arxiv.org·2026/10/26
为什么重要 为腿式机械臂部署VLA提供数据采集、模型适配与推理设计经验，演示覆盖和次级技能策略可指导真实机器人落地。
视觉-语言-动作（VLA）模型通常在广泛多样的数据集上进行预训练，随后在针对特定目标机器人的相对较小的数据集上进行微调。这种适配方法如今已被广泛确立，在固定基座机械臂上取得了最成熟的成果，因为其基座动力学是静态的，参考系也保持一致。将同样的机械臂安装到四足机器人上，会把工作空间扩展到整个地面，但会从根本上改变其实体形态：策略不能再孤立地控制机械臂，而必须联合协调基座导航、身体姿态和末端执行器运动。
在这项工作中，我们研究了这种实体形态变化在数据采集、模型适配和推理设计方面的实际需求，以在 Unitree Go2 与 AgileX Piper 机械臂上实现闭环移动操作。我们报告了使这一真实机器人部署可行的选择，包括模型变体、动作时域和演示惯例，并开展了一项针对抓取-放置任务的研究，将数据覆盖范围与模型容量分离开来。
我们表明，扩大演示覆盖范围可将留出末端执行器误差降低 30%，并且纳入次级技能（例如推动而非抓取）会引入新的行为模式，同时不会降低基线性能。最后，闭环真实世界实验展示了数据集构成如何塑造成功与失败，验证了该流程可用于更复杂的移动操作。
Vision-language-action (VLA) models are typically pretrained on broad, diverse datasets and subsequently fine-tuned on a comparatively small dataset tailored to a specific target robot.
Such adaptation recipe is now widely established, achieving its most mature results on fixed-base arms due to their static base dynamics and consistent reference frame.
Mounting the same arm on a quadruped extends the workspace to the entire floor, but fundamentally changes the embodiment: rather than controlling the arm in isolation, the policy must jointly coordinate base navigation, body posture, and end-effector motion.
In this work, we study the practical demands of this embodiment change across data collection, model adaptation, and inference design to achieve closed-loop loco-manipulation on a Unitree Go2 with an AgileX Piper arm.
We report the choices that made this real-robot deployment feasible, namely model variant, action horizon, and demonstration conventions, together with a study tailored to a pick-and-place task, separating data coverage from model capacity.
We show that expanding demonstration coverage reduces held-out end-effector error by 30%, and incorporating secondary skills (such as pushing rather than grasping) introduces new behavioral modes without degrading baseline performance.
Finally, closed-loop real-world experiments show how dataset composition shapes success and failure, validating the pipeline for more complex loco-manipulation.
原版排版与图表 · 手机无法内嵌时可在浏览器打开。PDF 插件内选文不能传给网页助手。