Drone alone: how airliners may lose their pilots

Uncrewed planes that can operate in civilian airspace are almost here – a first step on the road to pilotless airliners

WOULD you fly in an airliner knowing there were no pilots in the cockpit? This is no mere hypothetical question. The US Federal Aviation Administration this month kicked off what could be the first step in a journey towards the full automation of the airliners we all travel on.

The FAA commissioned the Boeing subsidiary Insitu, based in Bingen, Washington, and the New Jersey Air National Guard to begin investigating ways for civil aircraft to share their airspace with remotely piloted uncrewed aerial vehicles (UAVs). In the UK, a research programme called Astraea 2, led by BAE Systems and Airbus owner EADS, is pressing ahead with similar aims.

The goal on both sides of the Atlantic is to allow UAVs to share civilian airspace, rather than clearing a section of airspace for every UAV flight, as happens now. While this segregation of the sky has prevented collisions, arranging clearance for every flight is time-consuming and curtails potential fly-at-a-moment's-notice applications for UAVs.

If they are to share civilian airspace, uncrewed planes will need to be able to sense the presence of other aircraft and take evasive action. Air traffic controllers would also need robust ways to manage flocks of UAVs - whose remote pilots may be hundreds of kilometres away.

Why contemplate opening up the sky to drones? For one thing, they could provide a relatively cheap boost to border-monitoring efforts, says Lambert Dopping-Hepenstal, an Astraea 2 project chief at BAE. At the moment, even in a country the size of the US, he says, "there is not enough segregated airspace" to support such schemes. Other mooted applications include surveillance, search-and-rescue and crop monitoring.

You just don't need someone with a million dollars' worth of pilot training any more

It won't stop there. Once UAVs can avoid passenger jets, remotely piloted cargo planes are likely to take to the skies, pushed by some compelling economics. "The cargo airlines want very much to lose their pilots. The money that would be saved in salary and benefits, including retirement and healthcare costs, is pretty staggering," says Mary Cummings, a former US navy fighter pilot who now researches ways to automate aviation systems at the Massachusetts Institute of Technology.

Indeed, some UAVs are already big, cargo-capable beasts: the largest variant of the Northrop Grumman Global Hawk has a 38-metre wingspan, similar to that of the aviation industry's veteran workhorse, the Boeing 737.

The FAA's project will initially focus on air traffic control procedures, whereas Astraea 2 researchers are focusing on a sense-and-avoid system for UAVs that is comprised of "cooperative" and "non-cooperative" elements.

Today's airliners use a cooperative system called the Traffic Alert and Collision Avoidance System (TCAS), whereby radio transmitters on each plane announce its position, height and heading. The system constructs a picture of what's in a plane's airspace and calculates collision risks. If a risk is detected, a loud, synthesised voice tells the crew to climb or dive to avoid the danger.

"UAVs will have to respond to these TCAS alerts," says Dopping-Hepenstal. But when planes without TCAS venture close, things get tricky. Some of these "may be radio silent or have low electromagnetic signatures, making them difficult to detect", he says. This is where the non-cooperative elements kick in. Astraea is developing a battery of sensors, including infrared heat sensors, millimetre-wave radars and optical cameras, to ensure UAVs know if a plane is nearby.

While infrared sensors and cameras should spot a plane in open air, they may lose it in cloud. That's when millimetre-wave radar, which easily pierces fog, takes over.

"Using multiple collision-sensing methods is clearly the safest approach," says Noel Sharkey, an autonomous-systems engineer at the University of Sheffield, UK. "But what happens in unpredictable circumstances? Human creativity can avert serious problems that have never arisen before." And while UAVs will have a remote pilot, "what happens if the wireless link to the pilot gets lost?" he asks.

Even trusted technology like TCAS fails on occasion, which worries Peter Ladkin, a safety-critical systems engineer at Bielefeld University in Germany. "Flying UAVs in civilian airspace, and mandating safety devices for them and their airspace co-users, has large, maybe even overwhelming, political, legal and social dimensions. It is not just a technology issue," he says. The repercussions from fatalities caused by the failure of a UAV's sensing technology would be profound, Ladkin adds.

If round trips from Boston to LA on an unmanned jet cost only $50, people would stampede to get on board

Cummings is convinced that the cost-saving potential will prove compelling. Autopilots and auto-landing systems - by which aircraft automatically follow a radio beacon to the runway - already do so much of the actual flying that pilots will eventually be seen as an unnecessary expense, even in passenger planes. "You just don't need someone with a million dollars' worth of pilot training any more," she says.

Carolyn Evans of the British Airline Pilots Association says UAVs will "still need a remote pilot trained to the same level of expertise. But they might not need a co-pilot - so pilots, as you might expect, are not overly keen on the idea."

Would Evans fly on a remotely piloted plane? "It's like the spartan offerings of some budget airlines: it's not a product I like the sound of. So no, you won't find me catching one."

Cummings is unperturbed. "If round trips from Boston to Los Angeles on the first new unmanned airliner cost only $50, people would stampede to get on board," she says.

Issue 2766 of New Scientist magazine