DextrAH-G: Pixels-to-Action Dexterous Arm-Hand Grasping with Geometric Fabrics

A pivotal challenge in robotics is achieving fast, safe, and robust dexterous grasping across a diverse range of objects, an important goal within industrial applications. However, existing methods often have very limited speed, dexterity, and generality, along with limited or no hardware safety gua...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-10
Main Authors: Tyler Ga Wei Lum, Matak, Martin, Makoviychuk, Viktor, Handa, Ankur, Allshire, Arthur, Hermans, Tucker, Ratliff, Nathan D, Karl Van Wyk
Format: Article
Language:English
Subjects:
Collision avoidance
Distillation
Fabrics
Grasping (robotics)
Hand (anatomy)
Hardware
Industrial applications
Robot learning
Robotics
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A pivotal challenge in robotics is achieving fast, safe, and robust dexterous grasping across a diverse range of objects, an important goal within industrial applications. However, existing methods often have very limited speed, dexterity, and generality, along with limited or no hardware safety guarantees. In this work, we introduce DextrAH-G, a depth-based dexterous grasping policy trained entirely in simulation that combines reinforcement learning, geometric fabrics, and teacher-student distillation. We address key challenges in joint arm-hand policy learning, such as high-dimensional observation and action spaces, the sim2real gap, collision avoidance, and hardware constraints. DextrAH-G enables a 23 motor arm-hand robot to safely and continuously grasp and transport a large variety of objects at high speed using multi-modal inputs including depth images, allowing generalization across object geometry. Videos at https://sites.google.com/view/dextrah-g.
ISSN:2331-8422