ZEST: Zero-shot Embodied Skill Transfer for Athletic Robot Control
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912864192692224 |
|---|---|
| author | Sleiman, Jean Pierre Li, He Adu-Bredu, Alphonsus Deits, Robin Kumar, Arun Bergamin, Kevin Bhardwaj, Mohak Biddlestone, Scott Burger, Nicola Estrada, Matthew A. Iacobelli, Francesco Koolen, Twan Lambert, Alexander Lin, Erica Mungai, M. Eva Nobles, Zach Rozen-Levy, Shane Shi, Yuyao Wang, Jiashun Welner, Jakob Yu, Fangzhou Zhang, Mike Rizzi, Alfred Hodgins, Jessica Bertrand, Sylvain Abe, Yeuhi Kuindersma, Scott Farshidian, Farbod |
| author_facet | Sleiman, Jean Pierre Li, He Adu-Bredu, Alphonsus Deits, Robin Kumar, Arun Bergamin, Kevin Bhardwaj, Mohak Biddlestone, Scott Burger, Nicola Estrada, Matthew A. Iacobelli, Francesco Koolen, Twan Lambert, Alexander Lin, Erica Mungai, M. Eva Nobles, Zach Rozen-Levy, Shane Shi, Yuyao Wang, Jiashun Welner, Jakob Yu, Fangzhou Zhang, Mike Rizzi, Alfred Hodgins, Jessica Bertrand, Sylvain Abe, Yeuhi Kuindersma, Scott Farshidian, Farbod |
| contents | Achieving robust, human-like whole-body control on humanoid robots for agile, contact-rich behaviors remains a central challenge, demanding heavy per-skill engineering and a brittle process of tuning controllers. We introduce ZEST (Zero-shot Embodied Skill Transfer), a streamlined motion-imitation framework that trains policies via reinforcement learning from diverse sources -- high-fidelity motion capture, noisy monocular video, and non-physics-constrained animation -- and deploys them to hardware zero-shot. ZEST generalizes across behaviors and platforms while avoiding contact labels, reference or observation windows, state estimators, and extensive reward shaping. Its training pipeline combines adaptive sampling, which focuses training on difficult motion segments, and an automatic curriculum using a model-based assistive wrench, together enabling dynamic, long-horizon maneuvers. We further provide a procedure for selecting joint-level gains from approximate analytical armature values for closed-chain actuators, along with a refined model of actuators. Trained entirely in simulation with moderate domain randomization, ZEST demonstrates remarkable generality. On Boston Dynamics' Atlas humanoid, ZEST learns dynamic, multi-contact skills (e.g., army crawl, breakdancing) from motion capture. It transfers expressive dance and scene-interaction skills, such as box-climbing, directly from videos to Atlas and the Unitree G1. Furthermore, it extends across morphologies to the Spot quadruped, enabling acrobatics, such as a continuous backflip, through animation. Together, these results demonstrate robust zero-shot deployment across heterogeneous data sources and embodiments, establishing ZEST as a scalable interface between biological movements and their robotic counterparts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_00401 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ZEST: Zero-shot Embodied Skill Transfer for Athletic Robot Control Sleiman, Jean Pierre Li, He Adu-Bredu, Alphonsus Deits, Robin Kumar, Arun Bergamin, Kevin Bhardwaj, Mohak Biddlestone, Scott Burger, Nicola Estrada, Matthew A. Iacobelli, Francesco Koolen, Twan Lambert, Alexander Lin, Erica Mungai, M. Eva Nobles, Zach Rozen-Levy, Shane Shi, Yuyao Wang, Jiashun Welner, Jakob Yu, Fangzhou Zhang, Mike Rizzi, Alfred Hodgins, Jessica Bertrand, Sylvain Abe, Yeuhi Kuindersma, Scott Farshidian, Farbod Robotics Artificial Intelligence Achieving robust, human-like whole-body control on humanoid robots for agile, contact-rich behaviors remains a central challenge, demanding heavy per-skill engineering and a brittle process of tuning controllers. We introduce ZEST (Zero-shot Embodied Skill Transfer), a streamlined motion-imitation framework that trains policies via reinforcement learning from diverse sources -- high-fidelity motion capture, noisy monocular video, and non-physics-constrained animation -- and deploys them to hardware zero-shot. ZEST generalizes across behaviors and platforms while avoiding contact labels, reference or observation windows, state estimators, and extensive reward shaping. Its training pipeline combines adaptive sampling, which focuses training on difficult motion segments, and an automatic curriculum using a model-based assistive wrench, together enabling dynamic, long-horizon maneuvers. We further provide a procedure for selecting joint-level gains from approximate analytical armature values for closed-chain actuators, along with a refined model of actuators. Trained entirely in simulation with moderate domain randomization, ZEST demonstrates remarkable generality. On Boston Dynamics' Atlas humanoid, ZEST learns dynamic, multi-contact skills (e.g., army crawl, breakdancing) from motion capture. It transfers expressive dance and scene-interaction skills, such as box-climbing, directly from videos to Atlas and the Unitree G1. Furthermore, it extends across morphologies to the Spot quadruped, enabling acrobatics, such as a continuous backflip, through animation. Together, these results demonstrate robust zero-shot deployment across heterogeneous data sources and embodiments, establishing ZEST as a scalable interface between biological movements and their robotic counterparts. |
| title | ZEST: Zero-shot Embodied Skill Transfer for Athletic Robot Control |
| topic | Robotics Artificial Intelligence |
| url | https://arxiv.org/abs/2602.00401 |