Loon balloons are based on artificial intelligence from Google |
Loon Balloons, a subsidiary of Alphabet, is responsible for moving the Internet from helium balloons to Earth's stratosphere. Since the navigation system is no longer managed by an artificial program, the balloon has been directed to all parts of the world thanks to artificial intelligence which becomes a new stage.
The company is now using an improved, deep learning-based flight control system that is more efficient and effective than the old manual system.
The system is now operating a fleet of balloons in Kenya. For most of the past decade, Loon launched its first fleet in July after testing its fleet through a number of disaster relief initiatives and other testing environments. Commercial Internet services.
Like the startling progress of artificial intelligence researchers in teaching computers to run advanced video games, enhanced learning is a technology that allows programs to learn skills through trial and error.
Obviously, when it comes to high-rise balloons, this duplication is not possible in the real world. These balloons are expensive to operate and even more expensive to repair if they fail.
With the help of Google's AI team, Loon, like many other AI labs that have turned to reinforcement learning to develop advanced AI programs, has introduced its in-flight control system into how to use computer simulations to ride balloons.
Loon said: His system is the world's first artificial intelligence system for commercial flight systems to surpass human-designed systems.
During the first real system test in Peru in July 2019, the AI-controlled flight system encountered a conventional system that is controlled by an artificial algorithm called StationSeeker.
Loon now believes his system can prove deep learning useful for controlling complex real systems (mainly continuous and dynamic activities).