Robotic dexterous grasping is a challenging problem due to the high degree of freedom (DoF) and complex contacts of multi-fingered robotic hands. Existing deep reinforcement learning (DRL) based methods leverage human demonstrations to reduce sample complexity due to the high dimensional action space with dexterous grasping. However, less attention has been paid to hand-object interaction representations for high-level generalization. In this paper, we propose a novel geometric and spatial hand-object interaction representation, named DexRep, to capture dynamic object shape features and the spatial relations between hands and objects during grasping. DexRep comprises Occupancy Feature for rough shapes within sensing range by moving hands, Surface Feature for changing hand-object surface distances, and Local-Geo Feature for local geometric surface features most related to potential contacts. Based on the new representation, we propose a dexterous deep reinforcement learning method to learn a generalizable grasping policy DexRepNet. Experimental results show that our method outperforms baselines using existing representations for robotic grasping dramatically both in grasp success rate and convergence speed. It achieves a 93% grasping success rate on seen objects and higher than 80% grasping success rates on diverse objects of unseen categories in both simulation and real-world experiments.

翻译：机器人灵巧抓取因多指机械手的高自由度（DoF）与复杂接触而极具挑战性。现有基于深度强化学习（DRL）的方法利用人类示教来降低灵巧抓取中高维动作空间带来的样本复杂度，但针对高层次泛化的手物交互表征研究仍显不足。本文提出一种新颖的几何-空间手物交互表征DexRep，用于捕捉抓取过程中动态物体形状特征及手部与物体间的空间关系。DexRep包含三个组成部分：利用手部移动感知粗粒度形状的占据特征（Occupancy Feature）、表征手物表面距离动态变化的表面特征（Surface Feature），以及聚焦潜在接触点局部几何特性的局部几何特征（Local-Geo Feature）。基于该表征，我们提出一种灵巧深度强化学习方法，学习可泛化抓取策略DexRepNet。实验结果表明，在抓取成功率与收敛速度两方面，我们的方法均显著超越使用现有表征的基线方法。在仿真与真实实验中，该方法对已知物体的抓取成功率达93%，对未见类别的多样化物体亦保持超过80%的成功率。