发表机构
University of California, Berkeley（加州大学伯克利分校）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出形态度量模仿框架，通过形态优化、残差强化学习和策略蒸馏，将人手交互示范转化为零样本仿真到现实的视觉运动策略，显著提升接触精度和真实世界成功率。
AI 中文摘要
人手-物体交互（HOIs）为灵巧操作提供了丰富的示范来源，但直接从这些交互中学习面临着弥合形态差异、确保动力学可行性以及仿真到现实部署的挑战。我们提出了形态度量模仿（Morphometric Imitation），这是一个三阶段框架，将重建的人手-物体交互转化为零样本仿真到现实的视觉运动策略。首先，形态度量优化（MMO）在保留示范接触的同时，跨手部形态进行运动学重定向。其次，残差强化学习（RL）利用来自人类运动的物体姿态和接触信息来细化运动学参考，以生成动力学上可行的机器人示范。最后，这些示范被提炼为视觉运动策略。在三个机器人手和十个人手-物体交互中，MMO在五个基线中最强的基线上，每只手的接触F1分数至少提高8个百分点，同时下游动态重定向的成功率也提高了多达35个百分点。对残差RL的消融研究表明，使用物体姿态和接触信息具有互补的益处。最后，视觉运动策略在30个物体的300次真实世界试验中实现了89.3%的零样本成功率。项目页面：\href{此https URL}{此https URL}
英文摘要
Human hand-object interactions (HOIs) provide a rich source of demonstrations for dexterous manipulation, but learning directly from them presents challenges in bridging morphology gaps, ensuring dynamical feasibility, and sim-to-real deployment. We present Morphometric Imitation, a three-stage framework that transforms reconstructed HOIs into zero-shot sim-to-real visuomotor policies. First, morphometric optimization (MMO) kinematically retargets human motion across hand morphologies while preserving demonstrated contacts. Second, residual reinforcement learning (RL) refines the kinematic reference using object pose and contact information from the human motion to produce dynamically feasible robot demonstrations. Third, these demonstrations are distilled into visuomotor policies. Across three robot hands and ten HOIs, MMO outperforms five baselines in contact F1, improving on the strongest ones by 8 to 28 points, while improving the success rate of downstream dynamic retargeting by as much as 35 points. On a Sharpa hand, the visuomotor policies achieve 89.3% zero-shot success in 300 real-world trials on 30 objects spanning 10 categories. Project page: https://morphometricimitation.github.io
Comments24 pages, 10 figures
