Slashdot Mirror
Pilotless Planes Could Save Airlines $35 Billion Per Year, But Passengers Aren't Willing To Fly In Them Yet (fortune.com)
An anonymous reader shares a report from Fortune: Autopilot is hardly a rarity in the world of commercial air travel. But when it comes to a fully automated flight, most people say "hard pass," at least for now. The pilotless plane could save airlines as much as $35 billion per year, according to a new survey from UBS, reducing the cost of highly skilled employees ($31 billion), related training ($3 billion), and fuel ($1 billion). The deployment of autonomous technology could result in significant fare cuts, an estimated one-tenth of the total in the U.S. And yet 54% of passengers refuse to board a remote-controlled plane, according to the survey of 8,000 air passengers. That sentiment will change over time, the investment bank notes. By the middle of the century, the majority may be willing. But UBS said passengers won't do it today, even if ticket prices were lower -- a big hurdle to airlines, which the bank estimates could see profits double by using the technology. Much like the automotive industry, most passengers don't realize that there are quite a few autonomous systems already in place on today's aircraft -- including those that land the plane.
Technology can and does fail, due to bugs or intrusion. I want a human as a backup. Backup systems are usually a good thing, especially when you are thousands of feet high.
Table-ized A.I.
We all know the airlines will use this technology to replace the pilots, but they'll keep the fares the same, sell a few more seats in the cockpit, and then kill your dog just for laughs.
The cost of one crash where they were judged negligent could bankrupt just about any airline, so there is more to this than the cost of salaries.
This is the point. Why should it mean increased profit for airlines rather than lower prices for passengers?
I'd have thought the first step would be to relieve the need for a copilot?
Qantas Flight 32.
Autopilot is great when everything goes smoothly. But the moment things go pear shaped, it's not so good. Qantas actually ran the incident in their simulators after the event; pretty much everybody who tried to complete it ended up with a hull loss (aka: the plane crashed.) It was pure luck that they had one of their most experienced flight crew on board, who managed to land the craft with no loss of life.
That's the reason why I'm not comfortable with fully automated (no human pilot on board) flights. Yes, flights these days are mostly automated. But the pilots are trained to the Nth degree to handle matters when things go wrong, and that's why they're paid the big bucks. It's not like trains or automobiles, where if something goes wrong, you can just pull over and get out of the vehicle, after all...
But they WILL get rid of the co-pilot. Aircraft are all about the back up systems, and the human pilot is a good one. That's why they have the co-pilot now. They won't get rid of all people in the cockpit.
Instead they will have one pilot as an emergency back up, with the computer doing the flying 90% of the time.
excitingthingstodo.blogspot.com
I just solved that by inventing the Grope-A-Matic.
Table-ized A.I.
When a plane can land on the Hudson River without a pilot and without a loss of life, I'll be the first to buy a ticket.
Automation in the cockpit is great, except when something happens that wasn't expected, then, not having a pilot who has experience and skill is a death sentence. In fact, some would argue that we have already automated too much of the work pilots are doing, leaving them with few chances to actually practice their flying skills, increasing the danger should something unexpected actually happen.
How many times will we blithely assume that we can just automate complex tasks like flying passengers around? In the grand scheme of things labor costs of pilots is literally a drop in the bucket compared to fuel, logistics and maintenance. Are we going to do away with the cabin crew and their salary costs too? I mean they are only there to pass out peanuts and drinks (or the odd overpriced meal) to passengers. Why are THEY there? Oh wait, you say they have a safety component to their job? You don't say, and Pilots don't?
Personally, give me two happy pilots, well paid, well rested and well practiced who have actual flying skills sitting up front. I feel safer having to people who are likely going to be the first to die if we crash. I suspect they will put their best effort into saving us, given the situation.
Like my father, who worked for United Airlines for more than three decades said... "Pilots don't get paid for what they do, they get paid for what they can do when necessary." Stop trying to be cheap and pay up...
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
There are people that consistently fly the cheapest airlines, despite there being other options because they can't afford anything else. Flying for work was an eye opening experience into how the 'other half' lives.
Start up a pilotless airline that flies some core routes and consumers will beat a path to your door. And while you're at it hire some industrial engineers to improve loading and unloading times.
This could probably easily be done for FedEx, UPS, USPS, etc. Also, check your bags and they get put onto an auto-cargo jet while you get on a luxury cruiser with a pilot who parties with you. I'm into the future if this it.
Then 54% are ignorant about the operations of a modern commercial airliner.
The onboard computer systems already control the mechanical operation of the flaps, the rudder, ailerons, the stabalizers, the landing gear, the ventilation, hell, pretty much everything. A pilot's primary responsibility is managing and executing the decision-making. Yes, they can take manual control, but why, when the computer is much faster, more accurate, and more efficient? Just watch this video and hear a commercial pilot talk about how autopilot alone works. All the pilot does is input all the data into the autopilot, setting the course heading, the speed, and the altitude, and autopilot does the rest.
Pilot operations of commercial aircraft are very procedural, and it can very easily be converted into an algorithm managed by a computer. At the very least, it would not surprise me if, in the next 20 years, the FAA determines it's safe for computer automation to reduce the number of required pilots in an airplane from two to one.
Lower the prices, then slowly raise them back up again while dropping the option for tickets on human piloted planes.
“Common sense is not so common.” — Voltaire
If you got rid of the pilot and co pilot, you would barely see a blip in your ticket price.
Let's do the math.
2 pilots @ $200k each == $400k
Training etc for those pilots @ $200k each == $400k
1 trip per day assuming 3 weeks vacation a year: 5 * (52 - 3) == 245
Cost of pilot per trip: $800k / 245 == $3.27k
There are approximately 200 seats on a 737, so that's $3.27k / 200 == $16 per ticket potential savings
Now for an airline, that might make sense on a large scale because they'll reap millions a year in savings, but for consumers it's barely a blip on the radar.
These are with conservative estimates. The salary I took was the highest in the range on glassdoor, I'm assuming all their fancy simulator time doubles their salaries, and most pilots fly short haul flights so they rack up multiple flights a day. Wikipedia confirms the number of seats for a 737, but of course if you have a cabin of first class passengers there are less seats, but still it wouldn't matter.
Additionally, insurance companies will likely charge increased premiums for a pilotless craft, so at the end of day the savings will be considerably less.
The only time you would conceivably see a savings big enough to care would be with a transcontinental flight where you might have four or more pilots (because they sleep in shifts and rotate out). But, compared to the ticket price, I suspect the savings will be marginal.
I suspect there would also be additional overhead as pilots have other functions than flying. For instance, determining if a reroute is necessary or if a passenger is fit to fly.
Additional training and delegation of these duties would most likely raise the cost of other crew.
So, in the end, this is a non issue. Until AI auto pilot comes in a cheap as in uber quad copter that will taxi you where you want on demand, we won't see AI in the sky
References:
https://en.wikipedia.org/wiki/...
https://www.glassdoor.com/Sala...
Operate robotic cargo flights for a few years and a couple million takeoffs and landings, and there will be a quantifiable record of their performance in all sorts of conditions.
Pretty hard to argue with going robotic once the record shows computers consistently outperforming human pilots.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
I'm with you on planes, not so much on cars. With a car 'just stop here' is almost always a safe failure mode, and a significant percentage of the drivers out there are complete morons who have no business walking around unaided, let alone piloting a full ton of fiberglass and sheet metal.
I'm of the opinion we would do MUCH better with driverless cars because so many drivers are so damn bad that even with the inevitable failures of the AI, we'd kill far less people on the roads than we do now.
A thousand pounds of wood moving at 300 feet per minute. Don't get in the way.
Your estimate of pilot's salaries is way high, I think. Many of the smaller regional jets have pilots that earn 20-40K$/year, but those planes have something like 50 passengers. The ones earning the big bucks are a dying breed. You als need to account for the various benefits of course.
Earl Wiener, 55, a University of Miami professor of management science, telling the Airline Pilots Association (in jest) about 21st century aircraft:
"The crew will consist of one pilot and a dog. The pilot will nurture and feed the dog. the dog will be there to bite the pilot if he touches anything.
-- Fortune, Sept. 26, 1988
Welcome to modern aviation, where Dairy Queen is a better paying job until you have a decade of seniority at your current employer. New rules that went into effect after the Colgan 3407 crash mandate that airline first officers (co-pilots) must have a minimum of 1500 flight hours to be hired. That's 6 times the pre-2013 minimum of 250 hours, and the limit was imposed overnight. Those 1500 flight hours have to be earned somewhere.
The general career path for someone wanting to fly for a major airline is PPL (private pilot), a few years working as a CFI (flight instructor) to build time, getting a CPL (commercial pilot) license, flying commuter / cargo to get to 1500 hours, getting hired at a regional as a first officer, upgrading to captain, getting hired at a major airline as a first officer, upgrading to captain. That journey can take 15-20 years (there is no "pilot shortage" any more than there's a shortage of American IT workers), and you won't be earning very much along the way.
Pilots are desperate to get to 1500 hours so they can apply for a position at a regional airline and get their career ladder started. Commuter carriers know this and take full advantage of pilots. There are operations out there where you actually pay the company for a job in the cockpit.
As someone that works in the area of providing data for air navigation I can unreservedly say that the data to support completely automated flight from off-blocks to on-blocks, is simply not there yet. Aviation is chock full or rules, exceptions, regulations, grey areas, short term changes, and unexpected events; all currently best dealt with using a Mk 1 Brain (two in most cases). Not to say it couldn't happen for a good chunk of regular passenger/freight transport between major, 1st world, domestic centres in the fullness of time. Aviation change moves on timescales of a decade or more, not months.
Ultimately though, it is naive in the extreme to think you are going to save billions by not hiring pilots. All that experience has to get into aircraft and ground systems to make this work. That will not be happening as a matter of charity. What you save in pilots you lose in equipment costs, airway navigation and landing charges.
Patent litigation: A doctrine of Mutually Assured Destruction... in which everyone seems willing to push the button
I fly the airbus A320. The plane was developed 30 years ago and they are still finding software bugs in it. Every now and then some combination of conditions happens that nobody planned for and it's up to the pilots to fix it. Without pilots, the number of crashes would have been astronomical. Computers fail all the time.
And that's what I think people don't yet grasp with automating passenger-carrying aircraft. How many times has a circuit breaker popped and you couldn't reset a system? I've flown autopilots that were so sensitive that just the suggestion of turbulence was enough for them to transfer control to the human. Others could fly through Class 5 whitewater. Systems fail. Regularly. Systems fail in ways not predicted by manufacturers. I'd love to know if a computer could have landed a DC-10 in Sioux City, IA without knowing (or expecting) the failure that occurred in that accident. What does a computer do when there's a complete hydraulic failure in an aircraft that "will never have a complete hydraulic failure"? The big question is, how do you program a computer to handle a nearly infinite number of contingencies? I know AI will be the answer, but even AI must learn. If it hasn't seen a problem before, do you want it learning something new with 300 meat missiles in the back?
ATC and the tower will be automated too.
We got rid of lighthouse keepers long ago, if there had been an internet back then I'm sure we would have heard plenty of people telling how it would be a terrible idea to replace humans with automated lighthouses.
We humans like to think we are really exceptional thinkers who couldn't possibly be replaced by a machine. We're mostly wrong. Heck, even human trolls were mostly replaced by bots during the last election.
No, you are forgetting about all the incidents where a crash did not occur because the pilots were there to prevent it. Those are orders of magnitude more frequent, I have had several of those in my 22 years of flying,
Let me give you an example of "pilot error": a Turkish Airlines flight to Amsterdam. Automatic approach. More than 1000 ft above the ground, a radio altimeter error fools the autopilot into thinking the plane is just above the runway, so it pulls the throttles back to idle. The speed decreases and the plane stalls. Pilot error because the pilots should have been closely monitoring the speed and should have immediately reacted (as other crews have on numerous occasions). Without pilots, the plane would have crashed every time.
That's according to Boeing.
Flying high speed aircraft with meatbags behind the controls will someday disappear.
Anyone want to know how SpaceX controls their very high speed rockets? Answer, entirely with computers. People are too slow, make too many mistakes and are best suited for sitting back in their seats and letting the software run the ship.
"Most aircraft accidents are human error, so relative risk is likely lower."
That assumes that autopilot is at least as good as a human pilot in every situation.
What the concern many have is what happens when the autopilot fails? It's just absurd to think that such a complex program will never meet a real world situation it wasn't designed for. Or suffer damage to itself. Or have erroneous sensors. Or be hacked.
"The 9000 series is the most reliable computer ever made. No 9000 computer has ever made a mistake or distorted information. We are all, by any practical definition of the words, foolproof and incapable of error."
You really think that was trained for? No way. We don't train for it now after it's happened. We don't train for ditching in the river after Sully did it. All of that is accomplished by combining human experience with the human ability to improvise and adapt.
You know what's wrong with that statistic? Not that it's not true, but that it is misleading. In commercial aviation, crashes are vanishingly rare events. You are more likely to die on the way to the airport by a couple orders of magnitude than in a airplane accident. Pilots are SAFE, really really safe.
What statistic you need to make any reasonable comparison here is how many times pilots have "taken over" from the automation to avoid a crash and compare that to how many times pilots have "taken over" and THEN crashed a perfectly flyable aircraft. Problem is, you don't have ANY idea how many times pilots have prevented crashes because these events happen all the time and are not reported.
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101