The company stated that a set of algorithms written and executed by a deep reinforcement learning-based flight control system is more efficient and adept than the human model. The AI-powered system is currently managing balloons over Kenya where Loon launched its first commercial internet service in July 2020 following fleet tests in a series of disaster relief initiatives.

Reinforcement learning is a technique that allows software to teach itself skills through trial and error, like teaching computers how to play video games and assisting software to manipulate robotic limbs. However, such repetition is not possible in the real world while dealing with high-altitude balloons that are costly to operate and costlier to repair if damaged due to a crash.

Loon has taught its flight control system to pilot the balloons using computer simulation, with help from Google’s AI team out of Montreal. That way, the system could improve over time before being deployed on a real-world balloon fleet.

“While the promise of RL (reinforcement learning) for Loon was always large, when we first began exploring this technology it was not always clear that deep RL was practical or viable for high altitude platforms drifting through the stratosphere autonomously for long durations,” explains Sal Candido, Loon’s chief technology officer and co-author of a paper on the new flight control system published this week in the scientific journal Nature, in a blog post. “It turns out that RL is practical for a fleet of stratospheric balloons. These days, Loon’s navigation system’s most complex task is solved by an algorithm that is learned by a computer experimenting with balloon navigation in simulation.”

Loon says its system qualifies as the world’s first deployment of this variety of AI in a commercial aerospace system. “To be frank, we wanted to confirm that by using RL a machine could build a navigation system equal to what we ourselves had built,” Candido writes. “The learned deep neural network that specifies the flight controls is wrapped with an appropriate safety assurance layer to ensure the agent is always driving safely. Across our simulation benchmark we were able to not only replicate but dramatically improve our navigation system by utilizing RL.”

The AI-controlled system handily outperformed the human one by consistently staying closer to a device the team uses to measure LTE signals in the field, and that test paved the way for more experiments to prove the efficacy of the system before it formally replaced the one the team had spent years building by hand. Loon now thinks its system can “serve as a proof point that RL can be useful to control complicated, real world systems for fundamentally continual and dynamic activity.”