Will Passengers Ever Fly on Pilotless Planes?
(Inside Science) -- Autonomous cars from companies like Google, Uber and Tesla will soon become commonplace on our roads, according to some experts, and aircraft manufacturers are betting that it will only be a matter of time before the skies are filled with autonomous planes as well.
On Oct. 5, the aerospace giant Boeing announced plans to acquire Aurora Flight Sciences, a small Virginia-based company that specializes in autonomous flight systems. Aurora's Optionally Piloted Aircraft, called the Centaur, can be flown by a human pilot in the cockpit with help from the automated systems, or remotely by a pilot on the ground. The company also has a robotic copilot that has successfully landed a simulated Boeing 737, and will provide unmanned air taxis to Uber Elevate, the ride-sharing company's flying car concept.
Other companies are getting involved as well. Diamond, which makes small aircraft, has a system called eSafe, which can land a plane by itself if the pilot becomes incapacitated. It will choose the nearest airport, route around any bad weather in the area and control the flaps and landing gear during the descent using a GPS system enhanced with radar and laser altimeters to gauge the aircraft's distance to the ground and navigate the approach. And Boeing's major competitor, Airbus, recently tested a pilotless helicopter. Its A3 division plans to test its small passenger and cargo drone, known as Project Vahana, by the end of this year.
But how long will it be until passengers on commercial airliners are whisked across the world with no one in the cockpit?
The challenges facing autonomous planes differ from those facing autonomous vehicles on the ground. Although a self-driving car must usually deal with a more complicated environment than a plane -- in terms of other traffic and the physical landscape -- self-driving cars often have it easier when it comes to dealing with a sudden incident.
“For a car, it's usually safe to stop until the danger passes,” said Ella Atkins, an aerospace engineer at the University of Michigan. “An aircraft has to land, it's not safe until it's stopped on the ground. That can be a much harder problem.”
Still, the challenge seems surmountable, she said. She noted that today's aircraft already have many automated systems, which would not require much tinkering to take on even more responsibility in flight. “Flight management software was never designed to be fully autonomous,” she said. “But that doesn't mean that the very smart people at Boeing and Airbus couldn't do it.”
The remaining technical problems have more to do with improving the human-machine interface, said Stephen Rice, a psychologist who studies how people react to automation at Embry-Riddle Aeronautical University in Daytona Beach, Florida. A remote pilot responsible for several different craft lacks the feel for what's going on in each plane. If they have to intervene in an emergency it can take precious time to get back up to speed on the situation, he noted.
Atkins said a robot pilot might face problems dealing with human air traffic controllers using plain language, which can be difficult for computers to parse. A better option, she said, might be to replace the humans on the ground as well, and have everything controlled by automated systems communicating through secure data links.
Even once all the technical issues are ironed out, however, it will take time for the technology to be adopted. “Today's technology takes 15 years to get into the cockpit of a new plane, because of government regulations and the complexity of the product,” said Rice.
Mark Millam, vice president, technical, of the Flight Safety Foundation, a non-profit organization in Alexandria, Virginia, that provides safety guidance to the industry, said the regulations surrounding air travel are probably a bigger obstacle to automation than in other industries. “It's a big certification process” for new technologies, he said, that takes several years.
But neither bureaucratic red tape, nor technological gaps will likely be the main hurdle that pilotless planes need to cross.
Instead, said Rice, the issue is whether people are willing to travel on a plane flown remotely or by a robot. His research has shown that it will be a tough sell. “People are semi-OK with removing one pilot, but they are reluctant to get onto a commercial flight without any humans in the cockpit,” he said.
They will eventually come around, said Millam, as the improvements in safety -- and perhaps more importantly, cost -- become clear. “If you can show an improvement in safety and a reduction in cost compared with crewed planes, it's going to be hard for the public to walk away,” he said. And younger generations are already more comfortable with automated technologies, he noted.
Both Millam and Rice expect that pilotless air taxis, such as those being tested now in Dubai by the German company Volocopter, will be the first step towards automated commercial flight. Then Rice expects it to spread into commercial cargo delivery: “People don't care if their Amazon package goes down somewhere in Alaska,” he said. After that, short commercial passenger flights in remote areas would likely be next. “If there are no spectacular accidents, people will accept it,” he said.
It will still be a long while, however, before the technology makes the jump from small taxis and package delivery to large-scale commercial passenger flights. “I don't think we're going to see it even in the next 20 years,” said Rice. “It will probably be more like 30, and even then only a few.”