At Google's annual I/O developer conference this week, the company unveiled updates to its Android operating system, talked up new ways to turn smartphones into virtual reality headsets, and rebooted its attempt to break into digital payments with Android Pay. It didn't spend much time, however, discussing one of its most ambitious initiatives: Project Loon, Google's effort to beam broadband Internet access down to remote or rural regions of the globe from a network of stratosphere-roaming balloons.
Loon was started inside the company's Google X lab in 2011 and has accompanied other high-flying efforts, such as one to fly solar-powered drones transmitting wireless Internet signals. One drone crashed on May 1. (Facebook, as part of its Internet.org initiative, also has drones of its own.) At the time of Google's early Loon trials in 2013, balloons stayed in the air for about five days, and Google could keep only a few up there at a time. It also took more than a dozen employees to launch each balloon—not a system that could grow efficiently around the world in the way envisioned by Chief Executive Officer Larry Page and his co-founder and Google X chief, Sergey Brin.
But Loon has come a long way in the past two years, and Mike Cassidy, a vice president at Google and the project's leader, gave Bloomberg an update. He highlights two recent advancements that could help Project Loon finally reach commercial deployment as soon as next year.
First, Cassidy says Google has partially automated the balloon launching process with 50-foot-tall, cube-shaped units it calls the Autolauncher. It's also referred to, internally, as the Bird House, because various iterations of the balloons have been named after birds. (Old models were dubbed the Falcon and Grackle; the current one is called NightHawk.) The metal and canvas contraptions block the wind, clamp the balloons into place, and provide a perch for the antenna payload, which prevents it from swinging as the balloons take off. By using the Autolauncher, Cassidy says balloons can be launched with four people every 15 minutes and in winds of up to 15 miles per hour. Before, they could do one balloon every 45 minutes and were limited to wind speeds of 6 miles per hour or less. “It lets us launch a lot more balloons a lot more reliably,” Cassidy says. That's important, because Google will need hundreds in the air at one time to blanket a region with reliable Internet access.
The second advancement within Project Loon is an even bigger deal. Until now, each balloon had to link directly with a telephone company's ground stations up to 80 kilometers (50 miles) away to relay Internet signals. That suggested Google or its telco partners, such as Vodafone and Telefonica, might have to sprinkle hundreds of ground stations across continents, an expensive and inefficient proposition. Now Cassidy says they have devised a way to pass high-frequency Internet signals from balloon to balloon in midair, which allows individual balloons to roam 400 kilometers to 800 kilometers away from a ground station. One of the biggest technical challenges, he says, was finding reliable ways to point the wind-blown balloons toward each other at the right times and at a high altitude, where temperatures can reach minus-50 degrees Celsius (-58 Fahrenheit) causing such conventional components as metal gimbals and motors to malfunction. “We had to learn how to be very accurate with the pointing,” Cassidy says. By linking balloons together in what's called a mesh network, Google can now provide coverage for an entire region like West Africa with only about eight ground stations.
By the end of the year, Cassidy hopes to be able to provide a few days of continuous service in its tests. So far during trials in Australia, Chile, New Zealand, Brazil, and other countries, Google has succeeded only in providing intermittent access before the wind carries a balloon off. If it can overcome the remaining challenges, Cassidy is hoping to roll out the service more widely by the end of 2016 and is looking at underserved Internet markets such as Africa, Latin America, and Southeast Asia as the best places to start.