发表机构
University of Tehran（德黑兰大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出结合关键点表征与DQN的强化学习框架，通过迭代优化抓取位姿，在Dex-Net数据集300个物体上对几何判定不可抓取物体实现100%成功率，且可实现仿真到真实的迁移，为接触丰富的操作任务提供解决方案。
AI 中文摘要
开发具备物体理解与操作能力的机器人，需要紧凑、可解释且泛化性强的表征。本研究提出一种基于强化学习的机器人抓取优化框架，将基于关键点的物体表征与深度Q网络（DQN）相结合。利用仿真环境中采集的二维顶视图像，基于几何的算法生成初始抓取候选，所提框架对这些候选进行迭代优化，将失败的抓取转化为成功的抓取。使用UR5机械臂在Dex-Net数据集的300个物体上开展的实验验证了该框架的有效性，其对几何方法判定为不可抓取的物体实现了100%的成功率。通过在Delta并联机器人上开展的物理实验，进一步验证了该框架的仿真到真实（sim-to-real）迁移能力，优化后的抓取成功操作了之前无法抓取的物体。研究结果强调了强化学习在解决机器人抓取挑战中的有效性，为接触丰富的操作任务提供了可扩展且适应性强的解决方案。
英文摘要
Developing robots capable of understanding and manipulating objects requires compact, interpretable, and generalizable representations. This work proposes a reinforcement learning-based framework for robotic grasp refinement, integrating keypoint-based object representations with a Deep Q-Network (DQN). Using 2D overhead images captured in a simulated environment, a geometric-based algorithm generates initial grasp candidates, which are iteratively refined by the proposed framework, transforming failed grasps into successful ones. Experiments conducted on 300 objects from the Dex-Net dataset using a UR5 manipulator demonstrate the framework's effectiveness, achieving a 100% success rate on objects previously deemed ungraspable by geometrical methods. The framework's sim-to-real transferability is further validated through physical experiments on a Delta parallel robot, where a refined grasp successfully manipulates an object that was previously ungraspable. The findings underscore the effectiveness of reinforcement learning in addressing challenges in robotic grasping, offering a scalable and adaptable solution for contact-rich manipulation tasks.
