Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metrics

TL;DR

Dex-Net 2.0 leverages a large synthetic dataset to train a deep neural network that predicts grasp success, enabling fast and reliable robotic grasp planning on novel objects with high accuracy.

Contribution

This work introduces Dex-Net 2.0, a synthetic dataset and a GQ-CNN model that significantly improves grasp planning speed and success rate using only synthetic training data.

Findings

Achieves 93% success rate on known objects in 0.8 seconds

Outperforms existing methods in speed and accuracy

Maintains high precision on novel household objects

Abstract

To reduce data collection time for deep learning of robust robotic grasp plans, we explore training from a synthetic dataset of 6.7 million point clouds, grasps, and analytic grasp metrics generated from thousands of 3D models from Dex-Net 1.0 in randomized poses on a table. We use the resulting dataset, Dex-Net 2.0, to train a Grasp Quality Convolutional Neural Network (GQ-CNN) model that rapidly predicts the probability of success of grasps from depth images, where grasps are specified as the planar position, angle, and depth of a gripper relative to an RGB-D sensor. Experiments with over 1,000 trials on an ABB YuMi comparing grasp planning methods on singulated objects suggest that a GQ-CNN trained with only synthetic data from Dex-Net 2.0 can be used to plan grasps in 0.8sec with a success rate of 93% on eight known objects with adversarial geometry and is 3x faster than registering point clouds to a precomputed dataset of objects and indexing grasps. The Dex-Net 2.0 grasp planner also has the highest success rate on a dataset of 10 novel rigid objects and achieves 99% precision (one false positive out of 69 grasps classified as robust) on a dataset of 40 novel household objects, some of which are articulated or deformable. Code, datasets, videos, and supplementary material are available at http://berkeleyautomation.github.io/dex-net .