免费合规：通过双边遥操作学习可辨识阻抗

中文摘要 AI 辅助

视觉-语言-动作模型告诉机器人移动到哪里，但不告诉它该用多大的力去推。接触密集任务依赖于第二个量，即合规性，然而目前没有广泛使用的演示接口能够记录它。障碍在于可辨识性：由于实际姿态和测量到的力无法将操作者的预期平衡点与其刚度分离开来，因此VR控制器、SpaceMouse和手持式夹爪即使在原理上也无法提供合规性监督。先前的合规输出策略通过手工指定的任务结构、特权模拟接触状态或专用的力和触觉硬件来绕过这一问题。四通道双边遥操作通过使用主臂作为预期平衡点的独立测量直接消除了这种模糊性，使得仅利用机械臂上已有的关节力矩传感即可通过回归辨识出每个轴的刚度。这以零标注成本生成了每时间步、方向相关的合规标签，我们利用这些标签对VLA进行微调，使其在输出姿态的同时输出刚度。在Franka Research 3的擦拭任务中，在五个策略中，只有我们的策略在指令要求用力擦拭而非正常擦拭时，接触力发生了变化（均方根值从正常擦拭的6.4N变为用力擦拭的9.1N，Cohen's d = 0.89，p = 0.023）。

英文摘要

Vision-language-action models tell a robot where to move, but not how hard to push. Contact-rich tasks depend on that second quantity, compliance, yet no widely used demonstration interface records it. The obstacle is identifiability as realized pose and measured force cannot separate the operator's intended equilibrium from their stiffness, so VR controllers, SpaceMouse and handheld grippers cannot supply compliance supervision even in principle. Prior compliance-output policies work around this with hand-specified task structure, privileged simulation contact state, or dedicated force and tactile hardware. Four-channel bilateral teleoperation removes the ambiguity directly by using the leader arm as a separate measurement of the intended equilibrium, making per-axis stiffness identifiable by regression using only the joint-torque sensing already on the manipulator. This yields per-timestep, direction-dependent compliance labels at zero annotation cost, which we use to fine-tune a VLA to emit stiffness alongside pose. On a Franka Research 3 wiping task, ours is the only policy of five whose contact force changes when the instruction asks for a firm wipe rather than a normal one (6.4N (normal) to 9.1N (firm) RMS, Cohen's d = 0.89, p = 0.023

发表机构

• Institut de Robòtica i Informàtica Industrial, CSIC-UPC Barcelona（巴塞罗那工业机器人研究所（CSIC-UPC））

机构由 AI 辅助整理，请以论文原文为准。