Isn’t that what sort of… happened with the Air France A330? Otto’s probe got frost bite and that caused him to lose his shit…and the two guys up front couldn’t understand why the cockpit was lit up like Christmas tree, with Betty bitching away…?
I wonder if the last thing they heard it called out was "retard " right before it hit the water?
And with Asiana? If I recall the jist of the NTSB report was that the crew ( remember this was check ride so there was supposed to be experienced instructor rated senior training pilots up front watching.) was that not one of the flight crew had a real firm grasp of how the auto throttle and LOC/ APR systems work and interact with each-other in different flight configurations.
What about the elimination of human caused accidents?
It’s a balancing act. Planes are already highly automated and they can auto land if it’s got the equipment and the airport is set up for it. However, as we saw the other day at Laguardia it doesn’t take much for things to go south. It could have been a faulty thrust reverser or the anti-lock brake system but at 140 mph things happen quickly. A decision has to be made to mitigate the damage and in a fraction of a second the wrong decision can get exponentially worse. Sometimes skidding off the runway is better than the alternative if the computer isn’t programmed for the event.
That particular airport has a short runway and the landing has to be decisive. By that I mean the pilot has to plant it on the numbers and decelerate quickly. There’s no allowances for soft landings. Computers don’t understand the reason behind something, just the pre-programmed set of instructions.
And the crew miscalculated the conversion between lbs and liters when they were re-fuelling…
As much as most of us here are probably uncomfortable with the idea of autonomous transport, I suspect we are very likely to see it becoming common place within 20 years because the single biggest obstacle to it isn’t technological, it’s human. Planes already spend most of their flight flying (and landing) autonomously and as noted above, the pilots are there for “fuzzy” situations. Eliminate the pilots and you eliminate roughly 70% of the accident chain, cite.
I would argue that Gen Y and the Homeland Generation by and large view transport as something of a PITA rather than something to do and will have much less of an issue letting go of the controls, so to speak.
Incidentally, semi-autonomous cars are already on the roads and in the showrooms to some extent (Tesla, Mercedes and Volvo to name a few), and the market will only increase.
That’s really the thing. Total amount of pilot-error crashes - total amount of pilot-rescued incidents. Look for the total sum. I wouldn’t be surprised if the balance ends up in favor of automation. But I still think people would be dodgy on it.
I also suspect it’s coming whether we will or won’t. Some discount airline will make some sort of test run. After six months there’ll be no incidents. then tickets will be available cheaper and people will sign up. After the model gets proven - such as it is - one of the bigger carriers will give it a go on some regional hop and it’ll expand from there.
But it won’t be a fast process.
In before LSL guy lays down the law.
Apparently, modern airliners are nowhere remotely close to being fully automated. You know that blizzard of switches and knobs you see if you peek in the cockpit? Contrast that to a car, where there is a steering wheel, gas, brake, and ignition. (and about 4 gauges).
What do you think would be easier to automate?
So no. No fucking way. We’ll have automated cars for decades before automated airliners. Operating one is a complex, skilled task that requires human level judgement.
Sure, differing phases of flight can be done under computer control - but even drone aircraft aren’t really automated, they require extensive human supervision and an actual pilot lands and takes off.
It’s just a well spread myth, like the one about humans using 10% of their brains, that because autopilots can keep the aircraft on course in clear weather, and control some of the control surfaces during a landing, we’re somehow 1 step from not needing pilots at all.
Bzzt.
I go backwards and forwards over this issue. Sometimes I feel like it’s close, other times I don’t. Overall I feel that the major hurdles are not technical. A few points though.
There is a lot of information available about incidents where a human has been blamed for the accident, generally pilot error though in some cases the human is an engineer or it could be a human system such as the training structure of an airline that is at fault. There is very little information about the number of times a human saves the flight though. We known about the Sullenbergers of the world, Sully himself, the crew of the DC10 that lost hydraulic control and came breathtakingly close to saving everybody, the Gimli glider guys after their initial fuck up, but we don’t hear about all the times human intervention prevented the flight from even making the headlines. I suspect that even the pilots involved aren’t always aware of times where they did something very right and the end result was that nothing newsworthy happened.
I fly a 1990s aircraft and probably every day I fly I manage the aircraft out of a situation that its automatics are incapable of managing itself out of. That is in no way news, the automatics are a tool that I have at my disposal and because I know that I am there to fix things, I have no problem asking it to do things that I know it may have trouble with. I also fly a post 2000 aircraft and I also spend a bit of time “managing” its automatic systems.
No, they certainly can’t. The present state of the art is manual take-off for all aircraft at all airports and then automatic climb, cruise, and descent, for all but the most antiquated aircraft, and automatic landing for most aircraft into most major airports. As I understand it the lack of automatic take-off capability is to do with regulatory requirements, there is no technical problem with it.
This probably gives an unrealistic impression of the actual capabilities though. I could describe many modern cars as being able to cruise and park automatically, but the reality is that they can only cruise under the careful watching eye of a competent driver and they can only park themselves once the same driver has put the car into a position where its automatics can take over. The truth is that for all the automatic abilities of an airliner there is a lot of managing that happens, a lot of setting up, and a lot of ensuring the situation is suitable for the automatics to be allowed to do their thing.
That said, the same automatics are designed under the assumption that a crew of humans will be managing them. I guess my main point is that the current automatics aren’t as automatic as you might think, but that is not necessarily a point against a fully automatic aeroplane.
You know how there are people who refuse to buy version 1 of anything? I’m one of those people. I am also someone who would never fly on a pilotless airliner until I’d seen it proven in commercial operations for at least 10 years. I’m sure I’m not the only one. It would only take a relatively small proportion of people who think like I do to make the choice to go to pilotless airliners a commercially suicidal one for any airline. Don’t forget that airline’s profits, if there is any profit, are often measured in terms of cents or, if they’re lucky, dollars per ticket. the margins are tiny and the risks are huge.
I think that if it happens, it will happen to freighters first and then perhaps after a few decades of proven performance you might see it translate to passenger carrying aircraft. But if the freighters have a bad run, and maybe their maintenance and management won’t be as good because there aren’t any lives involved, that might kill the entire idea off for a human generation or two.
Summary: There probably isn’t any major technical issues with the idea however; it’s not known at the moment how often human’s save a flight so it’s hard to say what positive benefit the technology will have; the current state of the art probably isn’t as close to true pilotless airliners as the public thinks; and it will be hard to get a commercially viable proportion of people to actually fly on one.
That’s my current feeling on the issue, if you ask me again in a year I might feel differently.
I’ll eat my hat if it happens in the next 10-15 years. If Boeing/Airbus aren’t talking it up as a design concept now then it’s a few generations of aircraft away.
You also have to bear in mind that a very large percentage of passengers on commercial flights are there for leisure purposes (70% of passengers, according to Heathrow statistics). That means the airlines’ customers have a lot of flexibility in their choice about whether to fly with a particular airline or whether to fly at all rather than taking a different means of transport or just not making the trip at all.
Even if only half of those leisure passengers decided not to choose the pilotless airline for the first year it was in service - and that’s a very conservative figure for the first year when all the pilotless passengers would feel like guinea pigs - that would mean that airline would lose 35% of its potential passengers. I doubt any airline could cope with that loss.
Trains are the mode of passenger transport where doing without a driver is most feasible and indeed some passenger trains do not have drivers for at least some of their routes. However, they always (in London) have a member of staff on board, and on the DLR (Docklands Light Railway), which is definitely driverless for most of its route, a driver is still necessary for a couple of sections.
The “driverless” trains due to come into use on the tube in London in 2020 still have a driver sitting in the cabin because emergencies do happen - a driver would be able to brake if they saw a person about to fall onto a track, and currently trains can’t and probably never will be able to.
Tube passengers also tend to feel safer if there’s a driver on the train, but it’s not for trivial reasons. You really do need a member of staff (or two, preferably) of some sort, at least, on trains that go underground, for emergencies (sick passenger, train breakdown, power failure, etc), because it’s not like you can send someone in from outside, so it might as well be someone who can also drive the thing. The same would apply to planes.
Autonomous drone landing on a carrier. Autoland has been a feature and has been used on aircraft like the Caravelle since the 70s, and has allowed pilots to land in situations where they simply could not have safely without it.
GBAS and WAAS will only make flying autonomously easier, and once airlines realize how much money they can save by taking pilots out of the equation, you can bet the marketing will be intense to condition people to the idea. I’m betting the first autonomous transoceanic freighters will be in the air within ten years, and people haulers well within twenty. I would argue that attendants will still be onboard though, for the same reasons as the Tube above.
Excellent post.
I do think that the operative word is drone. And that means there is a pilot, just not in the plane.
With today’s remotely piloted drones, the ratio is, I think, one pilot per aircraft – half that of airliners. Over time, this ratio could decline as the pilots only get involved on takeoff/landing and when there is a special situation.
It’s easy to think of factors that will slow down the transition. If each big step has to be freighter-first, that means Boeing/Airbus have to, for several decades, design and manufacturer radically different product lines for freight and passengers.
I can’t see the ground control option ever being completely removed. So, while it should be harder to hijack, I don’t see how there could be a 100 no-hijacking-possible guarantee.
I remain unconvinced. I know there must be new freighters flying around somewhere but my experience is that in general freighters are a lot older than passenger jets. Nobody cares if they’re a bit shabby inside and out as long as they get the job done. Certainly all of the short haul freighters that I see around are very old. Our company were using B727s up until the mid 2000s. I also think airlines know exactly how much money they could save by not having pilots, where is this marketing campaign?
Still, lets say there is a company willing to take a punt on a brand new pilotless freighter. There’s a whole lot more infrastructure needed than just designing the aircraft, changing the ICAO rules, and developing whatever ground support would be needed for an automatic take-off system. There also has to be a complete overhaul of the ATC system so that ATC can provide instructions to the aircraft both in the air and on the ground. I think that technologically, designing a system that works adequately on the ground might be the major challenge, it may even be necessary to go to an external taxi device which would be operated by a person. That kind of major infrastructure change takes a very long time.
If they’re not talking about it seriously now then I’d say it is at least 20 years away for any kind of aircraft. And you still have the problem of getting people to fly on the passenger versions. I certainly wouldn’t, and I can’t be the only person who feels this way. That means that given the option of Airline A with pilotless jets and Airline B with conventional jets many more people will choose B which gives A a commercial incentive not to do it. In addition if it is initially a freight only proposition, do the freight companies have enough money and influence to make such a fundamental change to the way the aviation industry works?
A bit further on the technology required, it is a bit misleading to compare the concept to that of a driverless train. A train is effectively a one dimensional mode of transport. Its movement is restricted to either moving, or not moving, and the only choice beyond that is how fast to move. By comparison an airline pilot is making decisions all of the time. The vast majority of those decisions go unnoticed by passengers because the consequence of them is that the passengers experience a normal flight. It is telling that for such an incredibly simple system like the automated train, there is still somebody on board who’s job it is to take-over in the event of an emergency. This, for a vehicle that can do two things, start and stop, and you suggest that you could have a vehicle with complete freedom of movement flown entirely automatically within 10 years?
Airliner pilots spend a lot of time diverting around thunderstorms for instance, using a mixture of imperfect weather radar, experience, and gut feelings to decide what weather to avoid and what not to avoid, all while negotiating the deviations with ATC. It became apparent after the recent Asiana crash that the public have absolutely no idea what kind of weather related decisions are made on a daily basis in an airliner. Last night I spent the first sector of my flight almost entirely off track avoiding thunderstorms. The sector ended with an arrival and approach to the second biggest international airport in Australia to a runway that does not have an ILS or any other precision approach system in conditions that current state of the art jets would not have been able to auto-land in even if there was a suitable ILS. Auto-land is designed for low visibility, typically caused by fog, which is typically associated with light winds. As far as I know the wind limitations for an auto-land are still more restrictive than those for a manual landing.
How would an automatic weather avoidance system work exactly? What does it do if the weather radar paints thunderstorms in all directions? It is certainly possible to pick your way through storms only to find yourself faced with a wall of them and when you turn around you find the gaps have closed in behind you as well. When that happens the pilot is left “picking the path of least resistance” using a lot of experience and fuzzy logic that our brains are good at but that computer brains are not.
There is also the issue of what to do should the destination airport become unsuitable for some reason, perhaps there is a thunderstorm parked over the airfield. The choice might be to hold and wait for the storm to move on or to divert to another airport. If the only diversion airports available to the auto-jet are ones with the infrastructure to handle the auto-jet then, while the technology is new, you will be severely limited by where you can fly. Unless all of the airports upgrade at the same time, the auto-jet will be carrying a lot more diversion fuel than the piloted jets and carrying fuel costs money.
When the 9/11 attacks happened, every airliner in the US was immediately grounded. How do you think that kind of emergency would go down with automated jets? It would be a complete shambles, more so than it was.
TLDR: I disagree.
Yes, drone is a better word, and the concept of a ground controlled aircraft is a lot more realistic than that of a true automatic aircraft. One issue with ground control is that for take-off and landing you need the drone pilot to be on site to reduce the effect of transmission delays, this would also follow through to having drone staff at the diversion airports as well which would significantly offset the cost savings of not having onboard pilots.
On the subject of pilot numbers, you are correct that current airliners are designed to be flown by two pilots. Long haul flights typically need at least one relief pilot as well as it’s deemed to be unsafe expecting the crew to stay awake for 14 hours at a time. I’m not sure how minimal you could make your drone control staff. The truth is that you don’t really need two pilots anymore, I would be very comfortable flying around by myself most of the time. If I was to have a heart attack though I’d need someone to take over. That problem would also exist for drone pilots but drone pilots have the advantage that when they are not needed for one flight they could be controlling another flight.
One thing I’ve learned over the years, I have yet to come across an area of expertise that was as simple in reality as it seemed from the outside. I think most who consider pilotless airliners a realistic proposition within the next 20 years likely have no airline piloting experience, or no aircraft system design experience, or neither. It may seem like it’s not far off but the devil is in the details.
I wouldn’t be terribly surprised if the latest crash in Europe by a suicidal co-pilot might spur more reseach, and perhaps lead to autonomous or remote piloting.
Remote piloting might be an option, but I can’t see ever getting on a plane and letting a computer fly it, at least not with passengers on it.
That’s more of a psychological rather than a technological issue. Suicide or not, IIRC most chrashes are caused by pilot error.
However, I don’t want to suggest that technology will trump psychology – in fact, the opposite. So there’s a very good chance youre correct.
As with self-driving cars, a major push will probably come from the price of a ticket. If autopilot reduces crashes then it reduces the amount of insurance the airline has to pay. I have no idea how much that is but if saved you $50 would that be enough to get you on the plane?
Yes. Shit, I’d pay $50 more to get on an airplane if I was convinced it was safer.