Learning Continuous 3D Reconstructions for Geometrically Aware Grasping

TL;DR

This paper introduces a method that uses learned 3D reconstructions to improve robotic grasping by explicitly reasoning about object geometry, leading to more collision-aware and successful grasps.

Contribution

It presents a novel approach integrating continuous 3D reconstruction with grasp success prediction for geometrically aware robotic grasping.

Findings

Enhanced grasp success rates in experiments

Explicit geometric reasoning reduces undesired collisions

Improved grasp planning through learned 3D models

Abstract

Deep learning has enabled remarkable improvements in grasp synthesis for previously unseen objects from partial object views. However, existing approaches lack the ability to explicitly reason about the full 3D geometry of the object when selecting a grasp, relying on indirect geometric reasoning derived when learning grasp success networks. This abandons explicit geometric reasoning, such as avoiding undesired robot object collisions. We propose to utilize a novel, learned 3D reconstruction to enable geometric awareness in a grasping system. We leverage the structure of the reconstruction network to learn a grasp success classifier which serves as the objective function for a continuous grasp optimization. We additionally explicitly constrain the optimization to avoid undesired contact, directly using the reconstruction. We examine the role of geometry in grasping both in the training of grasp metrics and through 96 robot grasping trials. Our results can be found on https://sites.google.com/view/reconstruction-grasp/.