KungfuBot2: Learning Versatile Motion Skills for Humanoid Whole-Body Control
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_ | 1866909798531858432 |
|---|---|
| author | Han, Jinrui Xie, Weiji Zheng, Jiakun Shi, Jiyuan Zhang, Weinan Xiao, Ting Bai, Chenjia |
| author_facet | Han, Jinrui Xie, Weiji Zheng, Jiakun Shi, Jiyuan Zhang, Weinan Xiao, Ting Bai, Chenjia |
| contents | Learning versatile whole-body skills by tracking various human motions is a fundamental step toward general-purpose humanoid robots. This task is particularly challenging because a single policy must master a broad repertoire of motion skills while ensuring stability over long-horizon sequences. To this end, we present VMS, a unified whole-body controller that enables humanoid robots to learn diverse and dynamic behaviors within a single policy. Our framework integrates a hybrid tracking objective that balances local motion fidelity with global trajectory consistency, and an Orthogonal Mixture-of-Experts (OMoE) architecture that encourages skill specialization while enhancing generalization across motions. A segment-level tracking reward is further introduced to relax rigid step-wise matching, enhancing robustness when handling global displacements and transient inaccuracies. We validate VMS extensively in both simulation and real-world experiments, demonstrating accurate imitation of dynamic skills, stable performance over minute-long sequences, and strong generalization to unseen motions. These results highlight the potential of VMS as a scalable foundation for versatile humanoid whole-body control. The project page is available at https://kungfubot2-humanoid.github.io. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_16638 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | KungfuBot2: Learning Versatile Motion Skills for Humanoid Whole-Body Control Han, Jinrui Xie, Weiji Zheng, Jiakun Shi, Jiyuan Zhang, Weinan Xiao, Ting Bai, Chenjia Robotics Artificial Intelligence Learning versatile whole-body skills by tracking various human motions is a fundamental step toward general-purpose humanoid robots. This task is particularly challenging because a single policy must master a broad repertoire of motion skills while ensuring stability over long-horizon sequences. To this end, we present VMS, a unified whole-body controller that enables humanoid robots to learn diverse and dynamic behaviors within a single policy. Our framework integrates a hybrid tracking objective that balances local motion fidelity with global trajectory consistency, and an Orthogonal Mixture-of-Experts (OMoE) architecture that encourages skill specialization while enhancing generalization across motions. A segment-level tracking reward is further introduced to relax rigid step-wise matching, enhancing robustness when handling global displacements and transient inaccuracies. We validate VMS extensively in both simulation and real-world experiments, demonstrating accurate imitation of dynamic skills, stable performance over minute-long sequences, and strong generalization to unseen motions. These results highlight the potential of VMS as a scalable foundation for versatile humanoid whole-body control. The project page is available at https://kungfubot2-humanoid.github.io. |
| title | KungfuBot2: Learning Versatile Motion Skills for Humanoid Whole-Body Control |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2509.16638 |