Meet the most nimble-fingered robot ever built

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 robust 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 classifies grasps as robust from depth images and the position, angle, and height of the gripper above a table. 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 0.8s 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 GQ-CNN is also the highest performing method on a dataset of ten novel household objects, achieving 99% precision on test objects. (Video) Credit: Adriel Olmos, CITRIS Media

Berkeley professor Ken Goldberg, postdoctoral researcher Jeff Mahler and the Laboratory for Automation Science and Engineering (AUTOLAB) created the robot, called DexNet 2.0.

DexNet 2.0's high grasping success rate means that this technology could soon be applied in industry, with the potential to revolutionize manufacturing and the supply chain.

DexNet 2.0 gained its highly accurate dexterity through a process called deep learning. The researchers built a vast database of three-dimensional shapes — 6.7 million data points in total — that a neural network uses to learn grasps that will pick up and move objects with irregular shapes.

The neural network was then connected to a 3D sensor and a robotic arm. When an object is placed in front of DexNet 2.0, it quickly studies the shape and selects a grasp that will successfully pick up and move the object 99 percent of the time.

DexNet 2.0 is also three times faster than its previous version.

DexNet 2.0 was featured as the cover story of the latest issues of MIT Technology Review, which called DexNet 2.0 “the most nimble-fingered robot yet.”

The complete paper will be published in July.

Media Contact

Brett Israel EurekAlert!

All latest news from the category: Power and Electrical Engineering

This topic covers issues related to energy generation, conversion, transportation and consumption and how the industry is addressing the challenge of energy efficiency in general.

innovations-report provides in-depth and informative reports and articles on subjects ranging from wind energy, fuel cell technology, solar energy, geothermal energy, petroleum, gas, nuclear engineering, alternative energy and energy efficiency to fusion, hydrogen and superconductor technologies.

Back to home

Comments (0)

Write a comment

Newest articles

Lighting up the future

New multidisciplinary research from the University of St Andrews could lead to more efficient televisions, computer screens and lighting. Researchers at the Organic Semiconductor Centre in the School of Physics and…

Researchers crack sugarcane’s complex genetic code

Sweet success: Scientists created a highly accurate reference genome for one of the most important modern crops and found a rare example of how genes confer disease resistance in plants….

Evolution of the most powerful ocean current on Earth

The Antarctic Circumpolar Current plays an important part in global overturning circulation, the exchange of heat and CO2 between the ocean and atmosphere, and the stability of Antarctica’s ice sheets….

Partners & Sponsors