The robot works in the style of a neural network designed by Ken Goldberg of the University of California, Berkeley and colleagues. Amazon and other shipping companies have been looking for a solution for some time, as the dreaded epidemic of Code 19 has multiplied the trend of home shopping on the Internet.
Earlier, the same team of scientists made some changes to the software and hardware, which increased the grip capability of the robotic arm, and now they can see better with the help of a computer and three-dimensional (three-dimensional). Look at the objects in their original place in style. This way they can lift anything better without hitting obstacles.
The reason is that robotic arms move at lightning speed to lift objects. During this time, due to fluctuations in their speed, the object either rises correctly or collides and falls down. In the same way, the robot's arm is damaged by the shock.
Experts estimate that it is necessary to estimate the speed, speed and rotation of the robotic arm for each task. Only computers and neural networks can make better decisions at this point. That's why the team of scientists decided to control the robotic arm with a neural network. They kept thousands of items for months to teach the software and put them to practice.
That's how the robotic arm finally figured out what to do to save time and lift something safely. This way the computer can now make this decision in 80 milliseconds and before that it used to take 30 seconds to make the same decision. This means that the computer makes a quick decision and immediately commands the robotic arm to pick up the object. This way the packing speed can be increased hundreds of times.
It should be noted that if this technology is adopted by a company selling goods online, its work can be very fast.
0 Comments