Internet access is increasingly being recognized as a necessity like food, water, and shelter and Google doesn’t want to leave anybody out.
Before leaving office, President Barack Obama supported the notion that the internet should be easily accesible, while Canada’s government went as far as officially declaring broadband connections “necessary to the quality of life” of its citizens.
But the infrastructure for internet service isn’t in place everywhere, making truly universal access impossible at the moment. That’s why Google has been hard at work on a project to connect people living in underserved rural and remote regions with the rest of the world online.
Project Loon, which first launched in 2013 as part of the Google X (now just X), has been sending smart balloons up into the stratosphere for four years now with the aim of creating a floating, balloon-powered LTE network. Today, Project Loon’s top brass announced a major breakthrough.
Astro Teller, X’s “captain of moonshots” (yes, that’s really his title), penned a Medium post showcasing Project Loon’s big development, which he called a “magical, serendipitous” experience. The navigational systems that control the balloons have drastically improved beyond the wildest expectations of the project’s engineers something Teller admits happened totally on accident.
At first, Loon launched its balloons and literally threw caution to the wind, controlling the devices just enough to catch the right air currents to take them on their way. After the team’s navigation algorithms improved, flightpaths could be charted more accurately to plan return loops around oceans or continents after the balloons drifted away from their original targets, but it was still too tall a task to keep them in one place.
But early last year, the Loon team noticed a “weird” tendency in some of the balloons: they lingered in one area instead of drifting away on the winds.
This changed everything. When the project launched, the team assumed the balloons wouldn’t ever be able to stay in one place.
Now, their navigation algorithms had advanced to the point where they might set up consistent, solid networks of balloons that staid put.
This changed everything.
Throughout 2016, the Loon team tested the new methods in Peruvian airspace and found their hunch was correct: some of the balloons stuck around in the same place for as long as three months. After repeating their results with multiple launches, the team concluded they had found a way to reliably keep the balloons in one spot, clustering them in groups that “danc[ed] in small loops on the stratospheric winds, over a particular region.”
The team identified three areas of development that helped the balloons’ navigation system improve so rapidly.
The altitude control system was much more efficient after updating and streamlining its solar panels and air pumps, which allowed for more maneuvers during each flight.
Since the balloons’ algorithms depend on tons of data to develop, reaching a point where simulated flights were just as good as real ones to refine those algorithms was a big step forward, too.
Finally, using a machine learning system to pilot the balloons was key for traversing the unpredictable conditions of the stratosphere.
Relying on real-time maneuvers, rather than a predetermined course, made them way more likely to get to their destination without being blown away.
With the new clusters, setting up a Loon network can be done in weeks, rather than months.
More efficient control systems mean each balloon lasts longer, and less are needed to build a network. That drops the cost, which could make Loon much more attractive to business partners around the globe.
Next step: actually testing a network with real-life users in trials, rather than just Loon team members. Google says that will come “very soon,” making for a truly world wide web.