US 11,701,771 B2
Grasp generation using a variational autoencoder
Arsalan Mousavian, Seattle, WA (US); Clemens Eppner, Seattle, WA (US); and Dieter Fox, Seattle, WA (US)
Assigned to NVIDIA CORPORATION, Santa Clara, CA (US)
Filed by NVIDIA Corporation, Santa Clara, CA (US)
Filed on Mar. 4, 2020, as Appl. No. 16/809,087.
Claims priority of provisional application 62/848,521, filed on May 15, 2019.
Prior Publication US 2020/0361083 A1, Nov. 19, 2020
Int. Cl. B25J 9/16 (2006.01); G06N 3/045 (2023.01)
CPC B25J 9/1612 (2013.01) [B25J 9/1669 (2013.01); B25J 9/1697 (2013.01); G06N 3/045 (2023.01); B25J 9/161 (2013.01)] 27 Claims
OG exemplary drawing
 
1. A computer system comprising one or more processors and computer-readable memory storing executable instructions that, as a result of being executed by the one or more processors, cause the computer system to at least:
use a first neural network to generate, from a three dimensional point cloud of an object, a set of grasp poses that allow a robot to grasp the object;
use a second neural network to determine an evaluation of individual grasps in the set of grasp poses; and
refine the individual grasps in the set of grasp poses based at least in part on a gradient of the evaluation determined by the second neural network to produce a set of refined grasp poses.