ADAPT-Teleop: robotic hand with human matched embodiment enables dexterous teleoperated manipulation

摘要

Anthropomorphic robotic hands are essential for robots to learn from humans and operate in human environments. While most designs loosely mimic human hand kinematics and structure, achieving the dexterity and emergent behaviors present in human hands, anthropomorphic design must extend to also match passive compliant properties while simultaneously strictly having kinematic matching. We present ADAPT-Teleop, a system combining a robotic hand with human-matched kinematics, skin, and passive dynamics, along with a robotic arm for intuitive teleoperation. The hand is evaluated based on 2D/3D kinematic similarity, skin deformation, and compliance. Compared to state-of-the-art designs, ADAPT Hand 2 integrates the most comprehensive anthropomorphic features. To demonstrate the impact of this design, we implement a direct joint-to-joint teleoperation system, enabling dexterous and contact-rich tasks such as in-hand cube manipulation, multi-object grasping, and complex pick-and-place operations. The ADAPT-Teleop platform and the developed assessment metrics serve to advance the capabilities of anthropomorphic dexterous manipulators.