Design of Intelligent Control System of Manipulator Based on Deep Learning

Abstract

With the continuous development of artificial intelligence technology, the application of traditional manipulator becomes more and more intelligent. One of the key intelligent improvements is to enable the manipulator to obtain the ability to quickly sort a certain target object according to a certain physical attribute in an unstructured space environment. A robot arm intelligent control training method based on deep reinforcement learning is adopted. The control model obtained from the training is combined with the visual recognition algorithm to realize the grasping and sorting of objects in the space environment by the real robot arm. The feasibility and effectiveness of applying reinforcement learning to real manipulator control are verified.