Table of Contents
Fetching ...

Rotograb: Combining Biomimetic Hands with Industrial Grippers using a Rotating Thumb

Arnaud Bersier, Matteo Leonforte, Alessio Vanetta, Sarah Lia Andrea Wotke, Andrea Nappi, Yifan Zhou, Sebastiano Oliani, Alexander M. Kübler, Robert K. Katzschmann

TL;DR

Rotograb introduces a tendon-actuated hand with a rotating thumb to bridge the gap between human-like dexterity and industrial grippers. It uses 11 tendon-driven DoFs and a hollow rolling-contact finger joint with a center cutout to simplify kinematics. The rotating thumb expands the workspace and enables in-hand rotation; control combines teleoperation via depth sensing and reinforcement learning with proximal policy optimization (PPO) in simulation with domain randomization to reduce the sim-to-real gap. Experiments on YCB objects demonstrate versatile grasping across small to large items, ambidexterity in confined spaces, and effective in-hand rotation, underscoring Rotograb's potential for practical automation. Future directions include adding adduction/abduction and slimming fingers to further enhance dexterity and robustness.

Abstract

The development of robotic grippers and hands for automation aims to emulate human dexterity without sacrificing the efficiency of industrial grippers. This study introduces Rotograb, a tendon-actuated robotic hand featuring a novel rotating thumb. The aim is to combine the dexterity of human hands with the efficiency of industrial grippers. The rotating thumb enlarges the workspace and allows in-hand manipulation. A novel joint design minimizes movement interference and simplifies kinematics, using a cutout for tendon routing. We integrate teleoperation, using a depth camera for real-time tracking and autonomous manipulation powered by reinforcement learning with proximal policy optimization. Experimental evaluations demonstrate that Rotograb's rotating thumb greatly improves both operational versatility and workspace. It can handle various grasping and manipulation tasks with objects from the YCB dataset, with particularly good results when rotating objects within its grasp. Rotograb represents a notable step towards bridging the capability gap between human hands and industrial grippers. The tendon-routing and thumb-rotating mechanisms allow for a new level of control and dexterity. Integrating teleoperation and autonomous learning underscores Rotograb's adaptability and sophistication, promising substantial advancements in both robotics research and practical applications.

Rotograb: Combining Biomimetic Hands with Industrial Grippers using a Rotating Thumb

TL;DR

Rotograb introduces a tendon-actuated hand with a rotating thumb to bridge the gap between human-like dexterity and industrial grippers. It uses 11 tendon-driven DoFs and a hollow rolling-contact finger joint with a center cutout to simplify kinematics. The rotating thumb expands the workspace and enables in-hand rotation; control combines teleoperation via depth sensing and reinforcement learning with proximal policy optimization (PPO) in simulation with domain randomization to reduce the sim-to-real gap. Experiments on YCB objects demonstrate versatile grasping across small to large items, ambidexterity in confined spaces, and effective in-hand rotation, underscoring Rotograb's potential for practical automation. Future directions include adding adduction/abduction and slimming fingers to further enhance dexterity and robustness.

Abstract

The development of robotic grippers and hands for automation aims to emulate human dexterity without sacrificing the efficiency of industrial grippers. This study introduces Rotograb, a tendon-actuated robotic hand featuring a novel rotating thumb. The aim is to combine the dexterity of human hands with the efficiency of industrial grippers. The rotating thumb enlarges the workspace and allows in-hand manipulation. A novel joint design minimizes movement interference and simplifies kinematics, using a cutout for tendon routing. We integrate teleoperation, using a depth camera for real-time tracking and autonomous manipulation powered by reinforcement learning with proximal policy optimization. Experimental evaluations demonstrate that Rotograb's rotating thumb greatly improves both operational versatility and workspace. It can handle various grasping and manipulation tasks with objects from the YCB dataset, with particularly good results when rotating objects within its grasp. Rotograb represents a notable step towards bridging the capability gap between human hands and industrial grippers. The tendon-routing and thumb-rotating mechanisms allow for a new level of control and dexterity. Integrating teleoperation and autonomous learning underscores Rotograb's adaptability and sophistication, promising substantial advancements in both robotics research and practical applications.

Paper Structure

This paper contains 22 sections, 4 equations, 11 figures, 1 table.

Figures (11)

  • Figure 1: The Rotograb robotic gripper (c, d) enables precision and power grasps by integrating a rotating thumb that merges the dexterity of human hands (a) with the power grasp efficiency of soft grippers (b). Example grasps showing the dexterity of a human hand (a) and the power grasp capability of a soft fin ray gripper (b)appius2022raptor. Our new robotic gripper, Rotograb, uses a rotating thumb to combine the dexterity of robotic hands with industrial grippers to achieve precision grasps (c) and power grasps (d).
  • Figure 2: Rotograb's CAD model compared to its real-world instantiation. (a) CAD model of Rotograb with five fingers, including the rotating thumb and the actuation tower below the wrist with motors and spools. (b) Assembled 3D-printed Rotograb with silicon padding for increased contact friction.
  • Figure 3: Comparison of a normal rolling contact joint (a) and our rolling contact joint with the previously named cut-out (b). In the normal rolling contact joint, the tendon wraps around the joint while actuating the finger, extending the tendon length. In the adapted rolling contact joint, these critical non-linearities (yellow circles) are bypassed.
  • Figure 4: Uniform design and functionality of the hand's robotic finger with a base, three linked segments, rolling-contact joints, and a joint cut-out. (a) Side view of a finger, specifically the thumb. All fingers have the same design. The finger consists of a finger base and three links (lower link, upper link, and tip). The links are connected by rolling-contact joints and actuated through tendons. Mechanical stoppers limit the movement. (b) Close-up of the rolling contact joint with tendons and ligaments. The cut-out in the center of the rolling-contact joint simplifies the finger kinematics.
  • Figure 5: Detailed representation of a finger's internal flexor and extensor tendon routings, illustrating the coupled movement of two joints to demonstrate the complex mechanics of finger articulation. (a) CAD side view of the tendon routing inside the finger with three joints. Flexor tendons are drawn in red, extensor tendons in blue, while the green tendon couples the flexion of joints 2 and 3. (b) Illustration of the joint angles of a finger, with joints 2 and 3 being coupled.
  • ...and 6 more figures