外观
论文arXiv 2610.1048921 天前
HuMBLE框架实现人形机器人仿生步态与鲁棒转向控制统一
提出HuMBLE学习框架，通过教师-学生蒸馏与多任务RL微调，解决人形机器人仿生步态与指令跟踪冲突。在Atlas R1/D1及Unitree G1上验证，实现仅凭本体感知和躯干速度指令的实时可控仿生运动。
Boston Dynamics Atlas
宇树 G1本文由 AI 根据公开内容整理，可能有误，请以原文为准：HuMBLE: Human Motion-Driven Behavior Learning for Embodied Locomotion（arXiv 2610.10489）
核心内容
针对人形机器人运动中指令跟踪控制器步态机械、数据驱动策略泛化性差的矛盾，研究提出HuMBLE框架以合成实时可转向且鲁棒的仿生步态。该方法基于自研多速向运动数据集，利用强化学习训练全身参考条件策略，并通过教师-学生蒸馏生成仅依赖本体感知与平面躯干速度指令的轻量级先验策略，随后结合目标条件任务与参考引导风格正则化进行多任务微调，扩展了数据分布外的指令覆盖能力。
该框架已在Boston Dynamics Atlas R1、Atlas D1及Unitree G1三款人形机器人上完成实机验证。实验表明策略在室内外环境中均能响应用户直接控制并集成至分层控制栈；对比无人类数据的Tabula Rasa RL基线及消融实验证实，HuMBLE能以轻量化部署实现协调全身行为重建，在保留人类步态特征的同时保持全向可控性与环境鲁棒性，有效平衡了自然度与功能性。
关键看点
- 蒸馏架构：将全身参考条件RL策略蒸馏为仅依赖本体感知和躯干速度指令的轻量级先验策略，降低部署算力门槛
- 多任务微调：采用目标条件指令跟踪与参考引导风格正则化联合训练，突破单一数据集分布限制以增强鲁棒性
- 跨平台验证：在Atlas R1、Atlas D1及Unitree G1三型本体上实测，证实算法对不同动力学特性的泛化适配能力
- 性能对标：相比Tabula Rasa RL基线，该框架在保持人类步态协调性的同时实现了同等水平的指令跟踪精度与环境适应性
工程与产业观察
该工作为人形机器人运动控制提供了兼顾生物力学自然性与工程鲁棒性的范式参考。其「蒸馏+风格正则」技术路线解决了纯数据驱动策略难以落地的问题，证明了高质量人类运动数据经结构化处理后，可作为通用先验赋能异构本体，对具身智能底层运动层的标准化构建具有显著工程价值。
英文摘要 (Abstract)
Despite recent advances in humanoid locomotion, controllers optimized for command tracking and robustness tend to produce mechanical gaits, whereas controllers tied to human motion data often fail to generalize to commands outside the data distribution. This work introduces a learning framework that balances these competing objectives to synthesize real-time steerable, robust, and biomimetic locomotion policies from human data. Using an in-house curated locomotion dataset covering diverse speeds and directions, we first learn a natural locomotion prior policy through a teacher-student distillation process. Specifically, we train a full-body reference-conditioned policy with Reinforcement Learning (RL), then distill it into a lightweight prior policy conditioned solely on proprioception and a planar torso-velocity steering command. Next, we fine-tune the prior policy with multi-task RL to expand command coverage and robustness beyond the data distribution, pairing a goal-conditioned task that tracks arbitrary commands with a reference-guided task that tracks the human data as an explicit style regularizer. We validate our framework on three humanoid robots: the Boston Dynamics Atlas R1, Atlas D1, and Unitree G1. Experimental results demonstrate robust performance across real-world scenarios, including direct user-controlled locomotion in indoor and outdoor environments, and integration as the locomotion layer within hierarchical control stacks. Benchmarks against Tabula Rasa RL policies trained without human data and ablation studies confirm that our framework yields a lightweight, deployable policy that reconstructs coordinated whole-body behavior from a steering command, retaining the human gait characteristics while remaining robust and fully steerable.
作者：Mike Zhang, Dongho Kang, Kevin Bergamin, Nicola Burger, Robin Deits, Jonathan Foster, Bilal Hammoud, Katie Hughes, Francesco Iacobelli, Twan Koolen, M. Eva Mungai, Zach Nobles, Shane Rozen-Levy, Jean
来源
本页事实对应的公开页面，以原文为准。