sim2real gap and multiple real-world disturbances. We deploy Any2Track on Unitree G1 hardware and achieve a successful sim2real transfer in a zero-shot manner. Any2Track performs exceptionally well in tracking various motions under multiple real-world disturbances. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_13833 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Track Any Motions under Any Disturbances Zhang, Zhikai Guo, Jun Chen, Chao Wang, Jilong Lin, Chenghuai Lian, Yunrui Xue, Han Wang, Zhenrong Liu, Maoqi Lyu, Jiangran Liu, Huaping Wang, He Yi, Li Robotics A foundational humanoid motion tracker is expected to be able to track diverse, highly dynamic, and contact-rich motions. More importantly, it needs to operate stably in real-world scenarios against various dynamics disturbances, including terrains, external forces, and physical property changes for general practical use. To achieve this goal, we propose Any2Track (Track Any motions under Any disturbances), a two-stage RL framework to track various motions under multiple disturbances in the real world. Any2Track reformulates dynamics adaptability as an additional capability on top of basic action execution and consists of two key components: AnyTracker and AnyAdapter. AnyTracker is a general motion tracker with a series of careful designs to track various motions within a single policy. AnyAdapter is a history-informed adaptation module that endows the tracker with online dynamics adaptability to overcome the sim2real gap and multiple real-world disturbances. We deploy Any2Track on Unitree G1 hardware and achieve a successful sim2real transfer in a zero-shot manner. Any2Track performs exceptionally well in tracking various motions under multiple real-world disturbances. |
| title | Track Any Motions under Any Disturbances |
| topic | Robotics |
| url | https://arxiv.org/abs/2509.13833 |