论文/arXiv 2610.17999·15 天前

TRACE：面向跨类人机器人本体的神经全身控制高效迁移学习

TRACE通过运动学先验、蒸馏与低秩强化学习，将SONIC从Unitree G1迁移到H2并在评估动作和真机演示中验证。

AI 整理

• 本体 SDK 与控制
• 数据与采集
• 具身模型

宇树 G1

原文链接：TRACE: Efficient Transfer Learning for Neural Whole-Body Control Across Humanoid Embodiments（arXiv 2610.17999 · 发布日期：2026-10-13）

TRACE: Efficient Transfer Learning for Neural Whole-Body Control Across Humanoid Embodiments

TRACE通过运动学先验、蒸馏与低秩强化学习，将SONIC从Unitree G1迁移到H2并在评估动作和真机演示中验证。

核心要点

• 预训练类人控制器迁移到新本体时，受运动学与动力学不匹配影响仍具挑战。
• TRACE在投影层注入运动学先验并对齐名义关节动力学，再结合编码器蒸馏与低秩强化学习。
• 在10百万动作帧上约8小时训练后，适配控制器在67,953个评估动作上成功率为96.08%。
• 真机演示覆盖动态动作跟踪和全身遥操作，采集演示还支持下游VLA策略后训练。

验证范围与局限：验证范围为SONIC从Unitree G1到Unitree H2的迁移，包含评估动作和真机动态跟踪、全身遥操作演示。

英文摘要 (Abstract)

Humanoid behavior foundation models (BFMs) enable diverse whole-body control, but adapting a pretrained controller to a new embodiment remains challenging because of kinematic and dynamic mismatch. We present TRACE, a framework for efficient cross-embodiment adaptation that preserves pretrained motion representations while specializing feedback control to the target robot.TRACE first injects kinematic priors into input and output projection layers and aligns nominal joint dynamics to reduce embodiment mismatch and support stable fine-tuning. It then combines encoder distillation with low-rank reinforcement learning on a compact, quality-screened motion set, using a two-phase curriculum to establish stable tracking before refining difficult motions. We apply TRACE to transfer SONIC from Unitree G1 to Unitree H2 using 10 million motion frames and approximately eight hours of training on a single RTX PRO 6000 GPU, at less than 0.1% of the GPU-hour cost of training SONIC from scratch. The adapted controller achieves 96.08% success on 67,953 evaluation motions under a height-scaled tracking criterion. Real-robot demonstrations span dynamic motion tracking and whole-body teleoperation, with collected demonstrations supporting downstream vision-language-action policy post-training. These results demonstrate practical reuse of a pretrained humanoid controller across robots with different body scales and dynamics.

作者：Qingyu Lan, Zhihao Tian, Zifan Wang, Junwei Liang

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2610.17999：TRACE: Efficient Transfer Learning for Neural Whole-Body Control Across Humanoid Embodiments访问 2026-10-21