用物理理论约束强化学习，实现稳定抓握与精准操作。
Learning Stable In-Grasp Manipulation in a Non-Dropping Action Space
- 将复杂操作拆解为可分析的子技能，结合物理规律进行训练。
- 在不同噪声、摩擦和延迟条件下，学习效率与稳定性显著提升。
- 适合对机器人抓取与控制有高精度要求的研究者参考。
传统灵巧操作控制器依赖强假设下的解析模型，而强化学习虽能端到端探索技能，却因目标冲突导致学习不稳定且效率低下。本文通过将灵巧操作技能分解为多个更简单、可分析的组件，结合经典物理与控制理论对每个组件施加约束与引导，实现了高效稳定的技能学习。实验表明，在不同物体、传感器/电机噪声、延迟及摩擦条件下，该方法在保持抓握稳定的同时，显著提升了重定位与重定向操作的准确性和学习效率，验证了先验理论知识对强化学习的有效促进作用。
原文摘要 · Abstract (English)
Traditionally, dexterous manipulation controllers are designed using analytic models constrained by strong assumptions about the hand and the objects being manipulated. Reinforcement learning (RL) has become another common approach in which skills are explored openly in an end-to-end manner but is inefficient because of unnoticeable instability and conflicts in learning objectives. This paper attempts to efficiently explore stable and accurate manipulation skills by decomposing dexterous skills into multiple simpler/analyzable components. Each skill component is subsequently learned with constraints and guidance from classical physics and control theory. Our work shows that for stable grasp, in-grasp reposition/reorientation with different objects, sensor/motor noise, latency, and frictional conditions, skill learning becomes efficient and stable with prior knowledge from theory.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。