Researchers at the University of California, Berkeley (UC Berkeley) and Google Brain have developed a robotic system that can figure out at least the first principles of tool use, through a combination of experimentation and observation.
The system is made up of an off-the-shelf robot arm that can be controlled by a person or a computer, a camera that sees the environment within reach of the arm, and a neural network that lets the robot learn.
The robot was able to learn how to use simple tools, including a dustpan, broom, and duster, to move other objects around.
The project demonstrates how robots might someday be able to learn to perform sophisticated manipulations, and to solve abstract problems for themselves.
From Technology Review
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA