Grasp Diffusion Network: Learning Grasp Generators from Partial Point Clouds with Diffusion Models in SO(3)xR3

TL;DR

This paper introduces a diffusion-based generative model for robotic grasping from partial point clouds, effectively sampling collision-free grasp poses in SO(3)×R3, achieving high success rates in simulation and real-world tests.

Contribution

It proposes a novel diffusion model on the rotation manifold with collision avoidance guidance for grasp generation from partial point clouds.

Findings

Achieves 90% grasp success rate in real-world experiments.

Outperforms several baseline methods in benchmark tests.

Enables faster grasp sampling using advanced diffusion techniques.

Abstract

Grasping objects successfully from a single-view camera is crucial in many robot manipulation tasks. An approach to solve this problem is to leverage simulation to create large datasets of pairs of objects and grasp poses, and then learn a conditional generative model that can be prompted quickly during deployment. However, the grasp pose data is highly multimodal since there are several ways to grasp an object. Hence, in this work, we learn a grasp generative model with diffusion models to sample candidate grasp poses given a partial point cloud of an object. A novel aspect of our method is to consider diffusion in the manifold space of rotations and to propose a collision-avoidance cost guidance to improve the grasp success rate during inference. To accelerate grasp sampling we use recent techniques from the diffusion literature to achieve faster inference times. We show in simulation and real-world experiments that our approach can grasp several objects from raw depth images with $90\%$ success rate and benchmark it against several baselines.