The Google-Team, can robots learn to run (almost) like a real beast


Robots like Autonomous cars that take over in an unsupervised manner, certain tasks that are now well-known concept, but the advance of the learning process for this robot is quite different from that of independent, but supported by a lot of human intervention. Researchers at Google, we want to change this: they were able to develop the Algorithms in such a way that a four-legged robot learned how to go into a real-world environment, and with very little support from the outside. The statement said The Analysis Of The Technology the online “robot learning by myself and in the truth.

The work is based on a survey from about a year ago, and the Google-group and found for the first time, like a robot would in a real-world learning experiences. Reinforcement learning is usually carried out in a Simulation, a virtual DoppelgangeR for the robot to move through a virtual replica of the real-life environment to which the algorithm is robust enough to in real life. With this method, it is possible to prevent damage to the robot and the environment, but they are in need of an easy-to-simulated environment. As the sand is in place, it is a robot and put his foot on it, or how deep of a mattress this sinks in, it is not so difficult to calculate that in the simulations, to be of service to you here.

Against this backdrop, the Google researcher decided to dispense with the simulation, and to train in the real world. They were able to develop an algorithm that is more efficient, with less experience, and therefore, the less the error that it requires, and within a couple of hours, you have your together. Because of the environment that have had changes, he was able to respond quickly to similar environments, with slopes, obstacles, and stairs. With further modifications, the researchers ensured that the process of learning is self-employed and with no damage to the front end.

If you have any problems with the playback of the video, please enable JavaScript

(Source: Google)

Further, in the Technology of the Review online: