Microsoft's AI controlled sailplane soars like a bird but without motors

Building on its autonomous flight ambitions, Microsoft has tested a motor-less glider that borrows a simple technique from birds and captures invisible columns of rising hot air, called thermals, to soar through the sky, says a new report from the New York Times.

Tested in the Nevada desert, the new Microsoft sailplane boasts an AI controlled self-guidance system that predicts the movement of naturally occurring thermals and lets the machine ride on them, in a way similar to how falcons, hawks and other birds glide in the sky.

In a blog detailing the glide of the 16.5ft, 12.5- pound sailplane, Microsoft's principal researcher Ashish Kapoor said, "Birds do this seamlessly, and all they're doing is harnessing nature. And they do it with a peanut-sized brain". A machine, however, requires a complex set of AI algorithms to replicate the same.

Unlike most AI-controlled systems operating in the real world, the AI algorithms used in Microsoft's soaring machine perform multiple tasks and take action in real-time. Practically, it is impossible to predict weather and thermal pockets with 100% accuracy, but with advanced AI capabilities, the sailplane assesses things like air temperature, wind direction, and places where it's not supposed to fly. Then, it uses that data to make educated guesses on where to jump next (finds next upward hot air current), without any change in altitude.

It is probably one of the few AI systems that not only makes predictions but also takes action based on them, Kapoor notes. Though, for this specific test, they went with a backup option and had a motor attached to go manual if necessary.

Still, Microsoft's sailplane, or the system of riding winds, shows a lot of promise towards autonomous flights. As a future upgrade for this tech, the Redmond giant aims to deploy solar or wind energy to keep the glider up and running for as long as they want. This way, it could fly indefinitely and prove itself useful in a bunch of applications such as serving under-served areas with internet or monitoring ground and weather conditions in real time.