Friday, July 28, 2017

Will robots rescue, or threaten, the airline industry?


Yesterday it was reported that US airlines were suffering a "staggering pilot shortage".

Passenger and cargo airlines around the world are expected to buy 41,000 new airliners between 2017 and 2036. And they will need 637,000 new pilots to fly them, according to a forecast from Boeing released this week. That staggering figure is matched only by how many will leave the profession in the next decade -- particularly in the U.S.

Retirements at U.S. airlines will start to rise precipitously starting in 2021 as the current crop of pilots turns 65, the mandated age of retirement. More than 42% of active U.S. airline pilots at the biggest carriers will retire over the next 10 years, about 22,000, according to a recent report by Cowen & Company.

In the next 20 years, airlines in North America are going to need 117,000 new pilots, Boeing estimates. And the farm team for training and recruitment in the U.S. -- the military and regional carriers -- are already struggling to find and keep aviators.

The coming retirements exceed the active U.S. regional airline pilots corps, which stands around 19,000.

Without enough pilots, the amount airlines can fly will be capped. And an acute shortage may wreak havoc on air travel, grounding planes and reducing air service to some cities if routes are cut or curtailed.

It's already happening.

Last month, Horizon Air, the regional arm of Alaska Airlines, said it was canceling 6% of its schedule -- more than 300 flights -- from August to September because it doesn't have the pilots. And Republic Airways filed for Chapter 11 bankruptcy protection in 2016 in part because it was "grounding aircraft due to a lack of pilot resources".

There's more at the link.

This pilot shortage has been developing for several years, and airlines and the military have been devoting a lot of time and attention to dealing with it.  Perhaps the best-known technological approach is DARPA's Aircrew Labor In-Cockpit Automation System (ALIAS).  The program "envisions a tailorable, drop-in, removable kit that would promote the addition of high levels of automation into existing aircraft, enabling operation with reduced onboard crew".  An early iteration was flight-tested last year.

Yesterday, a turboprop plane took off from a small airport in Virginia that from the outside, looked fairly unremarkable.

But inside the cockpit, in the right seat, a robot with spindly metal tubes and rods for arms and legs and a claw hand grasping the throttle, was doing the flying.



The demonstration was part of a government and industry collaboration that is attempting to replace the second human pilot in two-person flight crews with robot co-pilots that never tire, get bored, feel stressed out or become distracted.

. . .

Sophisticated computers flying planes aren't new.

But the ALIAS robot goes steps further.

For example, an array of cameras allows the robot to see all the cockpit instruments and read the gauges.

It can recognise whether switches are in the on or off position, and can flip them to the desired position.

And it learns not only from its experience flying the plane, but also from the entire history of flight in that type of plane.

The robot 'can do everything a human can do' except look out the window, Langford said.

But give the programme time and maybe the robot can be adapted to do that too, he said.

The programme's leaders even envision a day when planes and helicopters, large and small, will fly people and cargo without any human pilot on board.

The programme, known as Aircrew Labor In-Cockpit Automation System (ALIAS), is funded by the Defense Advanced Research Projects Agency (DARPA) and run by Aurora Flight Sciences, a private contractor.

. . .

Elements of the ALIAS technology could be adopted within the next five years, officials said, much the way automakers are gradually adding automated safety features that are the building blocks of self-driving technology to cars today.

Again, more at the link.

Here's a video clip of the ALIAS system being demonstrated on a helicopter and a light aircraft.  I suggest watching it in full-screen mode, to see the smaller inset videos to best advantage.





There's even a possibility that aircraft might fly without co-pilots at all.  Instead, multiple aircraft might have a single pilot, with all of them being assisted by ALIAS-type systems in case of emergency, directed by a controller on the ground.

NASA is exploring a related possibility: moving the co-pilot out of the cockpit on commercial flights, and instead using a single remote operator to serve as co-pilot for multiple aircraft.

In this scenario, a ground controller might operate as a dispatcher managing a dozen or more flights simultaneously. It would be possible for the ground controller to “beam” into individual planes when needed and to land a plane remotely in the event that the pilot became incapacitated — or worse.

. . .

The potential savings from the move to more autonomous aircraft and air traffic control systems is enormous.

In 2007, a research report for NASA estimated that the labor costs related to the co-pilot position alone in the world’s passenger aircraft amounted to billions of dollars annually.

Automating that job may save money.

More at the link.

Here's another video, showing the ALIAS hardware and software controlling a Boeing 737-800NG simulator.  It's a very short step from this, to putting it aboard a real airliner (replacing the co-pilot's seat) and taking it flying.





Of course, this automation technology might also pose a real threat to airline operations as they're currently structured, because it can be applied to other modes of transport as well.  Karl Denninger hypothesizes:

Prediction: Within 10 years every single airline will be reduced to carriers that operate routes consisting entirely of flights of more than 1,000 miles, most over water.

Why?

Because self-driving cars.

. . .

Look folks, most cars today can be retrofitted ... Show me a $500 Lidar array that can do the job and suddenly that $2,500 retrofit becomes not only possible it's easy and it's an option roughly equivalent in cost to a leather seating package on new vehicles.  At that point the "take rate" will be 90%.

Today I can drive from my home to Atlanta in about 5 hours.  All-in, including "mandatory" 1 hour pre-take-off airport arrival requirements it takes me almost 4 hours assuming no weather or schedule delays to fly that same route ... Actual operating cost of said autonomous vehicle is materially cheaper than the flight is and I can take a nearly-unlimited amount of cargo with me at no additional cost ... The day I can get into the car at midnight in the back where I have equipped half the fold-down rear seat and trunk into a comfortable place to sleep, push the button, go to sleep and wake up at 6:00 AM (1 hour time zone shift) in Atlanta in time for two espressos before a business meeting Delta is bankrupt.

. . .

Folks, there is no business model for the airlines as they exist today once this becomes rationally expensive ... Not only is this more-convenient and "on demand" rather than on someone else's schedule nobody gets bumped, nobody gets groped, there's no "extra fee and insult" garbage the airline industry has turned into a maze of and it's cheaper on top of it.

The airlines have cut their own throats, in short, and technology is about to kill them all, with the exception of 3,000 mile flights and over-water segments where you simply can't do it any other reasonable way.  That's a fraction of their current capacity and operating schedule and I'm going to enjoy watching them all burn in bankruptcy court.

More at the link.

I must admit, the thought of being able to avoid almost all airline travel is a very welcome one.  I long ago grew sick and tired of airlines handling me as if I was a cow on the way to the slaughterhouse, cramming me into an aluminum tube with minimal space or facilities (not to mention the TSA treating me with utter disrespect in the process!).  I hope Mr. Denninger is right.

Peter

17 comments:

Eric Wilner said...

Hmmm.
Now I'm having a vision of no-human-pilot, family-size, aerial travel pods, coordinated through networked ATC.
If it's designed to fly itself, it doesn't need an instrument panel (the computer just gets data directly from the sensors). In the event of a loss of cabin pressure, the computer remains functional as it negotiates with ATC to reach a safe altitude for the passengers.
Not having that scarce and expensive pilot on board, the pod can be one of zillions.
I'm not so sure about the practicalities of range, refueling stops, maintenance, and so on - but advances in materials, propulsion, and monitoring should help with range and maintenance, and surely automated fuel bowsers aren't too much of a challenge.
The whole thing sounds like a cybersecurity nightmare, though. And it'd probably result in human-piloted general aviation being essentially shut down as an inconvenience to the System.

Quentin said...

Have you noticed how little pilots get paid? No wonder there's going to be a shortage.

jbrock said...

I'm with Denninger 100% on all that. I already refuse to fly anywhere I can drive to in 48 hours or less. Show me an affordable autopilot retrofit for (e.g.) a Mazda minivan and the airlines can commit multiple physically implausible acts. On themselves, not on me any morem

Anonymous said...

One a brighter note the Autopilot will not knock up any flight attendants, get a nasty divorced and end up a drunken wreck.

Gerry

Rusty Gunner said...

Most people would consider a shortage of staggering pilots to be a good thing.

mark leigh said...

If the hyper-loop concept works out well airlines will find a lot fewer passengers as travelers take the tube.

Billll said...

Instead, multiple aircraft might have a single pilot, with all of them being assisted by ALIAS-type systems in case of emergency, directed by a controller on the ground.

As long as the controller on the ground isn't named Mohamed, works from his basement, and has a chip on his shoulder.

tweell said...

Otto, from Airplane (the movie).

http://thebottomofabottle.files.wordpress.com/2012/10/airplane-autopilot.jpg

Anonymous said...

What Quinten said. I realize things have improved since the bad old days (mid to late 1980s) when commuter airline pilots got paid $12,000 year, less uniform and training costs, but 1) the military pipeline is disappearing, 2)the civilian pipeline of kids who learned to fly at Little Local Airport got choked closed by fuel costs and security concerns, among other things, and 3) you are out the cost of a college degree plus the cost of all the fight hours needed and the extra training, and even then it was still a cr@p-shoot if an airline's HR department would bother looking at your resume. I understand the need to gain experience and "pay dues," but a lot of younger people don't want that kind of loooong apprenticeship.

My $.02.

LittleRed1

Borepatch said...

Well, economic theory says that if you have a shortage of a resource that the price of that resource will go up. If it doesn't, you will get a shortage.

I know that's crazy talk, but still.

I don't have any sympathy for the airlines. It's a competitive market, including for labor.

And I would much rather drive to Atlanta than fly there - and it takes me 10 hours, so it's a fair amount faster to fly and probably no more expensive. But my Jeep doesn't grope my junk.

Anonymous said...

Hey, if it saves you $50 lets do it, right? Safety? Screw that, its $50. Look how many people still fly on korean to save $50.

And we all know those pilots are a bunch of oversexed drunks, right Rusty/Gerry?

I dont have the space or time to attempt to explain the career. I have been at it for 30 years now. Working for a paycheck, no retirement, crap benefits, no holidays, no family time, just as much fondling and groping as the pax (more actually). Trust me, you wouldnt understand how bad a job can be when 75% of the employees cant wait to retire (with zero, zip, nada retirement) Social security here I come These are folks who spent perhaps 10 years just becoming qualified to apply for the job; college, flight training, flying time. It pays less than most plumbers (in a metro area). No you cant just go to another company, its indentured servitude and the company knows it.
The job pretty much sucks, the companies are hostile to employees (and I work for one that has a great rep for family care- puke). The pay is OK if you work your ass off and therefore wind up divorced because of it.
I dont understand why anyone would choose this as a career, its garbage.

CDH said...

As a long time professional in the industrial controls field (mostly petrochemical, but the comparison is valid) I do everything I can to eliminate human factors and automate all safety critical functions. Not much in flying an airplane is not safety critical! With very few exceptions, I would feel much more comfortable with fully automated cockpits.

Now I'll hear the criticism "What about when an emergency happens outside the programming of the controllers?" Rare and becoming vanishingly rare. In the early days computers were not powerful enough to foresee or compensate for all possibilities. The current state of the art is so much past those early years there is no comparison. Even 20 years ago, pilot errors were more hazardous than automation failures. Now drones can handle combat missions (talk about adapting to random events on the fly!) better than humans, so scheduled flights are trivial by comparison.

Mike said...

I'm working on some of the tech for those self-driving cars. It's not going to happen nearly as fast as most people seem to think. There's still a long way to go, and I expect the first spate of major wrongful death lawsuits to put a serious damper on the enthusiasm of the industry for this tech.

VFM #7916 said...

Things will be peachy with automation until two factors come to the fore:

Weaponized cyberwarfare on a human scale.
Economic or political crash due to so many useless mouths.

!!FUN!!

Nate Winchester said...

@CDH - Yeah I'd believe that if our copiers wouldn't have a freakout every time someone changed the size of the paper in the tray.

Landing the plane is the real trick. We'll see if robots ever get to that, but you can go first.

Joseph Bridges said...

Eh...impressive technology usage, but for me, not really an issue - yet - mostly due to the simple fact that, except for some sort of unforeseeable (near-future) extreme-emergency situation, I will NEVER voluntarily board a commercial aircraft again - and that's the way it's been for me for going-on five years, now. Even though I've been "under-employed" now for a couple of years - to the point of being involuntarily "semi-retired" - I have actually turned-down opportunities to interview for job openings in my primary field (as a contract design engineer, mostly in aerospace mechanical/electromechanical work), because they were openings that would have necessitated airline travel to/from the job location (either that, or being away from home continuously for 6 to 8 - or more - months at a time, the complete term of the contract).

I simply decided to opt-out of dealing any further AT ALL with the essential insanity that is - and has been, for nearly a decade - U.S. airline travel. It had simply become too much...

If the job isn't within about 3 - 4 hours' driving-distance of home - I'm not interested, right up-front...yes, it got that bad. I've adjusted to no longer traveling by airline - even if it means a lowered level of employment (and, of course, income).

So...I really don't care who - or what - flies the plane(s).

It does strike me, though, that I'd really, really be interested in a truly safe, efficient self-driving car...

Consider this: If they actually start flying commercial aircraft using robotics/cybernetics like this - how long do you imagine it will be before they start doing the same thing with long-haul commercial trucks, where it's always a struggle to get good drivers who are able to deal efficiently with the demands of that job while getting the goods hauled to their required destination(s)? Seems to me like a very natural "fit", y'know?

Will said...

Last time I flew commercial was ten years ago, for a funeral. Not going to deal with that asinine security theater that is TSA any more. If I can't drive, I'm not going, and that would be coast to coast.

Normally I was flying one way, and driving the other. And with guns. Those clowns didn't seem to understand any of it, and I would become the center of focus every time I flew, to the point of nearly not making it to the gate, even though I was there a couple hours prior to boarding time. The gun handling/checkin changed every time. Obviously the airlines don't want the business, since they tolerate the TSA messing it up.

Add in how bad the airlines themselves treat their passengers, and they can FOAD.