Future Tense

The Air Traffic Control System of Our Low-Altitude Future

Something will have to stop all the drones, flying cars, and floating warehouses (floating warehouses!) from colliding.

A drone filled sky (animation).
Lisa Larson-Walker

Amazon CEO Jeff Bezos has been convinced for a while that one day flying robots will deliver packages to your home or office. No more driving to the store to pick up toothpaste; no more receiving human couriers toting important documents; no more mail carrier, no more mover, no more pizza guy. When Bezos described his vision of drone-filled skies in a 2013 60 Minutes interview, it may have been easy to scoff. “It drops the package. You come and get your package and we can do half-hour deliveries,” Bezos said, imagining that drone delivery could happen within five years. The biggest hurdle, he said, would be regulation.

Well, it’s five years later, and while drones haven’t taken over your mail delivery yet, the Department of Transportation says there are more than 1 million drones registered to fly in U.S. airspace (which is certainly a lowball figure, since one registration number can count for multiple drones from the same owner, and there are many drone owners who don’t register at all). The Federal Aviation Administration predicts that by 2021, the number of drones registered in its database could surge to more than 6 million. Alas, only about 70 of those currently listed in the drone registration database belong to Amazon.

But there’s a sense in which Bezos was right—and remains prescient. Drone delivery is coming, as could be some manner of flying cars or floating warehouses. But before that happens, authorities will have to figure out an entirely new system of low-altitude air traffic control, one that could be orders of magnitude more complicated than the system governing how airplanes travel. How do you control the skies that are mere inches above our heads?

For starters, the Federal Aviation Administration still hasn’t ironed out the rules that would allow drones to fly beyond the line of sight of the operator, which it needs to figure out before drone delivery can really take off in the United States. The whole point of drone delivery, after all, is that there’s no human doing the work; a person standing there watching the robot fly defeats the purpose. That regulatory uncertainty hasn’t necessarily stymied innovation, but it has limited the commercial uses of drones up until now.

Unlike airplanes, which take off from and land at airports, drones can begin and end their journeys anywhere. That’s an air traffic control nightmare. Since drones don’t have a pilot in a cockpit who can communicate with an air traffic control center about their flight plan and location, having millions of drones in the air flying without someone watching and piloting each aircraft is begging for disaster. All of which means that before drones can be let off the leash, they’re going to have to find a way to communicate—not only with each other but also with manned aircraft, the national air traffic control system, the fire department (if, say, there’s a house fire that a drone needs to avoid), and whoever else needs to know about a flying robot’s course and destination. In many ways, air traffic control for drones is the missing linchpin in Bezos’s dream of ubiquitous drone delivery.

One of the reasons why so many drones have been able to fly without complete pandemonium despite the lack of a national flight reporting system is that the aircraft don’t fly that high. Right now, the FAA allows for flying a drone within the line of sight of the operator only if its stays below 400 feet, which is unregulated airspace. Still, though airplanes fly higher than drones, at some point planes have to take off and land. With that in mind, the FAA has been working with NASA for years to research exactly how a low-altitude air traffic control that’s made to communicate with robots can come to fruition and work with how national airspace is already managed. (You can bet that low-altitude air traffic won’t be allowed anywhere near active runways.) The plan now is to finish that research by 2019 and hopefully have a system implemented by 2025.

That timeline, though, is a government research timeline, and meanwhile, the industry, with help from the White House, is aiming to have a solution much sooner. In October, the White House released an executive order inviting states and municipal governments to apply for broad waivers that would allow companies to conduct drone delivery trials and test air traffic control systems that go beyond what the current law allows. And according to Gregory McNeal, co-founder of AirMap, a company that provides drone traffic management solutions and has been working closely with the FAA on drone policy, hundreds of local governments have expressed interest in participating. Now the FAA and members of local and federal law enforcement agencies are reviewing those proposals. The Trump administration ordered that a minimum of at least five pilot sites for testing drone delivery be approved by the end of July. Which means that this year, drone delivery is likely to take off in certain areas across the country, including major cities, long before a national drone traffic management solution is finalized. All kinds of businesses have been eager to hop on the drone bandwagon. There was the dry cleaner in Philadelphia who hung his hanger on a drone in 2013, and in 2014, a flower delivery company outside of Detroit tried to ferry a bouquet through the skies, only to have the FAA ground the drone. There’s also (thank goodness) a company working on a drone burrito delivery service.

The new drone pilot programs could take any number of shapes. If there’s one thing the industry hasn’t been shy about, it’s imagining what a future with drone delivery might look like. Amazon, for example, has proposed an idea for beehive-like towers where drones can recharge and be restocked before going out for another delivery. The company also won a patent for its design of floating warehouse fulfillment centers (like, in the sky!), as well as an idea for smaller drones to assemble and link together to carry larger packages. There’s also Project Wing, the drone delivery program from Google’s parent company Alphabet, which has tested burrito delivery by drone and has a patent for a ground robot designed to catch packages from drones to bring them to people’s front doors. Amazon has also proposed an aerial highway system, where drones that are filming or surveying land fly lower than drones jetting across the sky at high speeds delivering packages.

And beyond the futuristic ideas proposed, it’s not like Amazon has been quietly waiting for the FAA’s green light. In 2014, tired of the slow pace of FAA approvals, the company decamped to the U.K. to start testing its drone delivery program and has likewise opened testing centers for Amazon Prime Air in Israel, Austria, and France. The company has tested multiple iterations of its delivery drone and has been steadily hiring engineers for years to make its dream of drone delivery reality. Still, Prime Air hasn’t rolled as a full program officially anywhere yet, and if the U.S. does end up following through with the executive order to open the skies in some parts of the country, America’s largest online retailer might roll out its full drone delivery program somewhere in the U.S. first.

Still, though the industry is likely pleased with the expedited timeline from the Trump administration, it’s hard to say how exactly each of the new drone pilot programs will roll out. Drones can be pushed off-course due to high winds, for example, which Parimal Kopardekar, the head of NASA’s drone air traffic control testing program, told me can derail a drone a couple of hundred feet off-course. That means that a drone could drift into an area where it’s prohibited from flying, like an airport, or even into another drone. But in order to hedge for a scenario like that, receiving real-time communication from the drone about where it’s headed will be essential to clear the way from other drones and also to help navigate the aircraft away from areas it’s supposed to avoid, like over wildfires, large sporting events, or jails, for example.

That kind of information sharing will require a level of interoperable communication that’s not yet standardized across drone-makers, the FAA, or law enforcement—something that would likely be a part of a national air traffic control system that integrated drones. Which means that the communication or data exchange protocols between drones, federal, and local agencies would have to be standardized in such a way that “you won’t be talking in French and I won’t be talking in English when we share information about co-location and our intent,” says Kopardekar. Though there has been extensive testing and research to figure out how such data-exchange protocols would work, there’s still no established national standard.

And there’s really no telling how the FAA’s pilot programs will look. One drone program could just be to allow Amazon to try its drone delivery idea by itself in just one city. Another might allow drone delivery broadly for a range of companies. All the programs that get approved have to show that they’ve fully considered how their programs will prioritize public safety and security. But the whole point of these pilot programs is to learn more about how to implement drone flights safely, meaning the federal government is opening up the skies to drones to fly potentially beyond the line of sight of the operator, at night, and over people, before there’s a clear understanding how to do this right. After all, we can’t mitigate for risks until we understand the risks.

Despite the excitement in industry and government for drone delivery to take flight, many people are still wary of drones, whether it’s due to the privacy implications of having an autonomous robot with high-definition cameras buzzing around overhead, or out of a fear that the aircraft may fall out of the sky. And one major accident could set off a cascade of bad public opinion that cements resistance to the high-tech delivery concept for years into the future, potentially opening the doors to major setbacks for the industry if politicians feel their constituents don’t want the robots flying over their backyards.

All of this is to say nothing of how more complicated things could get once other aircraft enter our low-altitude skies. Uber, for example, is working on a flying car service that the company says it expects to get off the ground by 2020, an unlikely proposition considering commercial drones have been around for decades and still can’t legally fly at night or over crowded areas without someone watching. At the CES trade show in Las Vegas this year, the company Workhorse was set to debut its passenger drone, but it never took off because of the rain. And then there’s the flyer from Kitty Hawk, which you can think of like an aerial motorcycle; the company is helmed by Sebastian Thurn, who formerly pioneered Google’s self-driving car project, and counts among its investors Google co-founder Larry Page. The vehicle seats one person and isn’t supposed to require a pilot’s license to fly. If any of these things have a hope of viability, it will rely in great part on how they fit into a new, highly dynamic air traffic system.

If things do go to plan, delivery drones have the potential to completely upend the way cities work and the way we receive the things we buy. Amazon has said that around 90 percent of its orders weigh less than 5 pounds, making them eligible for drone deliveries. When it comes to food orders, flying robots, soaring above traffic, may be able to ferry lightweight packages to people’s houses way faster than a delivery driver. Flirty, a Nevada-based drone startup, partnered with 7-Eleven to make 77 drone deliveries within the operator’s visual line of sight in 2016, moving things like coffee, a chicken sandwich, donuts, candy, and Slurpees, showing that the idea isn’t as far-fetched as one might think. Blood can be delivered to clinics short distances faster by drone than by a driver, as Zipline, a California-based drone company, has demonstrated in Rwanda. Drones could mean fewer cars on the road, which in turn could mean less emissions and less traffic. It could also mean fewer jobs for delivery drivers. (UPS has also been experimenting with drone delivery.)

While it’s hard to say how the trials will play out, autonomous drones are coming to American skies this year in a much bigger way. After that, the future is up in the air.