Pilotless Planes? Not So Fast.
UPDATE: April 21, 2015
ON APRIL 10th, The New York Times published an op-ed of mine about cockpit automation and the future role of the airline pilot. This was my sixth op-ed to run in the Times, but easily my favorite, given my long-simmering frustration over the topic in question.
As my regular readers are aware, one of my biggest pet peeves is the public’s widely held belief that jetliners essentially “fly themselves” — the result of a too-gullible media that takes at face value the claims of researchers, professors and tech writers who, valuable as their work may be, often have little sense of the operational realities of commercial flying. Consequently, travelers have come to have a vastly exaggerated sense of the capabilities of present-day cockpit technology, and they greatly misunderstand how pilots interface with that technology.
In the actual newspaper they gave it excellent placement, center of the page, with a graphic showing a Chinese military hat being buzzed by mosquitoes. I’m right between David Brooks and Paul Krugman. Brooks is pretty well-known. Krugman won a Nobel Prize. Not sure what my big claim to fame is. I hope you’ll share the link with friends, and feel free to jump in and leave your thoughts in the Times‘ comments section.
Really the piece is less about the future role of pilots than it is about our current role. My main point being is that it’s pilots, not some computer, that is flying your plane. I only wish they’d give me more space. There’s plenty more to say on the subject. Since the Germanwings disaster, the cable channels, op-ed pages and blog sites have been chattering away about the supposedly diminishing role of human beings in the cockpit.
For instance there was Flying magazine’s Peter Garrison writing in the Los Angeles Times. “From shortly after takeoff to shortly before touchdown,” explains Garrison, “airplanes fly themselves while pilots talk with controllers and one another and punch data into flight management systems.”
That’s up there among the most insulting and misleading characterization of how commercial airplanes are flown ever to appear in print. Garrison is an experienced pilot and should know better than to reinforce this pervasive mythology through such flip and deceptive descriptions. Pilots become their own worst enemies sometimes, not realizing how statements like this are interpreted by the public.
Not to be outdone was John Cassidy on the New Yorker website. “In some ways, human pilots have become systems managers,” Cassidy says. “They prepare the aircraft to depart, execute the takeoff and landing, and take the controls in an emergency. But for much of the time that a routine flight is in the air, a computer flies the plane.” That was good of him to remind us that pilots indeed “execute the takeoff and landing,” which is to say they perform them by hand, but the rest of it is more of the usual nonsense.
The photo accompanying Cassidy’s story shows a simple button marked “autopilot.” I’m not sure what that blue button is for, or what aircraft the picture is from, but the actual autoflight control panel on any jetliner is, suffice it to say, a lot more complex.
There’s more: In the Toronto Globe and Mail, reporter Paul Koring wrote an article called, “Aviation is Fast Approaching the Post-Pilot Era.” He quotes David Learmount, a “veteran aviation expert,” who predicts that “pilots won’t be in cockpits in 15 years but in an airline’s operations room, rather like the U.S. Air Force pilots flying Global Hawks [military drones].” What utter and shameless rubbish.
And my old friend Missy Cummings was at it again, this time fooling a reporter at CNN.com. “Pilots only spend three minutes per flight flying a plane anyway,” she spouts. That’s a disgusting and deceptive thing to say. What she might mean is that pilots spend a relatively little amount of time (though it’s more than three minutes) steering the plane by hand. But they very much are flying it for the entirety.
Occasionally, well-intentioned people will bring up U.S. Airways hero pilot Chesley Sullenberger as a good example of why pilots are still necessary. Point accepted, but still I don’t like this, because it implies that pilots are only valuable in the event of an emergency or some unusual circumstance. On the contrary, even the most routine flight remains a very organic, hands-on operation subject to almost limitless contingencies that require human input. Cockpit automation is merely a tool, and it needs to be told what to do, how to do it, when to do it and where. And though a pilot’s hands aren’t gripping the steering column for hours at a time, as it might have in the 1930s, they are manipulating, operating, and commanding the various systems and subsystems that carry you to your destination.
Here, let me give you a quick demonstration:
I was asked by somebody to talk them through a typical maneuver. A descent, for example. How would I descend my 767 from, say, 25,000 feet to seven thousand feet, with the autopilot on? Well, it’d happen as follows. This is going to be incomprehensible to most of you, but that’s part of the point:
After being cleared to the new altitude, in this case 7000 feet, I’ll first reach up and dial “7000” in the altitude window on the mode control panel. The other pilot will verify this. The next series of steps depends where exactly on the arrival profile we are, but it’s common to activate a VNAV descent using the DESCEND NOW prompt from the descent page of the FMS. Typically I’ll already have the page set up for maybe Mach .79 and maybe 315 knots. This will give you a pretty good rate of descent.
At around 11,000 feet or so, we need to slow down in order to hit the 250 knot restriction below ten thousand feet. You can let the plane do this on its own, in VNAV, but sometimes that carries you off the profile and creates more work, so I come out of VNAV by hitting the VERTICAL SPEED switch. The VS window opens and I dial it back to 1,000 feet-per-minute, or maybe less. The plane’s rate of descent immediately begins to slow. And the instant I hit the VERTICAL SPEED switch, the IAS window also opened, allowing me to set in 250 knots. The thrust levers come back and the plane decelerates.
Now, all I have to do is tweak the rate of descent until I safely hit 250 at or near the 10,000 foot target. I might use 1000 feet-per-minute initially, then reduce it to 500. Whatever it takes. Using the spoilers can be helpful here too (the rectangular panels that rise from the top of the wings). I may already have been using them earlier in the descent if VNAV wasn’t quite holding the profile, or if ATC seemed antsy, etc.
Then, at 10,000 feet and 250 knots, I select FLCH. The 250 knots is now locked in the window and the plane will now hold that speed. I can descend at idle, or use thrust to play with the vertical speed rate, speed-on-pitch style, depending. We’ll be issued several more altitude changes, and I’ll stay with FLCH the rest of the way down, at least until joining up with whatever instrument approach is being used. Some instrument approaches, though, are flown in VNAV, which I’ll reengage later, when it’s needed, and use the speed intervene function of the IAS control to maintain the approach and landing speeds.
And that’s just the altitude control. We’ll have a number of course changes as well, to be dialed in and flown using whatever methods are appropriate (LNAV, heading select, LOC or APP mode…)
And so on. So, why not have the autopilot do this? It is doing it. The autopilot has been on throughout this scenario. This is the automation at work. Point being: it’s the pilots, not a computer, that is controlling the operation. And this is why it is so infuriating when Missy Cummings says pilots are only flying the plane for three minutes.
Granted the 767 is an older plane. It was designed in the late 1970s. There have been a few minor upgrades to the plane’s avionics since then, but nothing major. The plane is still operated exactly as it was when the first 767s were delivered. Frankly, though, even on the newest models, the basics of cockpit automation haven’t changed that much in thirty or more years. The interface between pilot and technology on a 787 or an A350 isn’t drastically different from how it was on a DC-10 or an old 747-200 in 1972. And the Airbus A320, like the one in the Germanwings crash? Its platform technology was developed in the 80s.
People will speak of planes being “highly computerized.” In some respects that true, but not in the sense that people think, and in fact it’s pretty rare for a pilot to refer to any piece of cockpit equipment as a “computer.” That’s just not a word that we use. Obviously certain components are computerized, but look around the typical cockpit and what you’ll see are lots of levers, knobs and switches. Hands-on stuff. Even the most advanced flight deck is yet to have a QWERTY keyboard anywhere on board. Pilots still use an old, ABCD EFGH style keypad to enter data into the FMS or ACARS. Much of our allegedly high-tech equipment is clunky, old-fashioned and user-unfriendly.
To be clear, I’m not arguing the technological impossibility of a pilotless plane. Certainly we have the capability. Just as we have the capability to be living in domed cities on Mars. But because it’s possible doesn’t mean that it’s affordable, practical, safe, or even desirable. And the technological and logistical challenges are daunting to say the least.
To start with, it takes the better part of a decade to design, build, and test-fly a conventional commercial plane. Neither Boeing nor Airbus has a new aircraft platform currently under development, let alone one flyable by remote control.
Not only that, but pilotless planes would require a gigantic — and gigantically expensive — rebuilding of pretty much the entire civil aviation infrastructure, from a totally new air traffic control concept to the redesign of airports. How many hundreds of billions would that cost, globally? And that’s after developing a fleet of tens of thousands of aircraft that are safe and reliable enough for automomous operations. And you’d still need pilots to operate these aircraft from afar!
It’s not impossible, but it would be a hugely more formidable task than people are led to think.
As I write in the Times, though, it’s not the future that concerns me so much as the present. I’ll be long retired and probably long dead before anybody is zipping around on pilotless planes. What riles me up is the simple wrong-ness of so much of what people think they know about a profession that in fact they know very little about, how simple they think this concept would be to develop, and the smugness with which they pronounce pilots as all but obsolete. It really insults me. And the public, for its part, deserves an accurate understanding of how planes fly, and of what pilots actually do for a living.
SEE THE NEW YORK TIMES STORY HERE.
New York Times illustration by Alvaro Dominguez.
Patrick Smith is the author of COCKPIT CONFIDENTIAL.
Leave a Comment
Maximum 1500 characters. Watch your spelling and grammar. Poorly written posts will be deleted!
115 Responses to “Pilotless Planes? Not So Fast.”
You are viewing newest comments first. Click to reverse order
It’s not a great time for a young aviator to hope that they will have a long career in flying airliners or heavy corporate jets. I give it 20 years for those jobs to disappear along with ATC which will be a virtually ghost town as it will also become self sufficient. This has to happen to save the airlines money. The drones are living proof and also low or no visibility will mean nothing as the people operating these planes from the ground will be able to see the runway as a clear VFR day no matter what the weather. It will just be like a video game to them. A night landing is 300 times more dangerous than a daytime landing, but that will end also. Our advancement in this field as is automobiles and cell phones and any electronics with GPS and on and on is unlimited and moving at an unbelievable pace. It just might be the time to get back into flying small aircraft and helicopters if you really want to be a so called pilot. I have been flying a Gulfstream V as an SIC and from what I see is these guys can fly 6000 miles and actually only fly the plane for about 4 minutes so what is the fun in that. Pilots are just collecting the pay and walk around with their chests out but come on other than an occasional difficult cross wind landing where are the flying skills anymore.I have 17,000 hours in small jets, sabreliners, lear jets, citation jets, Bell helicopters and all cessna a and 400 series, and I had to fly all those planes and really was an aviator.
Did you actually read the article that I wrote?
You state that you were a Gulfstream pilot and have 17,000 hours. I really HOPE that’s not true, because this crap about how you “only fly the plane for four minutes”…. no pilot with an ounce of self-respect would type something so misleading, throwing his profession under the bus and giving people a completely distorted idea of what “flying” actually entails.
Meanwhile, for pilotless planes to become a reality would require a complete reinvention of the entire aviation infrastructure, from airports to ATC. That alone would equate to tens if not hundreds of billions of dollars and would take a LONG TIME.
Not to mention perfecting a technology that effectively doesn’t exist yet. Right now, it takes almost ten years to develop a “normal” jetliner. How long would it take to develop a reliable, equally safe pilotless one, and THEN work that plane into that currently nonexistent infrastructure?
And don’t start going on about drones? If you are going to equate drones or military UAVs to commercial aircraft, then it’s clear you have NO idea what you’re talking about and this entire conversation is pointless.
It looks like the next generation of Boeing airliners will be pilotless. Does Boeing know anything about flying a commercial airliner? Machine learning is disrupting the world. 50% of the jobs that exist now won’t exist in 20 years. Sorry, but that includes pilots.
I agree Steve, but I don’t think we have that much time left. The other day I saw Avianca make two go arounds in Medellin Colombia on a crystal clear day on a long runway with passengers on board. Other than pilot training I have only made one go around in my 50 year career of 17,000 hours so it just goes to show you the pilots especially in those third world countries are weak at best. Where we require a minimum of 1500 for SIC they put them up there with 250 hours. Come on man that’s dangerous crap.
Okay, I’m calling bullshit.
I have somewhere around 15,000 hours, over 13,000 of them with commercial airlines, and I’ve made more go-arounds than I can remember. Most of my colleagues would say the same. Once in a Beech-99, and once in a Boeing 757, we were forced to make TWO go-arounds on the same approach (both because of ATC spacing)!
And your remark about “third world airlines” is obnoxious. Avianca is one of the oldest airlines in the world (older than any U.S. carrier) and its pilots are highly skilled.
You make it sound as though you do not know how to program VNAV in your FMS. Fix ALT/SPD deletes everything you described in your descent. The FMS even advises if extra drag is required. I’m a pilot as well and can tell you all you’ve done here is further the case for automation. Piloting a plane is a relatively simple job to automate in all eventualities and has already been demonstrated. See the president of Boeing’s comments on pilotless flight- it is a matter of time. With more and more flights and less airspace to operate with, removing human delay and imprecision will become a necessity to operate in our nations airspace. If you don’t believe me ask yourself why you aren’t hand flying all the time; the computer is far more precise than a human can possibly be in every scenario. I would love to see pilots remain in the cockpit, but unfortunately the only real obstacle to removing them is public perception. It’s not quiet as hard as establishing an interplanetary colony.
On come on, Charles. If you’re really a pilot you’d have to know how silly this comment is.
For one, you’re vastly oversimplifying what “flying” actually entails. We’re not talking about the mere STEERING of the airplane. Of course the basic skeleton of a flight, including turns, climbs, and descents, could be automated. But what good is that? No flight has EVER played out in the manner of some carefully controlled, perfect-scenario laboratory experiment. Flying is very ORGANIC, for lack of a better term, and there are so many moving parts, so to speak, that affect the course of a typical flight, requiring countless subjective inputs.
And you haven’t even touched on the immense infrastructural issues: air traffic control, airports, etc.
Yeah yeah organic and fluid or any other adjectives you can conjur to romanticize flying. I get it. I love it too. However we’re already seeing ATC data link upgrades that will help eliminate read back hear back issues and allow for direct input from ATC, the biggest issue for CAT3c landings are emergency crews ability to assist in the event they’re needed on the ground. You sounded like a mid 1970s flight engineer when you talked about all of the moving parts, and where are the flight engineers now? They went the way of the navigator and radio operator. All I can really say is let’s meet back in 25 years and see how things look. My bet is that there will be far fewer of us on payroll with many more flights per day. Again, I want to agree with you but the heads of the major aircraft manufactures are already in preparation for the inevitable.
“…Yeah yeah organic and fluid or any other adjectives you can conjur to romanticize flying…”
It’s not to romanticize anything. I chose that word because it’s the best one I could find to underscore how complex and changing every flight is.
“…You sounded like a mid 1970s flight engineer when you talked about all of the moving parts, and where are the flight engineers now?…”
Oh my gawd, Charles. That’s the most irritating comparison I’ve ever heard: the idea that the elimination of the flight engineer is somehow analogous to the elimination of pilots altogether. Gimme a break. And when I said “moving parts,” I obviously didn’t mean it in the literal sense. I was talking about all the various factors that affect how a flight is operated, from gate to gate.
“…All I can really say is let’s meet back in 25 years and see how things look…”
Twenty-five years? Nah. Let’s make it forty. Planes under development right now, like the 777-X, etc., are forecasting out at least 25 years with no significant changes to design or operation. And we haven’t even STARTED working on the ATC and airport infrastructures.
Interesting article. People can certainly underestimate the amount of work involved, so a simpler measure that anyone can understand might be amount of time spent on average operating vs monitoring the equipment.
For the purposes of this, a pilot isn’t the proper expert here. It would be an engineer who programs fly by wire systems, an AI expert, or something similar. I’m a programmer personally and while AI is not my field of expertise, I do follow its progress. I’m no expert, but hopefully i can give some insight. the reality is somewhere in-between what you’re saying and what some others have said.
Abbreviated thoughts because of limit:
*practical implementation is probably 20-40 years ahead.
*primary hurdle is software.
*we should analyze best current tech as a starting point eg best autopilot, autoland, etc.
*infrastructure upgrade would include one real pilot for a number of planes as a backup. The changeover would be gradual.
*benefits are safer flying, better fuel conservation, smoother flying, less stress on plane, more flights, and overall great cost savings that would be passed on to the consumer.
Hi Nathan. Thanks for the reply. Let me take these points one at a time, and see what we come up with…
“…a simpler measure that anyone can understand might be amount of time spent on average operating vs monitoring the equipment…”
I don’t think so. For one, what does “operating” mean? You are constantly making changes and inputs to various systems, even during the quietest portions of a flight. The volume of these inputs fluctuates, and it can be impossible to predict when, or for what reasons, things suddenly become very busy: a routing change, a weather issue, a mechanical issue, a passenger issue. We were flying to Europe the other night. All was quiet and boring when the controllers suddenly assigned us a “track change” right before we were about to enter oceanic airspace (change of routing and altitude, etc. for the crossing). We also had some storms to navigate around. For a solid 35 minutes we barely had a second of free time. The autopilot was engaged all the while, but there was a huge amount of stuff to input, revise, reorganize, communicate and calculate. That’s just one example. There are hundreds of things that can, and do, alter the rhythm.
“…a pilot isn’t the proper expert here. It would be an engineer who programs fly by wire systems, an AI expert, or something similar…”
That’s like saying you could automate an organ transplant without needing the expertise of a doctor. Forgive me, but here is where the hubris of people who fetishize technology really shows. You CANNOT be integral to this conversation without a strong understanding of the existing operational realities of commercial aviation. Dovetailing with this is that the fact that people outside the industry don’t have an accurate STARTING point. People assume that flying is ALREADY a lot more automated than it really is. The baseline is way off. You’d be amazed, maybe, at just how low-tech so much of what we do is — by how much of it is, in essence, more art than science, often because it NEEDS to be. What’s the best thing to do, the right thing to do, in a given situation? It depends. It’s subjective. And I’m talking NORMAL operations, not unusual circumstances or emergencies.
“…practical implementation is probably 20-40 years ahead…”
I don’t know. It takes a solid decade to build and develop a “normal” jetliner. Current fleets and aircraft orders project through to around 50 years from now, with no drastic changes. Then you’ve got the infrastructure to figure out.
“…primary hurdle is software. A is slowest progressing field of computer science. sensor tech already exists & getting cheaper…”
Just as important would be how to reorganize and it some cases totally redesign the commercial aviation infrastructure: airplane manufacturing and testing, air traffic control, airline training and operations (pilots, cabin crew, dispatch, meteorology), security issues, and so on. And you’d still need pilots to operate these unmanned planes from afar.
Everybody is an expert when limitless information is available via the internet. Unfortunately, with access to limitless information comes boundless ignorance and there is plenty of ignorance in some of the comments below.
James C…your facts are wrong. The US military is NOT currently using pilotless aircraft to carry passengers. Also, according to Boeing 80% of aircraft accidents are caused by pilot error…not 90%.
Source: 24 years of aviation experience in the US Air Force.
Excellent points; I agree. We need pilots.
Great article. I was having a discussion with someone in a coffee shop about this just recently.
I was saying the same as you, pilotless (commercial) aircraft are way off in the future, if ever. I’ll email him this article.
However, military drones are piloted remotely of course. What are the main operating differences between flying a drone remotely and commercial aircraft?
Thanks,
Gavin.
What absolute dribble. Military applications are already utilising unnamed passager carrying aircraft in conflict zones. It’s only a matter of time before this happens in large scale commercial aviation and humans will no longer be required in the cockpit.
I would much rather get into a plane flow solely by a computer. Much safer overall when human error is removed – considering over 90% of all aircraft accidents/incidents are pilot error – plus the airlines will be able to accommodate more seating, meaning cheaper flights.
It’s amazing to me how people with ZERO commercial aviation expertise are so vehemently confident when it comes to this topic. Do you affect this same I-know-more-than-you attitude with your surgeon? Why, when it comes to flying, is EVERYBODY an expert except for the people, like me, who’ve spent their entire adult lives in the field? It’s almost sort of creepy that people can become so convinced of things when they have no real knowledge of, or background in the subject.
I’m sorry, but you are WRONG. You’re just WRONG, and if your best point of reference are military UAVs, you’ve underscored my point without even realizing it. Come for a ride in a jetliner cockpit some time.
How’s that self-driving car concept coming along?
And by the way, the word is drivel, not “dribble.”
James, you have no idea what you’re talking about.
Also, it’s DRIVEL, not “dribble.”
As to why that’s especially funny here, you need to see the move “Burn After Reading.”
I’m student who’s studying airline and airport management. I’m doing an assignment based “whats the public’s opinion on pilotless planes?”.
I want to get some of your opinions and thoughts on pilotless planes, what are your thoughts on a plane being flown by unmanned?
Thank you for your cooperation
Hi Denzel pick another career there will not be a job for you in the future as the pilotless planes will doom your chances of actually flying a plane. Sad but true. Already it’s only a video game at best. Good luck Denzel.
Only a video game? Really? I’m gonna remember that comment the next time I’m in in the simulator, or the next time I’m flying a complicated departure out of Bogota, or an arrival into SFO, or the RNAV to 13L at Kennedy with a 25-knot crosswind…. or a thousand other examples. If you really are a pilot, you ought to be ashamed for giving people such a misleading idea of what flying actually entails.
It’s exactly articles like that one (it’s from 2014 by the way) that give this baloney so much traction among the public. Well-intentioned authors who know nothing about commercial aviation eagerly believing whatever they are told, and relying on sources they shouldn’t be relying on. The fact that Missy Cummings, a pilot-hater whose statements to the media range from mildly misleading to outright bullshit (i.e. on the Jon Stewart show), was the author’s go-to source only amps up the story’s fantasy factor.
You pilots just need to get used to the idea that your days in the cockpit are numbered. Every time you post about the complicated stuff you do I just roll my eyes because its ALL stuff that a computer can do much more efficiently and safely. YES, even the emergency stuff. Can the car I own now drive itself? NO. Will the next one I own drive itself? YES. Same with your plane if somebody builds it.
My plan was to let this one go. I didn’t feel like dealing with the frustration.
But then, only a couple of hours ago, I got back from one of those insanely work-intensive flights, where my voice was hoarse and I was just exhausted. It was a flight to Mexico City and, in many ways, just so typical. There was weather to deal with — storms, icing, crosswinds on landing — complicated approach and departure patterns that ATC kept changing on us, two sick passengers, and so on. All of this crap changing by the second, and all of it requiring lots of very subjective decision-making on our part.
Afterwards, I decided that not responding to this comment would leave me more frustrated than simply ignoring it. Because I feel offended, for one. I don’t know what it is that you do for a living, but obviously you’re not a pilot and obviously you don’t know very much about commercial flying. And this is common: people with no expertise on the topic find it perfectly within their bounds to argue with a pilot about how planes are flown. And at the same time they’ll accuse me of being the arrogant one! Where does this audacity come from? Would you walk into an operating room and start arguing with a brain surgeon about how brain surgery is really done, and how soon computers will replace doctors? You have no place in this conversation without at least a reasonably professional understanding of how commercial planes are flown.
And don’t give me the Luddite argument. It’s not that I’m insecure or worried, as you say, that as a pilot my “days are numbered.” Please. It’s not a threat to me. Like I say in the post above, I’ll be long retired and probably long dead before such a thing is reality on any measurable scale. I personally don’t care if eventually it happens. And I’ve never claimed that it’s impossible. I’m just saying it’d be a much, much, much more substantial undertaking than people are led to believe.
One part of the problem, I think, is that you assume that flying is a very objective, step-by-step series of events that always unfolds in perfect order, and that any potential complication can be solved through pure digital logic. In reality the typical commercial flight is a highly organic event, subject to hundreds or even thousands of human decisions — many of them small, some of them big, and all of them important one way or the other. Dealing with all of this stuff from a room somewhere thousands of miles away… are you out of your mind? I cannot begin to tick off the potential complications — and disasters — that the pilotless plane concept invites.
You’ll be flying on a pilotless plan around the same time that you’ll be getting a doctorless kidney transplant.
And we haven’t even touched yet on the logistical and cost challenges of essentially having to rebuild the entire commercial aviation infrastructure from scratch, from our airports to air traffic control. And for what? So that the airlines could save money by having their pilots work out of remotely located control centers instead of being on board?
Patrick, you should point your readers to any of the thousands of airliner cockpit videos on YouTube…many of them illustrate the complexity of flying even a highly-automated airplane.
As a bonus, they often show the first officer as the PF…I know the misconception that FOs are “assistants” is another pet peeve of yours!
Read the latest Autoweek magazine, Nov. 9, 2015. JAY LENO wrote an article, “Autonomous Autos: They Might Save Lives”. I immediately thought of Patrick when I read these words, “The car proceeds by itself, but you’re there in case something bad happens. It’s like flying a jet. Mostly the pilot just sits there- the flying is done for him. He might be involved in takeoff and landing, but nearly everything else is automated.”
As soon as people start getting into driverless cars and stop crashing, and then stop worrying about crashing, and then stop worrying about everything they’ve ever worried about in a car except for how to shut the kids up, they’ll soon start worrying about the fact that there’s a human being in charge of their life whilst they’re in the air and won’t be very happy about it. The genie is half way out of the bottle already.
And to follow up on that – have all the engines fallen off? No problem, calculate (in perhaps 100 milliseconds) a glide path to the nearest suitable airstrip, and implement that plan immediately, whilst simultaneously sending the information to ATC. Automate ATC as well, to automatically clear the skies and allow the plane to land. No suitable airstrip? Find a field and, with no stress as a hinderance, land in it as carefully as possible, all the while broadcasting data to a server farm to run millions of simultaneous simulations in order to facilitate the best possible outcome for all concerned, retrieving the information back to put the best plan into action.
There are things pilots can never do. There are things pilots can do that they shouldn’t be able to do. Separate the flight control systems from the entertainment systems, use AES/SHA2 encryption and signatures to enforce the use of the correct parts in all parts of the plane to ensure the Lajos incident can’t happen again… the possibilities are limitless and probably only 15 years away.
Guy (May 18, 2015 at 11:02pm) makes some great points, but misses out on why computers will take over from pilots in reasonably short order. Many people have commented to say that the Qantas A380 which was safely landed back in Singapore wouldn’t have been landed remotely. Well, that’s true, because a remote pilot would have latency issues down to the finite speed of light which means remote pilots will never take control of commercial jets. However, whilst a pilot is reliant on space and their eyes for the amount of data that can be provided to him or her at any one time, computers don’t have this limitation. A computer can monitor altitude, speed, bearing, weather conditions, turbulence, engine thrust, RPM, fuel level and (literally) millions of other parameters all the time, without stopping for a sip of coffee. Instead of, what – 30, at best, for a pilot? Furthermore, there’s no limitation on the number of sensors you can have for each particular measurement. You don’t need to ask the air traffic controller what the air pressure is – you can measure it, and you can also receive it via satellite data. You can use LIDAR to keep a constant picture of everything around you and use some reasonably simple image analysis to work out what’s what and where it is, even if things are invisible to the naked eye.
Often I look for analogies to better understand a situation. In this case it proved easy. Consider another highly automated work environment: the control room of a nuclear power plant.
There are many similarities. The engineers at the plant don’t turn wheels to control the coolant flow, nor to vent steam, nor to regulate the speed of the generators, etc. Computers do all that, under the direction (not supervision) of several engineers. Usually the work proceeds normally, but there can be emergencies which, if not handled appropriately, can become life or death situations.
And the same goes for any other kind of power plant, even hydroelectric ones. And the operations management of power grids, subway systems, and several other types of jobs.
Do we need pilots in airplanes? Do we need engineers in nuclear plants? Do we even need to ask the question?
Replacing Humans as pilots isn’t really the way to go, rather they should supplement them. I mean lets think about unmanned military craft, how is it to tell the difference of a combatant and noncombatant on the ground? A pilot might be able to realize that the large group of people are civilians while an UCAV wouldn’t hesitate to bomb them as it was ordered to attack a grouping of enemy forces.
The author doesn’t know enough about computers. He has made the counterpoint without knowing it. Sure, a computer cant fly a 767 or Curtiss jenny out of the box. You need a new plane, or significantly modified one to make this work, just like how driverless cars are significantly modified.
In reality, all the procedural steps he details for his 767 maneuver are exactly what computers are good at, and unlike in your average car, in a relatively modern aircraft there is enough telemetry already built in to support robust logic driven operations. Also, the complexity in general of flying is much less than driving. Once airborne, you’re looking at maintaining altitude, airspeed and bearing. No children jumping out at you, complicated directions to follow, motorcycles cutting you off… So much easier for a computer to fly an airplane safely compared to driving a car safely.
Flying commercial airliners is about procedure and reacting to inputs in a controlled manner. This is cake for a computer.
Given the risks associated with human pilots deciding to fly in to mountains or ignoring their instruments, it is overdue.
http://www.planecrashinfo.com/cause.htm
Pilots that cling to their importance are in for a shock.
Flying an aeroplane is not a piece of cake for computers.
Modern PRIMary flight control computers have limited capability and are not as capable as you might expect. Triple redundant PRIMs must have slower bus and processor speeds to not overload the backbone bandwidth (for fail-operations transition to another PRIM).
Certified aviation software is always years behind the civil community. You can’t stop an aeroplane in flight! Data buses, security and reliability is a problem much greater for aircraft than office computers.
Garbage in – garbage out: Air Data Computers were unable to resolve the faulty inputs on QF72 — the FCCs sent the aircraft from +1 to -0.8g in 0.5 seconds. Multiple systems and checklists were faulty on QF32 due to sensor (parameter) failures.
The autopilots on Boeing and Airbus aircraft have less authority than the pilots, and can only be used for relatively smooth flight.
The current generation of serial computers might have reached their limit of complexity, a barrier where introducing more features introduces an exponential increase in potential failures and thus lower reliability. This might be acceptable for land based computers, but not for aircraft computers.
Conclusion: We will see pilotless passenger commercial aircraft only when we can build parallel computers that can replicate human thought, awareness, consciousness and prediction.
http://qf32.aero/2015/08/25/aviation-pathways/#PilotLessAircraft
Thank you so much for your comment and link shared. I was doing some research on the future of artificial intelligence / truer aviation automation (I’m a rotorcraft pilot, most of what we do isn’t ‘that’ automated, was interesting learning about fixed wing guys opinions). BUT that link to the QF32 pilot was really really enlightening. He had some great points and I appreciated your sharing it. I obviously like askthepilot.com as well, read a handful of articles here, but wanted to thank you for your input as well.
I’ve always used this example with my friends: when you’re driving in the fast lane at 65+ on Cruise Control, are you still driving? Of course you are. autopilot is Cruise Control in 3-4 axis.
“To be clear, I’m not arguing the technological impossibility of a pilotless plane. Certainly we have the capability. Just as we have the capability to be living in domed cities on Mars.”
There simply isn’t any evidence of any kind that we have the current capability of living in domed cities on Mars. The very few experiments in living in domed environment here on Earth have not been particularly successful. About the closest we’ve come is the Biosphere 2 project in Oracle, Arizona. There were only two missions completed and both had significant problems and the first one was the only one to run its full course. It is only about 1-½ acres and sustained only 8 people. Hardly a city. I know families in one house large than 8. This was all on Earth. We have not yet demonstrated the ability to land and construct anything on Mars, let alone something as complex as Biosphere 2, let alone a full city.
On the other hand, I can walk over to my local Walmart and a buy remote-controlled airplane. It isn’t a passenger plane, but it is a plane and it can be flown remotely. We don’t need to build a fully-functioning remote-controlled 787 to know the technology that allows for for such a plane is, if not fully available, well within site. I agree there is likely not much in the way of economic incentive to build a remote controlled 787, but it is much, much more feasible than a domed city on Mars.
A few thoughts:
First, there is a recurring thread which runs “Can a novice safely land an airliner?” Most folks agree that it would be a miracle.
This in itself ignores all that the flight crew do to get the aircraft in the air in the first place; inspections, pushback, taxiing, dealing with the local tower and departure control, and settling into normal flight.
In crowded commercial airspace, I’d rather have a human up there. Right now.
We now have driverless cars, and they seem to work OK. Recent tests involving long distance and unexpected situations have been hopeful, although I don’t believe any state has allowed them without a human emergency pilot.
I’m not sure how the car and the airliner stack up; certainly the airliner is larger and much more complex mechanism, but much of that complexity is the sort of thing that computer control systems are designed to do. Three dimensions vs. two dimensions poses some difficulties, but they’re not insurmountable. And there is always the issue of numbers; 3 lives at stake vs. 400 lives.
As the driverless car goes, so goes the pilotless plane.
I suspect your job and lifestyle are safe for your lifetime, but keep your eyes on driverless cars.
Pilotless airplanes — even if this were somehow deemed desirable (for the purpose of saving a coupla salaries) — would, as Patrick has pointed out, require a huge and costly overhaul of the entire System.
Driverless cars would require relatively little overhaul, maybe next to none.
People said that about the automobile. “you can find a blacksmith or a veterinarian in every town, but who knows how to fix a car”?
One of the massive differences which is impossible to escape from: cars have a default fail-safe, which is to stop. Planes simply do not have that option. Given that the main problem with switching to pilotless/driverless is what to do when things go wrong, this is probably the vital technical hurdle (if we ignore all the other problems of cost and so on).
Patrick,
According to your article you take umbrage (in part) with John Cassidy on the New Yorker website for stating, “In some ways, human pilots have become systems managers,”
Later you state that:
..they (pilots) are manipulating, operating, and commanding the various systems and subsystems that carry you to your destination.
Again, you give a somewhat detailed account of the operations required to take a 767 from 25000ft to 7000ft.
Your statements show the pilot is managing, controlling and operating the automated flight control systmes on the plane … You are a system manager. Not just in an abstract sense (everything is a system), but in the modern technological sense — You are interfacing to a computer system
that is running various system controllers to perform the tasks you tell it to do — in a controlled and stable fashion (most of the time). The computers have relieved the pilot of some duties.
These tasks require a vast amount of knowledge and experience about the flight systems — how they operate and what they can and can’t do. I don’t think the term “System Manager” takes away from work you are doing. I think it wonderfully illustrates the complex environment you are working in.
I have been reading and enjoying your work for years and think this semantic horse is dead. The “planes almost fly themselves” just equates to (or is a shorthand for) the “cockpit and avionics have been highly automated.”
Thanks!
I understand your point, Sean, but the thing is, while you seem to have a good understand of the topic, and of what I’m trying to say, the average person does not. They interpret “systems manager” a very different way. I’m also a little uncomfortable with the description of the pilot “interfacing with a computer system.” Although aspects of the flight are under guidance of a computer, it’s not a computer in the sense that most people think. Like I say in the story, “computer” is a term that pilots seldom use when describing cockpit components. We don’t think of the equipment that way. “Computer system” implies hands-off, but so much of what we do is hands-on. Switches, dials, speedbrake and flap handles… it’s a pretty manual operation, even with the automation on. The right language is, to me, important in giving people the correct sense of how it all works.
…. Garrison is an experienced pilot ….
Doesn’t mean he ever learned other than to set the aeroplane in motion and to react to whatever it next does.
Just like 85% of those I’ve flown with these past fifty-odd years.
Brian Richard Allen
Garrison has both designed and built two different aircraft from scratch, one of which he flew from California to Japan, like, 40+ years ago when it was basically dead reckoning over the Pacific. He has done a lot of different types of flying and possesses an impressive store of aerodynamic and aero-historical knowledge.
Which doesn’t mean that he knows a lot about the practical business of flying airliners. More’s the pity.
Patrick – – Always a pleasure to read you. Auto-Pilot is the same as Auto-Boiler(s) – which I used to operate. Otto runs them after we’re ‘on-line,’ but you must tell Otto what you want. Set points for pressure, flow, temperature, etc., and at the turbine, megawatts. At the Generator – voltage (the VAR schedule, I’m sure you recall). What? The frequency dives below 59.6?? AGC has tripped? Now you’re on ‘hand?’ Watch your levels, excess air and make ’em right!! Hours and hours of tedious boredom with moments of sheer – – – you’re heard that one. Step on the ball, Patrick – – if they still have those things. Best – C. Bryan
Patrick, I’m still one of your biggest fans, despite long-time silence, except for my recent defending you at FlyersRights.org, where I’ve been a major contributor and I proofread the weekly Newsletter.
Pilots and passengers have many common interests. Airlines have an interest in pitting us against each other, but we have much more in common. Mostly, what benefits one of us also helps the other.
So I suggest all of Patrick’s fans go to FlyersRights.org and join up, and read the weekly Newsletter. It costs nothing to join, and is member-supported. FR is not political, because people of all political stripes have to fly, and should have a voice.
Give it a try.
Dan Prall
It’s obvious that flying—and society—would be safer with an emergency backup pilot resource on the ground, for use in dire emergencies. You cite having two pilots in the cockpit as necessary redundancy. Why then is it not a good idea to extend redundancy so that a plane can be landed safely in the next Germanwings, or Egyptair, or Malaysia 370, or 9/11 situation?
Germanwings could have been much worse. We had a homicidal pilot in total control of an Airbus over Europe with 1000 miles of range. He could have destroyed any landmark or nuclear facility in Europe. He could have destroyed the Eiffel Tower, where an average of 25,000 people go every day. Do you really think existing or potential pilots with bad intent haven’t realized this?
Keep two pilots in the cockpit. But for the rare moments when help is needed from the ground it’s disingenuous to deny that we as a society have the capability and the responsibility to provide that help. 9/11 cost $3.3 trillion and two wars, crippled the airline industry and hobbled the world economy. Boeing says we will need 500,000 new pilots in the next 20 years. Can you guarantee that none of those new pilots, from all over the globe, will be a dedicated, quiet, intelligent, ATP-certified terrorist?
It’s not a good idea because that emergency backup control on the ground would override the pilots in the plane. So then the bad guys just take over the ground control station or hack into the communications link, and hey presto, they have control. And how can you ensure none of the gorund controllers would be dedicated, intelligent bad guys?
Ahhh, UNLESS you have an emergency-back-up BACK-UP control in a small drone that flies in tight formation with the airliner. That way if both pilots drop dead of heart attacks simultaneously AND the guy on the ground is a terrorist, you’re still OK.
Know what? That guy operating the crane on the construction site you’re passing in a bus could be a terrorist and about to drop 10 tons of steel-reinforced concrete on your bus.
Everybody needs to calm down about “terrorists” a little.
We have used securely controlled, encrypted satellite communications for many decades. That’s how Air Force nuclear bombers and Navy missile subs have been commanded since the early cold war. We can certainly protect secure locations on the ground better than we can protect isolated airplane cockpits that crash into the ground if control is lost.
A simple fail-safe system that would require independent second-person authorization and passwords before a rare control takeover would solve your problem with the idea.
By your logic, we should not have nuclear missiles or submarines controlled from the Pentagon, because somebody could just break into the Pentagon and order an attack. Or hack into the secure communications. Unlike rogue or incapacitated pilots—unlike Germanwings or 9/11—that has never, ever happened. If terrorists can take over secure nuclear facilities and launch attacks at will, we have much bigger problems than air crashes.
You believe in safety and redundancy, it seems, so long as that redundancy stops at your cockpit door.
Which scenario is more likely:
1. An armed and trained terrorist group takes over a securely hardened ground station, without setting off system alarms, navigates the fail-safe authorization and password system, and has the know-how to control an airplane from there, regardless of redundant security measures.
Or 2. Both pilots are slipped mickeys in their Cinnabons by a low-tech cell of airport restaurant workers.
No, we should not have nuclear weapons and other weapons of mass destruction because they are appalling and evil.
I think single pilot airplanes are closer to reality than pilotless airplanes. The Germanwings crash might have actually delayed that debate for a while longer.
ONE of the reasons for having two pilots is redundancy. Even if the workload is reduced enough for one pilot to comfortably do everything on a flight (and that includes emergency situations, not just normal flights), you still cannot lose that level of redundancy. The probability of “failure” of a pilot (loses conciousness, etc) is too high to consider dropping to a single pilot (this is because acceptable probabilities of failures of safety critical systems on an aircraft – even redundant ones – is VERY VERY low). You are technically right that we’re probably closER to single pilot flights, but we are still nowhere near it, and there should be no realistic discussion about doing so.
For all the weird solutions to problems that only exist in the imagination, the most important purpose to having a properly trained, experienced flight crew on the flight deck is that the Captain and First Officer both want to get home in one piece. That means when abnormal or emergency circumstances occur, the fight crew will use whatever training and experience they have to get the aircraft on the ground in one piece. Anyone who happens to be sitting behind them will get home too.
Obviously there have been suicidal pilots. But given the number of departures worldwide since the beginning of commercial aviation to the number of criminal acts, I’ll take my chances with a CA and an FO.
Good points!
You can check out FlightAware and listen to ATC Live to figure out how much is actually going on during a flight. Very often you will see a planned flight plan being changed due to reports from other pilots who ran into heavy turbulence, thunder storms etc. By the time they get to the airport area they may be put on a holding pattern due to various reasons.
I cant imagine how they are going to work that out if the planes are remotely controlled. And this was just a couple of examples of what might happen during a flight.
What baffles me are the fact that other commercial pilots are advocating for this solution. I must have missed something?
Any piece aviation software can get a plane back on the ground.
Getting it down in ONE piece with NO injuries is the trick!
That’s going to be a long time coming!
Just look at a similar situation, the train crash in Lac Megantic,Quebec. Some one who didn’t know what they were doing, fiddled with the controls, and quite a few people died because there was NO engineer(pilot) on board to deal with the situation when things started going awry.
I too am a pilot and have been a follower of yours for many years. I find the media loves hyperbole and usually gets the facts wrong since they are in a rush to throw out ‘theories’ to hold the public interest as long as possible.
However I have to disagree with you regarding automation and the act of the pilot flying. I fact, automation (computerization) flies commercial planes from shortly after take off to shorty before landing. Yes the pilots do input commands into the FMS as needed but that is a far cry from the art of flying as demonstrated by Asiana 214 and Air France 447. Both times the pilots were forced by circumstances to hand fly the aircraft and they could not perform basic airmanship skills. Both crews were efficient computer managers but were not able to demonstrate piloting skills when called upon.
Personally I mourn the loss of the art but I do recognize it has made commercial flying much safer
What sort of investment would be required of the industry to maintain stick-and-rudder skills? Not a lot, I would think. Certainly not the price of even one AF447.
The sidestick on the Airbus seems to me to be an entirely arbitrary move. Its Airbus’s way of thumbing its nose at hands-on flying. Also a rather French fascination with abstraction generally.
It’s also — given the fact that the pilot-not-flying can’t see what the pilot-flying is doing — stoooopid as well.
I think that Jeff Latten got it pretty much right. As Arthur C. Clarke said, “Any sufficiently advanced technology is indistinguishable from magic.” Unfortunately, it appears that cockpit automation, or pretty much anything involving computers, is “sufficiently advanced” for most people.
As a software developer, I struggle every day with managers and clients who somehow seem to think that, since we’re about to use a computer to do something, we can throw logic and the basic rules about how the world works out the window. My main job is not so much coding as resolving the contradictions and simple silliness in the things that people ask me to make a computer do.
This always reminds me of an old poem about a computer:
I really hate this damn machine;
I wish that they would sell it.
It never does just what I want
But only what I tell it.
NYT: “all the news that fits, we print”
Mark Twain said if you don’t read the newspaper you’re uninformed, if you do read the newspaper you’re misinformed.
And of course now they are talking about hackers can get remote access of the plane (and crash it) by hacking in to the flight controls via the WiFi and/or entertainment system on board. Apparently you don`t even have to be on board the plane to be able to do this.
If you cant scare the passengers by claiming all pilots are mental ticking bombs, there`s gotta be another way to do it, right?
However, this is extremely difficult, if not impossible. The WiFi and entertainment systems I know, are completely separated from the flight controls and can not be hacked into.
To put it this way; if you install a WiFi router in your home and someone hacks into it, they are still not able to start your washer and dryer. They are simply not connected in any way.
A plane’s entertainment and wi-fi are not linked into essential flights systems, and neither are flight controls run by “software” in the sense that this scenario implies. This entire conversation is wildly misleading and just reinforces people’s misunderstandings of how planes actually fly.
Maybe, conceivably, theoretically, somebody could hack into the plane’s FMS (flight management system) and change the routing or something? But the pilots would see this and simply change it back. In a worst case scenario, you could just ignore any suspicious routing or command and fly manually wherever and whenever the heck you want to.
People have become so infatuated with electronics and gadgetry that they no longer understand what is real and what is fiction.
I may have jumped to a conclusion, but I still don`t think this could be possible. And Patrick knows more about the internal infrastructure than I do, so he can clarify.
But, on older planes which did not have WiFi etc when delivered from the factory does not have the WiFi and entertainment system connected in any way to the flight control systems (as described above).
The WiFi is mounted later and is totally separated. You may have seen the little “bump” on top of the fuselage, that is the antenna receiving the satellite signal.
I am not 100% sure about the new planes, like the B787, A380 etc – but it would surprise me a lot, from a security point of view, if those systems were connected in any way.
The way we (my company) set up the satellite communication is never connected to the operational system. And even if it was, the hacker would have to find a way to hack into that system. The way its setup goes like this (to make it simple);
– Location of servers
– Then the teleport(s) which the plane is communicating with
– Then find the actual server that plane is connected to
– Then find the correct IP or MAC address for the device they want to hack
– Then try to get access to that system
Each of these systems have their own set of security rules, if the hacker would even know where to look to begin with.
Now, they could bypass some of these “obstacles” simply by being on board the plane. IF those systems are connected in any way.
THIS should be widely reported by the press – that pilots can always turn off the FMS if in doubt and there is no way to override this override?
There’s not an “off” or “on” for the FMS to begin with. The FMS IN NO WAY controls the airplane.
I don’t have first hand knowlege of this since I’m not an engineer, hacker, or pilot, but the security researcher Chris Roberts states specifically that he was able to get into the avionics network on a new Boeing:
” “We researched further than that,” he told WIRED. “We were within the fuel balancing system and the thrust control system. We watched the packets and data going across the network to see where it was going.”
Eventually, Roberts and his research partner determined that it would take a convoluted set of hacks to seriously subvert an avionics system, but they believed it could be done. He insisted, however, that they did not “mess around with that except on simulation systems.” In simulations, for example, Roberts says they were able to turn the engine controls from cruise to climb, “which definitely had the desired effect on the system—the plane sped up and the nose of the airplane went up.” ”
That may not have been enough to crash the aircraft, but it isn’t nothing. It also indicates that on the newest systems, the avionics and infotainment systems are not properly separated.
http://www.wired.com/2015/04/twitter-plane-chris-roberts-security-reasearch-cold-war/
Also:
http://www.wired.com/2015/04/hackers-commandeer-new-planes-passenger-wi-fi/
Well, here you go:
http://www.cnn.com/2015/04/14/politics/gao-newer-aircraft-vulnerable-to-hacking/index.html
Hard to know exactly what they are going on, all they say is that if there is not a physical barrier between systems then there is theoretically always way to circumvent software firewalls. Not a lot to go on but I’m waiting for TSA to now ban all laptop, tablet, smart watch, camera and all other WiFi enabled devices. They’ll probably end up with a hell of a garage sale from all they ones they confiscate to help fund themselves (along with all that spare change they get to keep . . .)
I guess the folks who write articles about “pilot-less planes” have watched “Top Gun” and “Flyboys” and think “ah, that’s what a pilot is needed for”. Then they think as this kind of work is not needed on a modern airliner, they can get rid of the pilot. They cannot comprehend that a pilot’s job is actually entirely different most of the times but still very complex and absolutely needed.
Today, a bad guy in the cockpit can crash a plane.
Tomorrow, when pilotless planes are controlled completely by software, a bad guy could crash _all_ the planes.
Great stuff, Captain!
My retort for layman who insist that “airplanes these days practically fly themselves, right?” is, “Does your Cruise Control drive your car for you?”
Autopilots are nothing more than a fancy, 3D Cruise Control. They may be able to do their job with uncanny precision, using gobs of ultra-fast processing power. But that’s just it. They can Process, but they can’t THINK.
I’m just writing up an interview for Airways magazine that I had with Captain Richard de Crespigny, the Qantas A380 captain that had the Number 2 engine explode on takeoff, damaging all but 1 system. He and his crew troubleshot over 100 failures before safely landing their crippled, vastly overweight A380 with 100 meters of runway to spare. Of the over 400 passengers and crew aboard, not a single injury.
Try giving that job to HAL 9000.
Pilots will continue to be needed in the cockpit well past our lifetime.
Eric “Cap’n Aux” Auxier
capnaux.com
As soon as I saw the title on the top of the front page of the NYT (online), I knew who had to be the author. Kudos for a great article that explained the logistics to all of us non-pilots.
I love technology as much as anyone, but how can anyone work on a computer for more than a week and not realize that there are as many bugs in software as there are on a plane’s windshield? And that’s for a Mac. With Windows, the buggy software realization can occur after only a few hours. There’s got to be humans in the cockpit to figure out the work around’s go arounds. (WOGA?)
It’s a bit picky, but you say that “Krugman won a Nobel Prize.” That implies that Krugman was competing for the prize which, of course, he was not. As with military medals, such honors are “awarded,” not “won.”
Picky to the point of being plumb wrong. “X won a Nobel Prize” is a common expression. And usage will beat analysis any day.
It may be a very common error, but it’s still an error. And it is a bit of an insult as well. Krugman did not do what he did in order to “win a medal,” he did it for the betterment of mankind. To suggest that he was competing in a contest to “win” a few thousand bucks and a hunk of metal is belittling.
How is this for picky:
Paul Krugman did not win or was not awarded a Nobel Prize in Economics — there is no such thing. Alfred Nobel did not originally dictate there be a prize catagory for economics. Paul Krugman won the Nobel Memorial Prize (in Economic Sciences) that was established by the Bank of Sweden in memorial to Alfred Nobel (many years after the establishment of the Nobel Prizes). This prize is commonly referred to as Nobel Prize in Economics, however, this is an error — or a shorthand.
Thanks for playing.
Great post. It’s been alarming how quickly mood has shifted — before AF447, talk of pilotless planes was very rare and left-field, now one pilot does the unthinkable and the human-free cockpit is being presented as a sober and reasonable thing to aspire to. I wonder if a part of it has been a similar big shift in the way driverless cars have moved into the mainstream. At any rate, it’s amazing how risk-averse people are when it comes to air travel — because of one black-swan event, it’s being proposed that we scrap the entire system, not just technologically but philosophically.
Superb comment, Jeff, thanks for leaving it.
My only dispute is the notion that driverless cars have made it anywhere near the mainstream.
Funny story. I gave up flying a few years ago, for pretty much the same reasons as Eric Holthaus. But I still read Patrick Smith religiously! You’re always a great read and fully deserve your place alongside Brooks and Krugman.
Patrick, your outstanding NYT piece keeps us grounded in all things aloft. BRAVO!
John Cassidy should spend a day watching pilots undergoing training and check rides in an airline simulator.
The main thing is that computers can’t think through problems and never will. I remember a line, maybe from a movie of a pilotless plane, with the computer making an in-flight announcement to passengers “Good morning ladies and gentlemen, we are scheduled to land in Los Angeles at 3PM, and everything is operating normally, operating normally, operating normally, operating normally.”
The tech-assisted doctor analogy is a good one, but perhaps as as esoteric as piloting to the layperson. I think a simpler, more relatable one might be baking.
In your great grandmother’s day, baking was a complicated art – managing everything from fuel type and input rate (especially in a wood-fired appliance) to draft and damper settings, and even judging temps without instruments. Now, you don’t need a chef, right? Input the parameters from the recipe and the oven does the rest! Set the temp and the oven holds it, set the timer and it tells you when to take the dish out. The chef is only “hands on” for 60 seconds of the actual baking time!
That works pretty well for frozen pizzas, or perhaps an industrial-style food service setting – military, school, heck even an airliner! The resulting food is usually safe to eat. Nutrition is conveyed. Hunger satiated.
But would those results meet the standards required in a high-end restaurant? Can a few chefs in a warehouse in Omaha manage a dozen ovens for a group fine dining restaurants nationwide?
Probably not. Modern cooking appliances are great, but the inputs still need to be managed in person. Decisions have to be made based on in situ criteria. The drudgery, uncertainty, and variability have been reduced. Efficiency has increased. Fewer bodies are needed in the kitchen. But the talent and skill still needs to be present for great results. And in passenger aviation, great results = people staying alive.
Good stuff as always from Patrick.
I am also an airline pilot and often get into discussions of how the airplane flies itself. Patrick has correctly pointed out that there is much more to actually controlling the automation than simply pressing a button.
I often use my ship analogy, or at least Hollywood’s version of how a ship’s bridge operates.
When you watch a ship movie, including space ships like the USS Enterprise in Start Trek, do you ever see the captain of the ship actually steering the ship? No, it is usually a junior officer or enlisted man who is actually “steering” the ship. The captain is busy commanding the whole operation of the ship while the helmsman steers. All the autopilot does on a modern jet is to take the job of the helmsman. There is as much if not more to the captains role than there ever has been.
Put another way: The autopilot systems do a lot of Sulu’s work, but little of Captain Kirk’s.
Plus, look what happened when THEY introduced complete automation with M-5 in the episode “The Ultimate Computer”! 😉
I expect to ride on a pilotless commercial airplane right after I get that flying car I’ve been promised for the last 50 years.
So glad for this good, clear post, and your NYT piece is excellent as well. Thanks for being the solo voice of reason, even if it does lead you to many headdesk moments. Computers can do it ALL, right? That’s why Cranky Concierge is going out of business… not.
The thing about Sully is that that was a super-rare scenario, but given that it’s likely his experience in the cockpit (flight hours) gave him the know-how, it’s an argument for MORE qualified pilots up front, not less-experienced pilots in room hundreds or thousands of miles away. The time lag between the ~whump~ of a bird strike and deciding what to do about it could easily be the difference between dramatic crash where everyone survives and dramatic crash where nobody even had situational awareness long enough to do anything.
Routine flight is no different. Sure, you can freeze food and reheat it, but what you get isn’t pretty, sometimes still has frozen spots, and isn’t a joy to eat. We probably could figure out how to bolt on some pilotless system to an already hyper-automated flight deck (I’m looking at Airbus). You’re likely to get a marginally safe, herky-jerky flight with no callouts of dramatic sights and limited deviation to avoid rough air, traffic tangles, or the like. A human pilot handles all these inputs with aplomb, making good split-second decisions that, while minor and invisible to the amateur, provide a much smoother/safer ride. Humans FTW!
There was a piece on CNN tonight;
https://www.youtube.com/watch?v=AmRUTLPeXRY
I liked the following comment;
“…the people (pilots) on board, only there in the event of an emergency…”
Ok, so there should still be pilots on board in case of an emergency?
Whats the point then.
And here’s the matching cnn money article:
http://money.cnn.com/2015/03/27/news/companies/pilotless-passenger-planes/index.html
I’m surprised that none of the non-aviation writers have found this to support their pilots-are-playing-video-games-while-the-computer-is-flying-the-plane flights of fantasy …
Perhaps because purpose-built unmanned air vehicles have become a fact of life in the skies over Afghanistan, the introduction last December of a full-size helicopter, converted into an “optionally piloted” airborne delivery truck, generated little attention. Yet the unmanned Kaman Aerospace K-MAX represents, arguably, not only a greater technical achievement but a bigger step towards a future where commercial aircraft with empty cockpits ply their trade.
Expensive piloting skills are no longer required within Marine Unmanned Aerial Vehicle Squadron 1. Whenever one of the two K-MAX currently attached to its Cargo Resupply component is tasked, an NCO (not pilot-qualified) climbs in and starts it up. He then switches control to another NCO at the nearby main operating base, climbs out and walks away. NCO No 2 presses a button to send the helicopter off on its flight plan. At its destination, a grunt in a foxhole guides it to the landing zone with a games console. When it lands back at base, another operator gets back in and shuts it down.
http://www.flightglobal.com/news/articles/in-focus-k-max-variant-offers-glimpse-of-pilotless-373066/
[Paul Koring] quotes David Learmount [sic], a “veteran aviation expert,” who predicts that “pilots won’t be in cockpits in 15 years but in an airline’s operations room, rather like the U.S. Air Force pilots flying Global Hawks [military drones].” What utter and shameless rubbish.
Also from David Learmont …
… at 06:13 [the pilot] radioed air traffic control saying the right engine had failed. Around 13min later, at 32,000ft and with 100nm (185km) to go to Lajes’ runway, the left engine stopped. The crew prepared the aircraft for ditching, but 20min later, landed the A330 heavily but safely at Lajes.
http://www.flightglobal.com/news/articles/a330-lands-safely-after-gliding-for-20-minutes-135586/
This highlights another problem that must be solved before the pilots get to sit in an industrial park somewhere “flying” passenger aircraft between LAX and ORD — what to do when (not if) the engines quit. A train just stops. A car pulls over to the side of the road. An airplane needs to find a place to land and then execute the landing. The first time. It’s a problem that will be solved but not for a long time.
Another point (which you sort of eluded to), where’s all the wireless bandwidth to actually remotely fly these planes going to come from? The military has more wireless spectrum than anyone and they only fly a few drones at a time. Imagine 30K planes, all transmitting multiple HD full motion video feeds, haptic feedback, and receiving flight control data without excessive latency. Sure, you can run fiber to ground antenna sites, but there’s still a finite amount of spectrum available, and we all want it for our phones.
That said, I do think eventually we’ll see automation advance to the point that a single pilot will become the norm in 10 to 15 years, sooner on cargo flights.
I don’t see how. You can’t have a single-pilot plane until you have a way to land that plane remotely should the lone pilot become incapacitated. That will take a LOT more than just tweaking some software. You’d need a whole new airframe — or a radically redesigned existing one — for starters. It takes the better part of a decade to design, build, test and deliver a plane, and there are none of these even on the drawing board. And then who, exactly, is going to do the remote controlling, and how? That’s another gigantic and expensive challenge fraught with complications.
I think people really don’t understand the time scales involved in aircraft design. I’m currently working on software for the 787-10 – which is just the same as an existing plane but a bit longer. The project goes for years.
As you say, a new aircraft (such as the A380) takes close to a decade (over a decade depending on how you look at it).
The 787 took longer, and that was mostly because of new materials. In fact that took 4 years from roll out – pretty late in the development – to first delivery.
I just can’t fathom how long the design and testing would be for an aircraft intended to be capable of fully automated or remote operated flight (even with one pilot). It might eventually happen – though I can’t see how I’d be convinced to get on on in my lifetime – but the flight testing alone for something like that will probably take a decade or more.
If you look at it from a systems point of view, rather than just a “how do you write a computer program” point of view, no, we’ll never have single-pilot commercial passenger flights.
One of the very, very basic things we know is that humans make errors. If you want things to function reliably, you need systems that catch and mitigate these inevitable errors. And that’s a good part of the reason why you have two pilots on an aircraft: because each can check what the other is doing.
I doubt there’s ever been a commercial passenger flight where one of the pilots hasn’t made an error of some sort. It’s the whole system of checklists, cross-checking, and so on that makes the error nothing to worry about.
In a previous life, I did engineering design on electronic systems of DOD and business aircraft. Redundant systems were always a major design element. Two capable pilots are another redundant system.
Good luck charging at the windmills. When people mention pilotless airplanes, you should raise the issue of driverless subways. Controlling a subway is child’s play compared to an airplane, a fixed route, no weather, preplanned stops, yet it will be decades before we have driverless subways. Why should anyone expect pilotless planes to arrive earlier?
Tom S wrote, ” … it will be decades before we have driverless subways.”
AnsaldoBreda Driverless Metro
Although the trains are not equipped with drivers, there are stewards that help passengers, perform ticket controls and assist in emergency situations.
http://en.wikipedia.org/wiki/AnsaldoBreda_Driverless_Metro
Many airports, SEA included, have driverless subways on dedicated trackways. Computerized, non-operator trains are already in existence and are not that special. They are a far different level of complexity than commercial flight, however. The one in Seattle has been in operation since 1969. http://en.wikipedia.org/wiki/Satellite_Transit_System
The city of Lausanne, Switzerland has a driverless subway system, though hardly extensive. (Possibly other cities do too, for all I know.) As you say, we’re talking an entirely different level of complexity. Still, they are full of passengers who would be vulnerable in the event of an accident.
Driverless trains have been around for a long time now. Vancouver’s Skytrain (currently with 41 stations and about 70 km of track) has been in operation for thirty years now without an accident that a driver could have prevented. (54 people have been killed jumping or falling on to the track in front of a train; they were detected by track sensors but the train simply couldn’t brake in time.) There are at least two other installations of this same technology in Detroit and Scarborough, Ontario, as well.
Driverless trains of this sort are of course a completely different problem from driverless airplanes. For a start, the correct response to an emergency situation in an airplane is not usually, “stop, shut everything down, and wait for staff to come help.”
Nice commentary in the NY Times. I hope this will help quell some of the other so-called “experts”.
Excellent work, Patrick!
Although other have already mentioned, I just want to repeat it;
– Since you need a pilot anyway, why not keep him or her in the cockpit instead of on the ground?
– As Patrick said; the logistic to make this work is an enormous task. How many planes is each pilot on the ground suppose to control?
– As I mentioned in another blog post here; there are approx 50.000 commercial flights every day, operated by the 3 major airline alliances.
Good luck in remote controlling all of that.
– And, no surprise, the main reason for this discussion popping up again is the Germanwings crash. Ok, the pilot was crazy. So a pilot on the ground will never flip out and do the same thing? I`d like to see a guarantee that that will never happen.
In addition to the great work pilots do every day, there are lots of examples where a remote controlled plane would never make a safe landing. The one that stands out most (as I can come up with now anyway), is the Qantas A380 where an engine blew up and they returned to Singapore. You can watch the Air Crash Investigation episode on YouTube, just search for “QF32 titanic in the sky”. I can also recommend the book called QF32 written by the Captain on that flight. When you think of all the procedures and “what if scenarios” they had to go through to make a safe landing, there is absolutely no way a remote pilot would have landed that plane.
WOW, what a story – thank you for commenting it, Eirik. If I knew of it originally, I’d forgotten. To compare & contrast these pilots’ collective efforts, versus MH370 or GermanWings, completely emphasizes why we have, and will continue to have, human beings in charge & on board.
In particular, this Captain’s skill set harkens back to something Patrick’s written about for years: our global, generational loss of commercial pilots with substantial military experience. Trouble-shooting & real time problem solving are so woven into these folks’ DNA!! 650 (whatever) non-military flight hours, not so much.
Hi Patrick,
One thing that immediately occurs to me as a physicist who occasionally writes software that operates hardware and read in data taken by hardware, is that if some how you could actually build a piece of software on top of the hardware running the plane to fly the plane with no pilots, you leave an enormous hole for hackers. A smart terrorist (or murderer) could put his or her own version of parts of the software stack operating the plane. It doesn’t take much imagination to think many horrific possibilities. A similar comment would apply to emergencies too, I just don’t see how any software could deal with emergencies that were not in its playbook of known problems. And don’t tell me it can’t be done, if Airbus and Boeing can get code onto the plane so could a motivated attacker. Look what happened to those centrifuges in the Iranian enrichment program. Somebody made dang sure that software was loaded into them that caused to very bad things to happen to those centrifuges.
I also must say as some who has written code to run devices and read back data, the complexity of writing and debugging (it really is the debugging) seems very, very daunting.
Fred
I guess the main point is: If the pilot will still be a very busy person either way, what possible advantage is there to locating him/her at a remote location? What does that buy anybody? What’s the point?
There are very good reasons for an Air Force drone pilot to be located remotely. The aircraft can be much smaller, lighter, and cheaper, and if shot out of the sky, the pilot gets to go home at the end of the day instead of being injured, captured, or killed.
On a civilian airliner, you’ve gained nothing. You still need the same number of pilots (the workload can get too intense for just one), and you’ve managed to eliminate the cockpit. Big whoop. You’ve now lost a lot of emergency response capability (and $B’s) as your tradeoff.
Maybe one day in the far future we won’t need a pilot for a lot of this work, but that would require across-the-board automation between ATC, airlines and aviation companies. Not happening any time soon.
SirWired asked, “If the pilot will still be a very busy person either way, what possible advantage is there to locating him/her at a remote location?”
How much could they charge for those two great seats up front?
Great idea! “Cockpit Class” seating.
I read with great interest your OpEd piece in The New York Times. While I totally sympathize and agree with all your points, I think you are talking to a wall, so to speak. The public is full of Luddites who believe whatever they hear on the mainstream media, all of it, whether it’s a rightwing rag like FoxNoise, or a liberal outlet like MSNBC. Most of it is drivel spoken mainly by people who’ve never flown anything other than as a passenger, and who have zero understanding of the complexities of flight, whether it’s a local daytime VFR hop in a Cessna 152 or an overnight transoceanic flight in a big ‘un. I think there’s been a minor breakthrough in the use of “First Officer” instead of “co-pilot” but progress comes slowly.
Anyhow, keep saying it…sooner or later it will get through.
Two beautifully thought-out and written pieces.
I wonder whether Peter Garrison will rise to this bait.
I found it especially disappointing that Peter Garrison would perpetuate that myth.
Yes, if only he’d not written that one fatal sentence. Otherwise I found his piece OK.
I’ve always admired Garrison for his sheer aeronautical knowledge, elegant writing style and apparently wide range of interests. Oh well. He’s no spring chicken — maybe age is getting to him.
When you’re descending at .79/315, what’s the N1, typically ?
Actually, it turns out the N1 would be close to Idle Power, and the
Descent Rate between 3000-3500 ft/min :
Typically shooting for a 3500-4000 fpm descent at idle thrust which will give a descent just a few knots shy of 330kts.
If ATC wants 280kts then adjust the descent to around 2800-3000fpm to give the idle descent at 280kts
Or 2000fpm to give an idle descent at 250kts.
But usually will be able to descend from Top Of Descent at idle at 3500-4000fpm then decelerate from 330 to 250kts between 11000 and 10000 ft and continue the descent at 2000fpm at 250kts at idle until its time to start throwing out flaps and gear, etc.
That description of a “simple” descent really made your point. Thanks for continuing to fight the good fight! Now I’m off to read your op-ed.
Nicely done, Patrick. Every time I hear somebody say, “airplanes fly themselves,” my blood boils. What I really want to say is, “OK, if you really think that, on your next flight we’ll put you up on the flight deck alone with the tasty beverage of your choice and you can watch in pure amazement as the airplane ‘flies itself’ to your destination.”