Control and Grasping of an Anthropomorphic Robotic Hand Built by Bio-inspired Robotic Fingers

Abstract

Dexterous manipulation and grasping pose challenges for robots. Drawing inspiration from the human hand, this paper introduces a bio-inspired design for a robot finger and a robot hand, with the aim of creating an anthropomorphic robotic hand prototype. The new finger design seamlessly integrates rigid elements and soft materials. In this paper, we present a complete solution to model such bio-inspired robotic fingers leading to a real-time model-based pose controller. All five fingers of the robot hand are mounted on a carefully designed palm, with additional degrees of freedom (DoFs) for the index finger and the thumb finger to enhance dexterity. The robotic finger responds fast to control, resulting in a robotic hand capable of tracking human hand motion in real time, using a hand tracker for the hand pose. The robotic hand can be controlled to achieve most human grasp types. Several manipulation and grasping tasks are demonstrated using teleoperation.

Chao Liu

Postdoctoral Associate

My research interests include manipulation and tactile sensing, swarm and modular robotics, soft robotics, parallel robotics, control and motion planning.