TTT-Parkour: Rapid Test-Time Training for Perceptive Robot Parkour
Shaoting Zhu, Baijun Ye, Jiaxuan Wang, Jiakang Chen, Ziwen Zhuang, Linzhan Mou, Runhan Huang, Hang Zhao · 2026
Read TTT-Parkour: Rapid Test-Time Training for Perceptive Robot Parkour by Shaoting Zhu, Baijun Ye, Jiaxuan Wang, Jiakang Chen, Ziwen Zhuang, Linzhan Mou, Runhan Huang, Hang Zhao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-02331.txt.
Description
Shaoting Zhu and colleagues propose TTT-Parkour, rapid test-time training for perceptive humanoid parkour on unseen complex terrain. The framework fine-tunes an environment-specific policy in simulation before deployment on newly encountered terrain. A policy first pre-trains on diverse procedurally generated terrains, then fine-tunes rapidly on high-fidelity meshes reconstructed from real captures. Humanoids then master wedges, stakes, boxes, trapezoids, and narrow beams, with the pipeline taking about ten minutes per terrain and sim-to-real deployment without further on-robot updates.
Catalog record
| CREATOR | Shaoting Zhu, Baijun Ye, Jiaxuan Wang, Jiakang Chen, Ziwen Zhuang, Linzhan Mou, Runhan Huang, Hang Zhao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41377 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602.02331