BeyondMimic: From motion tracking to versatile humanoid control via guided diffusion
Science Robotics · 26.08.2026 · 11 (117)
DOI: 10.1126/scirobotics.adx8924 ↗Kurz gefasst
Fachartikel in der Zeitschrift Science Robotics, erschienen am 26.08.2026. Verfasst von Qiayuan Liao und 6 weiteren Autor:innen. Beteiligte Einrichtung: University of California, Berkeley, Berkeley, CA 94720, USA. Hauptthemen laut MEDLINE: Robotics. Die Arbeit wurde bisher 21-mal zitiert (Crossref) und nennt 99 Literaturangaben. Gefördert von: NSF Office of the Director, Defense Advanced Research Projects Agency, Wu-Tsai Human Performance Alliance.
Abstract (Originalsprache)
The humanlike form of humanoid robots uniquely positions them to achieve the agility and versatility in motor skills that humans have. Learning from human demonstrations offers a scalable approach to acquiring these capabilities. However, prior works either produced unnatural motions or relied on motion-specific tuning to achieve satisfactory naturalness. Furthermore, these methods are often motion or goal specific, lacking the versatility to compose diverse skills, especially when solving unseen tasks. We present BeyondMimic, a framework that scales to diverse motions and carries the versatility to compose them seamlessly in tackling unseen downstream tasks. A compact motion tracking formulation enables mastery of a wide range of highly agile behaviors, including aerial cartwheels, spin kicks, flip kicks, and sprinting, with a single setup and shared hyperparameters, all while achieving humanlike performance. Moving beyond the mere imitation of existing motions, we propose a unified latent diffusion model that empowers versatile goal specification, seamless task switching, and dynamic composition of these agile behaviors. Leveraging classifier guidance, a diffusion-specific technique for test-time optimization toward unseen objectives, our model extended its capability to solve downstream tasks never encountered during training, including motion inpainting, joystick teleoperation, and obstacle avoidance, and transferred these skills zero-shot to real hardware. Together, these components enable scalable acquisition of humanlike motor skills from human motion and motion synthesis that generalizes and adapts beyond the training setup.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- University of California, Berkeley, Berkeley, CA 94720, USA.
- Stanford University, Stanford, CA 94305, USA.
ORCID-Profile: Qiayuan Liao, Takara E Truong, Xiaoyu Huang, Yuman Gao, Guy Tevet, Koushil Sreenath, C Karen Liu
Förderung
- NSF Office of the Director
- Defense Advanced Research Projects Agency
- Wu-Tsai Human Performance Alliance
- Stanford Institute for Human-Centered AI
- The Robotics and AI Institute
- BAIR Humanoid Intelligence Center
- Design of Robustly Implementable Autonomous and Intelligent Machines
Zitieren
APA
Liao, Q., Truong, T. E., Huang, X., Gao, Y., Tevet, G., Sreenath, K., & Liu, C. K. (2026). BeyondMimic: From motion tracking to versatile humanoid control via guided diffusion. Science Robotics, 11(117). https://doi.org/10.1126/scirobotics.adx8924
Vancouver
Liao Q, Truong TE, Huang X, Gao Y, Tevet G, Sreenath K, et al. BeyondMimic: From motion tracking to versatile humanoid control via guided diffusion. Science Robotics. 2026;11(117). doi:10.1126/scirobotics.adx8924
BibTeX
@article{Liao2026,
author = {Liao, Qiayuan and Truong, Takara E and Huang, Xiaoyu and Gao, Yuman and Tevet, Guy and Sreenath, Koushil and Liu, C Karen},
title = {BeyondMimic: From motion tracking to versatile humanoid control via guided diffusion},
journal = {Science Robotics},
year = {2026},
volume = {11},
number = {117},
doi = {10.1126/scirobotics.adx8924}
}
EUVIMED