[论文] PolicyIK: High-Precision Inverse Kinematics for Robot Arms with Reinfo...

论文概要

研究领域: ML 作者: Yuhan Liu, Xinqi Liu, Zixuan Dong, Keith W. Ross, Shenji Wan 发布时间: 2026-11-03 arXiv: 2611.05372

中文摘要

逆运动学（IK）把末端执行器位姿指令转换为关节运动，许多视觉-语言-动作（VLA）策略依赖它来作用于特定机械臂并跨形态迁移。精确 IK 仍然费力：冗余与奇异点使逆映射复杂，执行依赖手工调参的控制器，而物理机械臂上的训练数据代价高昂。我们提出 PolicyIK——一个目标条件强化学习策略，用于高精度 IK。它根据位姿反馈更新关节位置参考，并观测其累积参考与实测关节的相对量，从而能在静止时保留位置控制器所需的偏移。两尺度位姿误差配合随误差变化的动作增益，在保留大幅度运动的同时消除毫米级残差。进展奖励与停留奖励教会它既抵达容差又守在容差之内。仿真中，训练后的 PPO 策略单次尝试即在 93.5% 的留出任务上把最后一秒的平均误差控制在 1mm 和 0.2° 内；部署协议把成功率提升至 98.9%。直接在物理机械臂上以离线 RL 训练时，PolicyIK 在相同时间预算下的到达与保持成功率超过全部四种经典 IK 控制器。在仿真 pick-and-place 中，PolicyIK 以 95.7% 的成功率执行 π_0.5 VLA 策略的末端执行器指令。这些结果表明，强化学习可以为机械臂提供高精度逆运动学。代码已开源。

原文摘要

Inverse kinematics (IK) converts end-effector pose commands into joint motion, and many vision-language-action (VLA) policies rely on it to act on a given arm and to transfer across embodiments. Precise IK remains laborious: redundancy and singularities complicate the inverse mapping, execution depends on hand-tuned controllers, and training data on physical robots are costly to collect. We present PolicyIK, a goal-conditioned reinforcement learning policy for high-precision IK. It updates joint-position references from pose feedback and observes its accumulated reference relative to the measured joints, so it can retain the offset that the position controller needs at rest. Two-scale pose errors with an error-dependent action gain resolve millimeter residuals while preserving large motions. Progress and dwell rewards teach it to both reach the tolerance and stay within it. In simulation, a single attempt by the trained PPO policy keeps the mean errors over the final second within 1 mm and 0.2° on 93.5% of held-out tasks, and a deployment protocol raises success to 98.9%. When trained directly on a physical arm with an off-policy RL learner, PolicyIK achieves higher reaching and holding success than all four classical IK controllers under the same time budgets. In simulated pick-and-place, PolicyIK executes the end-effector commands of the \(π_{0.5}\) VLA policy with 95.7% success. These results show that reinforcement learning can provide high-precision inverse kinematics for robot arms. Code is available at https://github.com/Tinaliu0123/PolicyIK.

---

*自动采集于 2026-11-05*

#论文 #arXiv #AI #小凯