DextrAH-RGB: Visuomotor Policies to Grasp Anything with Dexterous Hands

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Singh, Ritvik, Allshire, Arthur, Handa, Ankur, Ratliff, Nathan, Van Wyk, Karl
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910807976050688
author Singh, Ritvik
Allshire, Arthur
Handa, Ankur
Ratliff, Nathan
Van Wyk, Karl
author_facet Singh, Ritvik
Allshire, Arthur
Handa, Ankur
Ratliff, Nathan
Van Wyk, Karl
contents One of the most important, yet challenging, skills for a dexterous robot is grasping a diverse range of objects. Much of the prior work has been limited by speed, generality, or reliance on depth maps and object poses. In this paper, we introduce DextrAH-RGB, a system that can perform dexterous arm-hand grasping end-to-end from RGB image input. We train a privileged fabric-guided policy (FGP) in simulation through reinforcement learning that acts on a geometric fabric controller to dexterously grasp a wide variety of objects. We then distill this privileged FGP into a RGB-based FGP strictly in simulation using photorealistic tiled rendering. To our knowledge, this is the first work that is able to demonstrate robust sim2real transfer of an end2end RGB-based policy for complex, dynamic, contact-rich tasks such as dexterous grasping. DextrAH-RGB is competitive with depth-based dexterous grasping policies, and generalizes to novel objects with unseen geometry, texture, and lighting conditions in the real world. Videos of our system grasping a diverse range of unseen objects are available at \url{https://dextrah-rgb.github.io/}.
format Preprint
id arxiv_https___arxiv_org_abs_2412_01791
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DextrAH-RGB: Visuomotor Policies to Grasp Anything with Dexterous Hands
Singh, Ritvik
Allshire, Arthur
Handa, Ankur
Ratliff, Nathan
Van Wyk, Karl
Robotics
One of the most important, yet challenging, skills for a dexterous robot is grasping a diverse range of objects. Much of the prior work has been limited by speed, generality, or reliance on depth maps and object poses. In this paper, we introduce DextrAH-RGB, a system that can perform dexterous arm-hand grasping end-to-end from RGB image input. We train a privileged fabric-guided policy (FGP) in simulation through reinforcement learning that acts on a geometric fabric controller to dexterously grasp a wide variety of objects. We then distill this privileged FGP into a RGB-based FGP strictly in simulation using photorealistic tiled rendering. To our knowledge, this is the first work that is able to demonstrate robust sim2real transfer of an end2end RGB-based policy for complex, dynamic, contact-rich tasks such as dexterous grasping. DextrAH-RGB is competitive with depth-based dexterous grasping policies, and generalizes to novel objects with unseen geometry, texture, and lighting conditions in the real world. Videos of our system grasping a diverse range of unseen objects are available at \url{https://dextrah-rgb.github.io/}.
title DextrAH-RGB: Visuomotor Policies to Grasp Anything with Dexterous Hands
topic Robotics
url https://arxiv.org/abs/2412.01791