Saved in:
Bibliographic Details
Main Authors: Feng, Yusen, Wang, Xiang, Yao, Heyuan, Kang, Zixi, Huo, Xinyu, Yu, Boyang, Qiu, Pengyun, Zhao, Ruijie, Chen, Baoquan, Liu, Libin
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.24592
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913159295533056
author Feng, Yusen
Wang, Xiang
Yao, Heyuan
Kang, Zixi
Huo, Xinyu
Yu, Boyang
Qiu, Pengyun
Zhao, Ruijie
Chen, Baoquan
Liu, Libin
author_facet Feng, Yusen
Wang, Xiang
Yao, Heyuan
Kang, Zixi
Huo, Xinyu
Yu, Boyang
Qiu, Pengyun
Zhao, Ruijie
Chen, Baoquan
Liu, Libin
contents This paper presents MuGen, a data-driven framework for learning and deploying multi-skill locomotion on humanoid robots. MuGen enables a robot to perform expressive motions like humans under the guidance of example motion sequences. To achieve this, we employ vector-quantized autoencoders (VQ-VAEs) trained with model-based reinforcement learning, resulting in a generative representation of locomotion that captures key patterns of human motion from hours of heterogeneous human performance data. We employ a teacher-student learning framework and develop a new policy distillation strategy to enable a deployable student policy learning this efficient latent representation. This policy allows the robot to track and mimic unseen human motions and further enables the robot to reuse the learned latent space for other tasks. We demonstrate the effectiveness of our framework through a diverse set of motions and accurate execution.
format Preprint
id arxiv_https___arxiv_org_abs_2605_24592
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MuGen: Multi-Skill Generative Locomotion Controller for Humanoid Robots
Feng, Yusen
Wang, Xiang
Yao, Heyuan
Kang, Zixi
Huo, Xinyu
Yu, Boyang
Qiu, Pengyun
Zhao, Ruijie
Chen, Baoquan
Liu, Libin
Robotics
This paper presents MuGen, a data-driven framework for learning and deploying multi-skill locomotion on humanoid robots. MuGen enables a robot to perform expressive motions like humans under the guidance of example motion sequences. To achieve this, we employ vector-quantized autoencoders (VQ-VAEs) trained with model-based reinforcement learning, resulting in a generative representation of locomotion that captures key patterns of human motion from hours of heterogeneous human performance data. We employ a teacher-student learning framework and develop a new policy distillation strategy to enable a deployable student policy learning this efficient latent representation. This policy allows the robot to track and mimic unseen human motions and further enables the robot to reuse the learned latent space for other tasks. We demonstrate the effectiveness of our framework through a diverse set of motions and accurate execution.
title MuGen: Multi-Skill Generative Locomotion Controller for Humanoid Robots
topic Robotics
url https://arxiv.org/abs/2605.24592