| |||||||
ShanghaiTech University Knowledge Management System
| Humanoid Parkour Learning | |
| 2024-09-26 | |
| 状态 | 已发表 |
| 摘要 | Parkour is a grand challenge for legged locomotion, even for quadruped robots, requiring active perception and various maneuvers to overcome multiple challenging obstacles. Existing methods for humanoid locomotion either optimize a trajectory for a single parkour track or train a reinforcement learning policy only to walk with a significant amount of motion references. In this work, we propose a framework for learning an end-to-end vision-based whole-body-control parkour policy for humanoid robots that overcomes multiple parkour skills without any motion prior. Using the parkour policy, the humanoid robot can jump on a 0.42m platform, leap over hurdles, 0.8m gaps, and much more. It can also run at 1.8m/s in the wild and walk robustly on different terrains. We test our policy in indoor and outdoor environments to demonstrate that it can autonomously select parkour skills while following the rotation command of the joystick. We override the arm actions and show that this framework can easily transfer to humanoid mobile manipulation tasks.  |
| 关键词 | Humanoid Agile Locomotion Visuomotor Control Sim-to-Real Transfer |
| 语种 | 英语 |
| DOI | arXiv:2406.10759 |
| 相关网址 | 查看原文 |
| 出处 | Arxiv |
| 收录类别 | PPRN.PPRN |
| WOS记录号 | PPRN:89352573 |
| WOS类目 | Computer Science, Artificial Intelligence |
| 文献类型 | 预印本 |
| 条目标识符 | https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/433536 |
| 专题 | 信息科学与技术学院_硕士生 信息科学与技术学院_本科生 |
| 通讯作者 | Zhao, Hang |
| 作者单位 | 1.Shanghai Qi Zhi Inst, Shanghai, Peoples R China 2.ShanghaiTech Univ, Shanghai, Peoples R China 3.Tsinghua Univ, Beijing, Peoples R China |
| 推荐引用方式 GB/T 7714 | Zhuang, Ziwen,Yao, Shenzhe,Zhao, Hang. Humanoid Parkour Learning. 2024. |
| 条目包含的文件 | ||||||
| 条目无相关文件。 | ||||||
修改评论
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。