“We are rapidly coming to the point where airplanes can autonomously take off, fly to their destinations and land without human intervention” from: http://marshallbrain.com/robotic-nation.htm
So how close (or far) are we to a pilotless airliner? Is the aviation industry doing any research in this area or is too far off?
thanks
Bob Z
I’m an airline pilot. Airline management would love to replace us with something cheaper, but the BIG issue for them is whether the public would be willing to get on the thing.
Right now the military has almost fully automated airplanes flown by a combo of computers on board and pilots in vans on the ground. So far they’re used for reconnaisance. Google for [UAV] to learn more about these. The first prototypes intended for bomb-dropping / missile launching are being built now. Google for [UCAV] to learn more about these.
In a combat zone, collisions and crashes are undesirable, but tolerable. In peacetime travel, that’s not acceptable. The FAA and its international counterparts are struggling with figuring out how to mix manned, fully automated, and remotely-piloted aircraft in the same airspace. These are not simple problems.
One of the major civilian issues to deal with is that collision avoidance is still a matter of looking out the window and not hitting anyone you see. An automated airplane is blind, at least for now. Certainly with enough technology they could eventually have cameras and computer vision, or at least pretty good radar, but not any time soon.
So bottom line: the technology for a fully pilotless airliner-sized plane is probably 15-20 years away. The ATC infrastructure and regulatory framework to permit it to haul commercial passengers is maybe 30-40 years out, and the acceptance by the public may be sooner, may be never, depending on whether they have any spectacular screwups early in the program.
I know the idea makes me very uncomfortable, and not just in terms of my career. The variety of problems we deal with is boggling; I have a hard time buying into a computer program dealing with them all with the 6 sigma plus reliability of the current human system.
Aircraft are being flown and landed automatically every day, overseen by the flight crew of course. I was told years ago that it was technically possible to make the whole flight automatic.
The problem arises when something goes wrong.
How nice it is as a passenger, to have a human pilot up there with you, with a vested interest in getting that aircraft down safely.
V
I am a pilot. Have been for 15 years. But even if I werent, I would never board a plane without a pilot on board. Never, regardless of the technology.
This is more opinion on my part since I don’t know if there is a definitive answer.
That being said: A long time, if it EVER happens.
Airliners today are very automated, and many can land themselves. In fact, for the lowest-visibility landings an auto-land is required (the notable exception being the NG 737 with a HUD, where hand-flown Cat III approaches are done). However, this does not mean the airliners can do everything themselves.
Often the most difficult part of a flight is negotiating the taxiways at a busy airport. La Guardia and O-Hare are particularly fun at night and in the rain. How does an “auto-taxier” do this? You could embed transmitters in the taxiways and sensors in the nosewheels, keeping the aircraft on centerline. You would also need auto-throttles and autobrakes that work on the ground. These would have to operate verrry smoothly to get the results you have today. It could be done, but it would be very expensive. You would also essentially cede all control on the ground to the ATC folks, because they would issue instructions to the computers on the airplanes who would then follow the assigned taxi route. If controller one tells Airliner X to cross runway 17R while controller two issues a landing clearance to Airliner Y for the same runway, what happens?
The airplanes would also need sensing equipment to detect other airplanes taxiing around. Just because Low-Cost airways eliminates the pilots doesn’t mean that their won’t be ANY pilots out there. Pleasure pilots will still fly, and they won’t be installing any of the fancy “tracking” equipment that does something that they can do themselves. You now have to separate fully automated airplanes from pilots flying for joy of it.
Once you get airborne, who decides which way to deviate around the inevitable thunderstorms? Who decides that the ride is too bumpy and climbs or descends to a smoother alititude? When the (also inevitable) traffic jams happen, who programs the airplanes to enter holding patterns? The controllers again?
What happens when something goes wrong? I’ve had passengers have heart attacks in flight. Who decides when to divert and where you are going? These kind of decisions are dynamic and need a human involved at some point. If the human is not on-scene (in the airplane) you incur some delay. That delay could be critical for the passenger.
As for mechanical failures, most can be dealt with by a sophisticated enough computer, but there are ALWAYS the ones that no one ever though of. Before Capt Al Haynes brought his DC-10 into Sioux City, IA everyone had said that you could NEVER lose all 3 hydraulic systems on that airplane. They did, and then they flew it down using asymmetric thrust from the two working engines.
But all of these technical issues still miss the biggest barrier to this ever happening: who’s going to fly on these airliners? Millions of people are afraid to fly today. How many are willing to fly across the Atlantic with only a computer guiding them? Computers can (and do) fail. Anyone who has ever seen a “Blue screen of death” or who still has “12:00” flashing on their VCR will have a tough time climbing aboard a fully automated airliner.
So, my opinion is that this will not happen for a very long time. Two generations would be my earliest guess, and that assumes that it ever happens at all.
And yes, I know that my job depends on this NOT happening, but I hope that I’ve not let that cloud the discussion.
A common aviation joke:
The “cockpit of the future” will include a computer, a pilot, and a dog. The computer is there to fly the airplane, the pilot to monitor the computer’s performance. The dog’s job is bite the pilot if he tries to touch anything.
I was on an early '90’s flight into Atlanta. A thunderstorm was raging outside the plane as we descended into airport.
The pilot comes on and tells us how they’d be using their auto-land feature for this approach.
Hmm. Lightning… Thunder… Crosswinds… and a TRS-80 is landing the plane.
I could’ve done without that piece of info.
I think this is a particularly strong barrier to complete automation. While the pilot may not be needed for actually flying the plane, some sorts of decision-making seem to require a human. In the case where physical distress of a passenger may warrant a change in flight plan, how do you tell the computer what the situation is, and how does the computer decide whether to alter course?
This isn’t so bad on, say, and automated commuter train, where the train will stop at the next station in a few minutes. But on a transcontinental or transoceanic flight, the next stop may be hours away, and a passenger’s condition may warrant aborting the flight.
(I’m not an airline pilot, but I am a private pilot and instructor.)
[A joke from my youth]
"Welcome aboard the first fully automated passenger flight from coast to coast. There is no pilot aboard, everything is run by the computer. Don’t worry, nothing can go wrong, go wrong, go wrong, go wrong…[/AJFMY]
I’ve heard it seriously proposed that airliners could fly with a single pilot, who is viewed as a “backup system.” The pilot would be responsible for taxiing, but the flight computer would do the takeoff, the entire flight, and the landing.
If the (presumably triply redundant) computer identified a problem, or if the pilot detected a condition that qualified for special treatment, then the pilot would take over. His job would be not to complete the flight as scheduled, but simply to land the plane at the most sensible airfield within reach, where the problem could be dealt with.
I’m not sure if this is a workable scheme, though it appears to require no technologies that don’t already exist. It seems much more realistic than not having a pilot on board.
A potential problem with a pilot only as backup is that the pilot doesn’t maintain proficiency for when human intervention is really needed.
Well, you don’t get much proficiency in managing emergencies from even a large number of routine flights. To deal with this, airlines have advanced simulators that all pilots must “fly” regularly. This is where you become proficient in dealing with two engines out, a cockpit fire, and a piece of the rudder falling off.