DexGrasp Anything: Towards Universal Robotic Dexterous Grasping with Physics Awareness

2025-03-11

AI Summary

Addressing the challenge of generalizable dexterous grasp pose generation arising from high degrees of freedom in robotic hands and object diversity, this paper introduces the first physics-constrained, deep-coupled diffusion generative model. It explicitly incorporates contact mechanics, static friction, and stability constraints throughout both training and sampling of the diffusion process. We construct the largest-scale dexterous grasping dataset to date—comprising over 3.4 million high-quality grasp poses across 15,000+ heterogeneous objects. The method integrates multimodal object representations with a differentiable grasp quality evaluator. Our approach achieves state-of-the-art performance across all major benchmarks, significantly improving grasp robustness and cross-object generalization. To foster reproducibility and community advancement, we publicly release both the code and the dataset, establishing foundational resources for research and applications in general dexterous manipulation.

Technology Category

• Intelligent Robots: Manipulation
• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

A dexterous hand capable of grasping any object is essential for the development of general-purpose embodied intelligent robots. However, due to the high degree of freedom in dexterous hands and the vast diversity of objects, generating high-quality, usable grasping poses in a robust manner is a significant challenge. In this paper, we introduce DexGrasp Anything, a method that effectively integrates physical constraints into both the training and sampling phases of a diffusion-based generative model, achieving state-of-the-art performance across nearly all open datasets. Additionally, we present a new dexterous grasping dataset containing over 3.4 million diverse grasping poses for more than 15k different objects, demonstrating its potential to advance universal dexterous grasping. The code of our method and our dataset will be publicly released soon.

Problem

• Develop universal robotic dexterous grasping for general-purpose robots.
• Overcome challenges in generating robust grasping poses for diverse objects.
• Integrate physical constraints into a diffusion-based generative model.

Innovation

• Integrates physical constraints into diffusion model
• Creates 3.4M grasping poses dataset
• Achieves state-of-the-art grasping performance