Kinematic self-grasping learning method and system based on simulated industrial robot

The invention discloses a kinematic self-grasping learning method and system based on a simulated industrial robot, and belongs to the field of computer-aided manufacturing. In the method, robot grasping training is conducted based on a simulation environment with a reinforcement learning theory, an...

Full description

Saved in:
Bibliographic Details
Main Authors WANG TIANZHENG, YANG JIANZHONG, WU JUNXIONG, HUANG SI, XIANG DANQI
Format Patent
LanguageChinese
English
Published 07.08.2020
Subjects
CHAMBERS PROVIDED WITH MANIPULATION DEVICES
HAND TOOLS
MANIPULATORS
PERFORMING OPERATIONS
PORTABLE POWER-DRIVEN TOOLS
TRANSPORTING
Online AccessGet full text

Cover

More Information
Summary:The invention discloses a kinematic self-grasping learning method and system based on a simulated industrial robot, and belongs to the field of computer-aided manufacturing. In the method, robot grasping training is conducted based on a simulation environment with a reinforcement learning theory, and the simulation robot automatically acquires the position information of an object through an imagetaken by a camera to determine the grasping position of a robot end grasping tool; and meanwhile, the posture of the grasping tool is determined according to the shape and placement state of the to-be-grasped object in the observed image with an image processing method based on reinforcement learning, and finally objects which have different shapes and are randomly placed are successfully grasped. The grasping technology can be applied to many industry and life scenes and can reduce the complexity of grasping work programming of a traditional robot and improve the expansibility of a robot program, thereby greatly enl
Bibliography:Application Number: CN202010354236