DexDiffuser: Generating Dexterous Grasps with Diffusion Models

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Weng, Zehang, Lu, Haofei, Kragic, Danica, Lundell, Jens
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912106500063232
author Weng, Zehang
Lu, Haofei
Kragic, Danica
Lundell, Jens
author_facet Weng, Zehang
Lu, Haofei
Kragic, Danica
Lundell, Jens
contents We introduce DexDiffuser, a novel dexterous grasping method that generates, evaluates, and refines grasps on partial object point clouds. DexDiffuser includes the conditional diffusion-based grasp sampler DexSampler and the dexterous grasp evaluator DexEvaluator. DexSampler generates high-quality grasps conditioned on object point clouds by iterative denoising of randomly sampled grasps. We also introduce two grasp refinement strategies: Evaluator-Guided Diffusion (EGD) and Evaluator-based Sampling Refinement (ESR). The experiment results demonstrate that DexDiffuser consistently outperforms the state-of-the-art multi-finger grasp generation method FFHNet with an, on average, 9.12% and 19.44% higher grasp success rate in simulation and real robot experiments, respectively. Supplementary materials are available at https://yulihn.github.io/DexDiffuser_page/
format Preprint
id arxiv_https___arxiv_org_abs_2402_02989
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DexDiffuser: Generating Dexterous Grasps with Diffusion Models
Weng, Zehang
Lu, Haofei
Kragic, Danica
Lundell, Jens
Robotics
Machine Learning
We introduce DexDiffuser, a novel dexterous grasping method that generates, evaluates, and refines grasps on partial object point clouds. DexDiffuser includes the conditional diffusion-based grasp sampler DexSampler and the dexterous grasp evaluator DexEvaluator. DexSampler generates high-quality grasps conditioned on object point clouds by iterative denoising of randomly sampled grasps. We also introduce two grasp refinement strategies: Evaluator-Guided Diffusion (EGD) and Evaluator-based Sampling Refinement (ESR). The experiment results demonstrate that DexDiffuser consistently outperforms the state-of-the-art multi-finger grasp generation method FFHNet with an, on average, 9.12% and 19.44% higher grasp success rate in simulation and real robot experiments, respectively. Supplementary materials are available at https://yulihn.github.io/DexDiffuser_page/
title DexDiffuser: Generating Dexterous Grasps with Diffusion Models
topic Robotics
Machine Learning
url https://arxiv.org/abs/2402.02989