Scaling Behavior Foundation Model for Humanoid Robots

Source

Source name: arXiv
Source kind: paper
Authors / organization: Weishuai Zeng, Kangning Yin, Xiaojie Niu, Shunlin Lu, Weixiang Zhong, Jiahe Chen, Feiyu Jia, Xiao Chen, Zirui Wang, Furui Xu, Ming Zhou, Kailin Li, Weinan Zhang, He Wang, Li Yi, Dahua Lin, Jiangmiao Pang, Jingbo Wang
Venue: arXiv: cs.RO
Published: 2026-07-16
Collected: 2026-07-17
URL: http://arxiv.org/abs/2607.15163v1
PDF: https://arxiv.org/pdf/2607.15163v1
DOI: UNKNOWN

Plain-Language Brief

Researchers propose a scaling recipe for Behavior Foundation Models (BFMs) to improve the capabilities of humanoid robots, focusing on the learning paradigm, synergy between rollout quantity and reference motion diversity, and an expressive model architecture called Humanoid Transformer. The approach showed significant improvements in control fidelity and task generalization in both simulation and real-world experiments.

What the Document Says

The text discusses the use of Behavior Foundation Models (BFMs) to address challenges in humanoid robot control such as natural whole-body coordination, precise responses, and generalization across diverse environments.
The authors claim that substantial performance gains can be achieved through the coordination of three core components: motion tracking learning paradigm, on-policy rollout quantity and reference motion diversity synergy, and Humanoid Transformer model architecture.
The study demonstrates improvements in control fidelity and task generalization through extensive experiments, resulting in reduced Mean Per-Keypoint Position Error (MPKPE) compared to existing humanoid controllers.

Key Concepts Mentioned

Behavior Foundation Models (BFMs)
Humanoid robots
Learning paradigm
On-policy rollout quantity
Reference motion diversity
Humanoid Transformer model architecture
Mean Per-Keypoint Position Error (MPKPE)

Claims Made by the Document

BFMs can improve humanoid robot capabilities.
Performance gains can be achieved by coordinating the learning paradigm, on-policy rollout quantity, and reference motion diversity.
The Humanoid Transformer model architecture facilitates structured behavioral representations.

Evidence Used

Improved control fidelity and task generalization as demonstrated through extensive experiments in simulation and real-world deployment.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Humanoid control requires natural whole-body coordination, precise real-time responses to control signals, and robust generalization across diverse environmental contexts, making it a cornerstone for generalist embodied agents. Behavior Foundation Models (BFMs) have recently emerged as a promising solution to address these challenges by leveraging large-scale behavioral data to achieve superior expressiveness, versatility and generalization. However, despite growing interest in scaling BFMs to further improve their capabilities, it remains unclear how key factors, including the learning paradigm, behavioral data and model architecture should be coordinated to enable effective scaling. In this work, we revisit the scaling recipe for BFMs and demonstrate that substantial performance gains can be achieved through the coordination of three core components: 1) the learning paradigm of motion tracking that reformulates diverse humanoid control problems as the reproduction of integrated whole-body behaviors in the global frame; 2) the strategic synergy between on-policy rollout quantity and reference motion diversity; and 3) the expressive and scalable model architecture termed Humanoid Transformer that facilitates the natural emergence of structured behavioral representations. Through extensive experiments in both simulation and real-world deployment, we demonstrate that our approach yields significant improvements in control fidelity and task generalization, reducing Mean Per-Keypoint Position Error (MPKPE) on the test set by over 10% in local mode and 82% in global mode compared with existing humanoid controllers. These results establish BFM as a principled and effective foundation for scalable and general-purpose humanoid control.