Remote Towers

As a pilot, I'm a little freaked out by the idea and way more than a little intrigued. On the one hand, tower operations haven't changed much since the days of NDB approaches. A pilot calls in location and intentions, and a controller typically reaches for a pair of binoculars to identify the aircraft. Sure (in some towers), there might be a radar screen to help out or a paper flight strip with IFR routing information (this is just in the TRACON, from my understanding of it). At its most basic, though, the interaction is visual. That's a workflow ripe for modernization.

On the other hand, you're going to take away the one person who can actually see me---who can check that my gear is down, warn me about that fox on the runway or keep an eye on my progressive taxi---and you're going to put that worker in the blue-white Matrix glow of a warehouse cubicle somewhere off the interstate?

Remotely operated control towers aren't just a vision for the future, though. They're rapidly becoming the present. To see how, let's take a trip to Scandinavia. The town of Örnsköldsvik, Sweden, population 29,000, isn't where you'd imagine revolutions begin. On the clear, blue day of April 21, 2015, pilots were flying visually. But, for the first time ever, the controllers worked by instrument as they cleared an aircraft to land from about 90 miles away at the larger Sundsvall-Timrå Regional Airport.

That scenario is exactly the point: Airports in remote locations make one of what turns out to be many great arguments for remote towers. In the case of Örnsköldsvik, it would cost about $1 million per year to hire a crew of six full-time, local controllers. In their place is a system designed by Saab (yes, that Saab). It features a camera tower with 14 high-definition cameras that capture a seamless panorama of the field. Those are synched to airport weather and operations data, all of which is relayed to the Sundsvall crew. For the 12 or so flights that land daily at Örnsköldsvik, it's an efficient solution that brings the resources of a larger ATC facility to a small airport that might otherwise go un-towered.

That's not to say there weren't technical problems to overcome. Even a tiny speck on a camera lens becomes a huge blot on TV, so the Saab system encloses the cameras in a glass housing, with high-pressure air running continuously over the exterior surface to keep the view clean and clear. And, when Saab first began testing the system, controllers were spooked by the absence of aircraft noise. Now, those sounds are piped in to provide aural clues.

Despite these learning experiences, the expense to install a remote tower is much lower. While Saab hasn't released exact cost, Norwegian competitor Kongsberg Defence Systems booked a sale to provide remote tower services to 15 airports at a total cost of $48 million, an average of just over $3 million per facility. In any case, remote towers are a small fraction of the $10 million or more that a small U.S. airport might spend on a manned tower, let alone the $120 million that a major hub like San Francisco International Airport recently spent on a new tower facility.

We might not completely shut down tower operations at a major airport yet, but that doesn't mean there aren't immediate opportunities at large facilities. Take Searidge's remote tower for HungaroControl in Budapest, which has nearly 100,000 movements per year. The system there has been testing with both live movements and in shadow mode on multiple runways. The goal is to operate as a contingency tower this year and a full remote tower by 2018.

Per Ahl, Saab's Vice President Head of Marketing and Sales, Digital Air Traffic Solutions, says it's not uncommon in Europe for a private company to own a major airport and several smaller airfields in a region. The Saab solution allows the operator to pool resources and still maintain high levels of service at all facilities.

In the U.S., Saab has an ongoing remote tower demonstration project at Leesburg, Virginia. Since 2015, the project, in partnership with the Virginia SATSLab, has tested high-definition fixed and pan-tilt-zoom cameras, signal light guns and microphones. FAA-certified control tower operators continue to staff a traditional tower, and there's no plan yet for a cut-over. If remote technologies can work in a busy regional airport inside the Washington, D.C. Special Flight Rules Area and just 10 miles from the commercial hub of Dulles International Airport, well, you have to think changes might come quickly, even in a world of FAA review and approval.

It may be that the idea of providing tower service to isolated locations helps to explain why the three leading companies competing hail from Sweden (Saab), Canada (Searidge) and Norway (Kongsberg Defence Systems). Sure, there's a cost savings, but for pilots perhaps also a promise of services where they don't currently exist.

"It's going to be hard [for remote towers] to be accepted just because an accountant thought it was a good idea," says Searidge's Sauriol. In addition to pop-up service at backcountry airstrips, he offers the idea of seasonal service. "You have vacation places based on summer- or wintertime, and you can only justify the tower and being able to provide services in a seasonal setting."

Remote service also creates the possibility to "move" controllers to areas where there's a shortage. Say you have a surplus of workers in Colorado and a shortage in California, due to the high cost of living there. The Rocky Mountain crew can take the Coast shifts. They can even potentially sit in a common facility with approach (TRACON) controllers to better coordinate handoffs.

"In every situation that I'm aware of, remote towers are an upgrade of one sort or another," says Sauriol, citing a project at Dallas/Fort Worth Airport with the Massachusetts Institute of Technology and the FAA's NextGen group. The short-term goal was getting all controllers co-located in a single tower, instead of spread to multiple facilities around the field. That's a value because it frees tower real estate for other uses, while reducing or eliminating the visual blind spots that make multiple towers necessary in the first place.

But the real, long-term idea was to find out whether they could use technology to maintain the traffic flow that controllers achieve in visual conditions when they can't see out the window due to weather.

Now, we're on to something exciting: We've ripped out the windshield, but we're replacing it with something that shows you more and better information.

"We can do better with software," Sauriol says. "A lot of that is bringing in data, augmenting what you see. So, you may click on a radar target and the video will actually zoom onto it---what we call digital binoculars."

Likewise, radar or ADS-B data-block information with the aircraft's N-number, and movement of ground and air vehicles, can be displayed next to their images. In reduced visibility, infrared might be overlaid, as well, and automatic tracking can follow objects as they move.

This is where thinking big and moving slow may in the end yield revolutionary results.

"The idea that, given a white sheet of paper and transitioning to a remote tower, you'd re-create that exact same thing just isn't right," says Sauriol. "You can do the 360-degree view, and we're doing that in a few places," he says. "But that's mostly the exception."

Instead, Searidge is working toward a completely new concept of airport control. It's not a re-creation of reality; it's a flexible platform that serves up the information needed at any given time.

Here's where the rethink of tower control goes beyond just hardware: Like so many modern, big-data solutions, the Searidge solution relies on artificial intelligence and neural networks that help the software learn and adapt.

Instead of writing rules that say, for instance, that a jet with a red tail is Air Canada, the operator just points out to the system, "This is an Air Canada plane. There's another one. And there's a third one." The application, itself, makes the connections to help it draw out the similarities. So, it quickly learns to identify planes with images degraded by bad weather. And, it can figure out what kinds of ground movement are likely to pose a threat to an incoming plane.

Sauriol imagines a future in which the remote tower software might predict that a landing aircraft has a high probability of needing to go around and then help the controller to automatically reposition other traffic in anticipation of that.

It all begins with getting the software in the field and allowing controllers to begin adapting it to their needs. "That's the real trick," Sauriol says. "The traditional cycle is that a tower tells the FAA their operational conditions, and the FAA says, ’Okay, we're going to build you a solution.' The cycle time for that is many years.

"What we want to say is, ’What if we were to give you the tools to build your own system, because it's artificial-intelligence based?' Local operators would suddenly build their own applications and solve problems in months that would have taken years and millions of dollars."

It's a utopian vision, to be sure. Maybe when all is said and done, remote towers will just be about saving money, consolidating facilities instead of just replacing them one-for-one.

I suppose even that would be okay. But, if it turns out that in a few decades I'm approaching a snowy, backwoods airstrip at night and a quick radio call activates an on-demand team of professionals to help me land safely, that's a future I'll embrace.

A commercial pilot with instrument privileges, Grant Opperman is a writer and business strategy consultant who flies himself to more than 20 states across the U.S. for business and pleasure.