通过视觉提示引导双臂拆装，提升复杂零件非破坏性拆卸效率
Affordance-Guided Dual-Armed Disassembly Teleoperation for Mating Parts
- 基于物体几何生成可抓取位姿和拆卸方向，可视化辅助操作
- 实测任务成功率提升，零件姿态偏差减少，稳定性更好
- 适合需要精准拆装的工业场景，如精密设备维护
由于需要灵活操作且内部结构可视性差，机器人对配合件的非破坏性拆卸仍具挑战。本文提出一种基于可及性提示的遥操作系统，支持双臂固定-拆卸任务的人类直观示范。系统在虚拟环境中可视化由物体几何决定的可行抓取位姿与拆卸方向，以应对遮挡与结构复杂性问题。为避免负载下位置跟踪过度，引入融合位置控制与阻抗控制的混合控制器。真实实验验证了该系统的有效性，任务成功率提高，对象姿态偏差显著降低。
原文摘要 · Abstract (English)
Robotic non-destructive disassembly of mating parts remains challenging due to the need for flexible manipulation and the limited visibility of internal structures. This study presents an affordance-guided teleoperation system that enables intuitive human demonstrations for dual-arm fix-and-disassemble tasks for mating parts. The system visualizes feasible grasp poses and disassembly directions in a virtual environment, both derived from the object's geometry, to address occlusions and structural complexity. To prevent excessive position tracking under load when following the affordance, we integrate a hybrid controller that combines position and impedance control into the teleoperated disassembly arm. Real-world experiments validate the effectiveness of the proposed system, showing improved task success rates and reduced object pose deviation.
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。