"Pilotless airliners safer" - London Times article
Join Date: Sep 1999
Location: Limoges/Sussex
Posts: 21
Likes: 0
Received 0 Likes
on
0 Posts
A quick reminder
Great thread.
Consider, though - no human has ever actually "touched" the control surfaces of any Airbus product since the A320 first took to the skies in the 80's.
Even ECAM drills are a matter of putting the switches into the position which the software has already decided is safest.
So much for decision-making...
Consider, though - no human has ever actually "touched" the control surfaces of any Airbus product since the A320 first took to the skies in the 80's.
Even ECAM drills are a matter of putting the switches into the position which the software has already decided is safest.
So much for decision-making...
Join Date: Apr 1999
Location: Hong Kong
Posts: 305
Likes: 0
Received 0 Likes
on
0 Posts
Even ECAM drills are a matter of putting the switches into the position which the software has already decided is safest.
Join Date: Jan 2001
Location: Home
Posts: 3,399
Likes: 0
Received 0 Likes
on
0 Posts
RetiredF4
There is an accepted methodology in scientific debate.
You put forward an idea.
You say why you think it and try to put forward evidence to support it, either experimental or references.
An example.
You say "why has nobody invented a system that cuts in when the sh1t hits the fan?"
You put forward no evidence to support this, however in the spirit of debate...
I say...
NASA-Pioneered Automatic Ground-Collision Avoidance System Operational | NASA
This is where you should come in and either say "oh I'm terribly sorry, I'll do some research in future before making statements" or find some alternative evidence to support your belief.
Or you could just move on and not register your mistake ....
The Air France example and others like it do not help your case. That aircraft was designed that the failure mode is to hand it to the pilot. As has been stated repeatedly, if you were pilotless you would have another option. It just so happens that pitch and power would be more than enough.
Piltdown
I don't disagree with anything you say, except that I think it will be done in the next 20yrs, and since I'm not the one who will be making billions from it, I respectfully decline the offer to speak to grieving relatives.
There is an accepted methodology in scientific debate.
You put forward an idea.
You say why you think it and try to put forward evidence to support it, either experimental or references.
An example.
You say "why has nobody invented a system that cuts in when the sh1t hits the fan?"
You put forward no evidence to support this, however in the spirit of debate...
I say...
NASA-Pioneered Automatic Ground-Collision Avoidance System Operational | NASA
This is where you should come in and either say "oh I'm terribly sorry, I'll do some research in future before making statements" or find some alternative evidence to support your belief.
Or you could just move on and not register your mistake ....
The Air France example and others like it do not help your case. That aircraft was designed that the failure mode is to hand it to the pilot. As has been stated repeatedly, if you were pilotless you would have another option. It just so happens that pitch and power would be more than enough.
Piltdown
I don't disagree with anything you say, except that I think it will be done in the next 20yrs, and since I'm not the one who will be making billions from it, I respectfully decline the offer to speak to grieving relatives.
Last edited by Tourist; 5th Dec 2014 at 05:34.
Even ECAM drills are a matter of putting the switches into the position which the software has already decided is safest.
So much for decision-making...
So much for decision-making...
Having gone through Uni in the 70@s when AI was supposed to be around the corner forgive me for being a bit sceptical about some of the current claims.
I'm also well aware (from my friends that stayed in academics,) that these days it's all about funding and getting your project into the public eye...so I don't think providing links to any PR about some blue sky project gives the general public any real idea about the state of play ( putting forward links to projects is not "scientific debate"...but on that subject I could provide some great links on cold fusion BTW...).
IMHO reckon you might just see regular unmanned freight across the pond, between specific freight hubs, in the next 20 -30 years.
Routine unpiloted passenger flights between the like of CDG, LHR, JFK...at the very least 50 years off...but maybe that's me being a luddite.
Tourist
This is where you should come in and either say "oh I'm terribly sorry, I'll do some research in future before making statements"
This is where you should come in and either say "oh I'm terribly sorry, I'll do some research in future before making statements"
You would wonder how often ECAM procedures include the famous word "if" in wording or in sense. All piloted aircraft are designed for the failure mode to hand it to the pilot, my words. The computers do the easy stuff, easy to design, easy to implement, and easy to drop out when the sh*t hits the fan. That is the present concept. Your unmanned drones have a bit of explosives for those cases when the computers run out of ideas. What proof do you have, that this failure fall back mode of yours concerning AF447 ( you brought that example yourselve) is anywhere available in close future?
I think you got too far detached from real flying, leading a meeting in a comfortable chair is different to flying aircraft, even when computers help in doing the job.
Last edited by RetiredF4; 5th Dec 2014 at 11:12.
All piloted aircraft are designed for the failure mode to hand it to the pilot, my words.
There are good arguments against autonomous flight but that is not one of them. Every failure mode could ultimately be automated. In each case a fair sized mountain of a safety case would need to be addressed, but it could be done. The big question - is it "worth the candle".
Join Date: Dec 2006
Location: KUL
Posts: 431
Likes: 0
Received 0 Likes
on
0 Posts
He is rights
F4 dude is right. In 5 years your car will offer you to drive you home from the pub, or your kid to school. Not much later the public will demand the same technology on airliners. The pilot of the future will reside over a cubicle in flight ops and control 20 or more aircraft during one shift.
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,574
Received 422 Likes
on
222 Posts
The deciding factor in most things is cost. For example, it's all very well to design an automated car but it relies on an infrastructure. That infrastructure costs money. The way things are going round these parts, they aren't even repairing the road surface, due to lack of funding, let alone building in the necessary new equipment for autonomous cars to be commonplace.
Same with new medical research. The manufacturers can produce miracle drugs but they are often too expensive to be put in common use.
It's no different for airborne cars.
There remains the problem of customer resistance. If the customer won't buy the product, the designers have wasted their time.
Same with new medical research. The manufacturers can produce miracle drugs but they are often too expensive to be put in common use.
It's no different for airborne cars.
There remains the problem of customer resistance. If the customer won't buy the product, the designers have wasted their time.
Nemo Me Impune Lacessit
Join Date: Jun 2004
Location: Derbyshire, England.
Posts: 4,091
Received 0 Likes
on
0 Posts
F4 dude is right. In 5 years your car will offer you to drive you home from the pub, or your kid to school. Not much later the public will demand the same technology on airliners. The pilot of the future will reside over a cubicle in flight ops and control 20 or more aircraft during one shift.
Spoke to a good friend yesterday who is an underwriter and partner for one of the major aviation insurance syndicates at Lloyds, he has followed PPRuNe for many years, mainly for entertainment, but loves the threads like this one, as I mentioned in an earlier post, "It aint going to happen"
Nemo Me Impune Lacessit
Join Date: Jun 2004
Location: Derbyshire, England.
Posts: 4,091
Received 0 Likes
on
0 Posts
Even ECAM drills are a matter of putting the switches into the position which the software has already decided is safest.
So much for decision-making...
So much for decision-making...
Join Date: Aug 2007
Location: UK
Posts: 0
Likes: 0
Received 0 Likes
on
0 Posts
Parabellum, of course you are right, all the professional pilots on this forum will agree and so will the non-pilots who do not have a commercial agenda. I like the troll bit too as many bit the bullet and have waded in, much to the delight, no doubt, of the envious poor souls out there who would love a leather jacket and big watch !
Sunning myself with an old mate who told me of the glory days when we would allow pax visits to the Flight Deck. One guest was a train driver who suddenly remarked, out of the blue, that he felt my colleague was not worth the money he was getting as he appeared to be "not doin nuffink!". Skipper replied, "Well, Driver West, at the moment we are around 5 miles above the planet earth", 'Yeah ?', replied West, in quizzical fashion, "and", continued Commander," we are doing, about, eight miles a minute," ' Yeah ?', replied West, in quizzical fashion, "and" continued Boss, "if anything fails, we can't stop !". West went ashen-faced, bleated 'F..K !!' and scurried back into the cabin.
We are talking about Air Transport here with all the regulation it requires. Of course the paying public will never stump up the cash to set foot in a pilotless, public transport aircraft. The very notion beggars belief. I agree with Parabellum as I , also stated in previous posts, It aint gonna happen.
Sunning myself with an old mate who told me of the glory days when we would allow pax visits to the Flight Deck. One guest was a train driver who suddenly remarked, out of the blue, that he felt my colleague was not worth the money he was getting as he appeared to be "not doin nuffink!". Skipper replied, "Well, Driver West, at the moment we are around 5 miles above the planet earth", 'Yeah ?', replied West, in quizzical fashion, "and", continued Commander," we are doing, about, eight miles a minute," ' Yeah ?', replied West, in quizzical fashion, "and" continued Boss, "if anything fails, we can't stop !". West went ashen-faced, bleated 'F..K !!' and scurried back into the cabin.
We are talking about Air Transport here with all the regulation it requires. Of course the paying public will never stump up the cash to set foot in a pilotless, public transport aircraft. The very notion beggars belief. I agree with Parabellum as I , also stated in previous posts, It aint gonna happen.
Join Date: Sep 2013
Location: UK
Posts: 730
Likes: 0
Received 0 Likes
on
0 Posts
Sorry, Tourist, but your faith in the infallibility of machines and computers is more than a little disturbing, and your stated view that all modern pilots are incapable of light without the aid of computers is as insulting as it is plain wrong. There are some pilots who have been lead into automation dependency because of dire airline management philosophies, and I am proud to count myself as one amongst many who stand against that deterioration and encourage as much hand, visual and raw data flying as circumstances prudently permit. Many companies are trying to reverse the rot, and there are a lot of skilled operators out there, just as before.
If all pilots were inept and all computers infallible, then you'd be making sense, Tourist, but you come across as a theoretical engineer and someone who has never flown the line. Even line engineers, despite all the banter, know full well that computers aren't reliable enough to be left unsupervised and in total control. And a man far more intelligent than any of us on this forum just made a public statement that AI will be the death of humanity - a point often cheesily made in Hollywood, but true all the same.
If all pilots were inept and all computers infallible, then you'd be making sense, Tourist, but you come across as a theoretical engineer and someone who has never flown the line. Even line engineers, despite all the banter, know full well that computers aren't reliable enough to be left unsupervised and in total control. And a man far more intelligent than any of us on this forum just made a public statement that AI will be the death of humanity - a point often cheesily made in Hollywood, but true all the same.
Join Date: Jun 2001
Location: UTC +8
Posts: 2,626
Likes: 0
Received 0 Likes
on
0 Posts
Every failure mode could ultimately be automated.
Of course it will never happen. I've just had a long argument on a science forum with some science folk who reckon that personal, automatic pilotless cars will be both possible and available in the near future.
They say the machines will "....have 100% redundancy and be computer controlled so they won't be able to crash."
The only 'evidence' for their assertions is a single driverless car that Google is developing.
When I asked how such a machine would be able to look down and assess the landing site on someone's driveway for cross wind, turbulence, fixed hazards, loose hazards, loose pets and children, they told me "360 degree cameras, range finders, infra red, and anyway the technology is already used in CGI films". I then asked how any computer would be able to collect all that information and assess the risks by PREDICTING what might happen, in the way a helicopter pilot does.
After several carefully reasoned posts from me I have given up. Thing is they read too many comics and believe science fiction programs.
I don't see how there can ever be pilotless passenger aircraft. Before there could be, the human brain would need to be completely modelled. As of now, only a small part of the decision making process can be simulated. What about icing, avoiding thunderstorms, lightning, engine failures, gear failures, flap jams, generator failures, medical emergencies, bird strikes, explosive decompression, cabin fires etc. etc. ???? No computer or software could ever be developed to consider and react to the massive amount of variables and potential problems in even the most 'simple' passenger flight.
And before pilotless airplanes, we would need to have developed and proven driverless trains. (Operating in one dimension: forward speed) I know there are some, but we would need to operate and prove 100% of them. Then we would need to develop and prove driverless cars and lorries - 100% of those too. (Operating in two dimensions on defined roads). Only then could we even begin to dream of the concept of pilotless commercial aircraft. (Operating in three dimensions in the air and also unable to stop within a few vehicle lengths at any time).
The concept of pilotless computer controlled aircraft has arisen because some pilots make accidents, and some people's knee-jerk reaction is to say "well, obviously we must get rid of the pilots then, they are the problem" NOOOO! What is actually the root cause of this? = Bad training. Bad working conditions. Long working hours and working during the circadian low points. Minimum possible training hours to just pass then onto the line in an A320 with less than 200 hours actual flying experience in an actual aeroplane and NO experience in actual multi engined aircraft or of flying in bad weather. Or onto a Dash 8 without possessing the correct instinctive reaction to a stall for example.
Instead of all this nonsense, let's go back to some decent pilot training, some decent hand flying practice and some decent conditions and decent hours and decent contracts for pilots. That's what is really needed.
They say the machines will "....have 100% redundancy and be computer controlled so they won't be able to crash."
The only 'evidence' for their assertions is a single driverless car that Google is developing.
When I asked how such a machine would be able to look down and assess the landing site on someone's driveway for cross wind, turbulence, fixed hazards, loose hazards, loose pets and children, they told me "360 degree cameras, range finders, infra red, and anyway the technology is already used in CGI films". I then asked how any computer would be able to collect all that information and assess the risks by PREDICTING what might happen, in the way a helicopter pilot does.
After several carefully reasoned posts from me I have given up. Thing is they read too many comics and believe science fiction programs.
I don't see how there can ever be pilotless passenger aircraft. Before there could be, the human brain would need to be completely modelled. As of now, only a small part of the decision making process can be simulated. What about icing, avoiding thunderstorms, lightning, engine failures, gear failures, flap jams, generator failures, medical emergencies, bird strikes, explosive decompression, cabin fires etc. etc. ???? No computer or software could ever be developed to consider and react to the massive amount of variables and potential problems in even the most 'simple' passenger flight.
And before pilotless airplanes, we would need to have developed and proven driverless trains. (Operating in one dimension: forward speed) I know there are some, but we would need to operate and prove 100% of them. Then we would need to develop and prove driverless cars and lorries - 100% of those too. (Operating in two dimensions on defined roads). Only then could we even begin to dream of the concept of pilotless commercial aircraft. (Operating in three dimensions in the air and also unable to stop within a few vehicle lengths at any time).
The concept of pilotless computer controlled aircraft has arisen because some pilots make accidents, and some people's knee-jerk reaction is to say "well, obviously we must get rid of the pilots then, they are the problem" NOOOO! What is actually the root cause of this? = Bad training. Bad working conditions. Long working hours and working during the circadian low points. Minimum possible training hours to just pass then onto the line in an A320 with less than 200 hours actual flying experience in an actual aeroplane and NO experience in actual multi engined aircraft or of flying in bad weather. Or onto a Dash 8 without possessing the correct instinctive reaction to a stall for example.
Instead of all this nonsense, let's go back to some decent pilot training, some decent hand flying practice and some decent conditions and decent hours and decent contracts for pilots. That's what is really needed.
Last edited by Uplinker; 5th Dec 2014 at 11:31.
Join Date: Feb 2014
Location: England
Posts: 2
Likes: 0
Received 0 Likes
on
0 Posts
What makes you think you were in a driverless train ?
Trains - at LHR well there is the HEX and HEC the passenger trains, both have drivers, well a bag of meat at the front. (Self described driver). Guys and Gals who all to frequently incur SPADS. (Mostly in the shunting yards) - I guess akin to ground movements.
Then the disused PAD - LHR baggage only line - and last (that I am aware of) is the very busy secured baggage system, so unless you talking about the staff car park bubbles ? More of a Disney ride I don't see it as these I don't think go between the terminals,
On the subject of automation, I consider nuke power stations would be a better system to start to automate, remove the people from the control room - consider the risk of an aircraft hitting a full stadium vs the risk of issue at a nuke power plant, - of course worse case would be the aircraft hitting the said control room.
Then the disused PAD - LHR baggage only line - and last (that I am aware of) is the very busy secured baggage system, so unless you talking about the staff car park bubbles ? More of a Disney ride I don't see it as these I don't think go between the terminals,
On the subject of automation, I consider nuke power stations would be a better system to start to automate, remove the people from the control room - consider the risk of an aircraft hitting a full stadium vs the risk of issue at a nuke power plant, - of course worse case would be the aircraft hitting the said control room.
Join Date: Sep 2013
Location: UK
Posts: 730
Likes: 0
Received 0 Likes
on
0 Posts
To programme computers to handle every eventuality, the engineers first have to envisage every eventuality, and we see time and time again new and interesting scenarios with subtleties previously unconsidered. And of course, the fallible human engineers have to build and programme these wonder machines...
Rotate on this!
Join Date: Jan 2004
Location: Aberdeen
Age: 64
Posts: 403
Likes: 0
Received 0 Likes
on
0 Posts
"It aint going to happen"
Join Date: Aug 2013
Location: FL060
Posts: 147
Likes: 0
Received 0 Likes
on
0 Posts
Not really a comparison...
Ummm, Me 262 was a wartime aircraft flying in a war in the '40s. Please do not compare that to a commerical air service in peace time. There are no similies or metaphors to be used for flight safety.
My favourite example of how humans can save an a/c is illustrated in this clip (skip to 28:05). Please write me an algorithum to fix the problem at 34:05.
https://www.youtube.com/watch?v=WfNB...pb1dg&index=14
Sorry, don't know yet how to embed the veiwer...
My favourite example of how humans can save an a/c is illustrated in this clip (skip to 28:05). Please write me an algorithum to fix the problem at 34:05.
https://www.youtube.com/watch?v=WfNB...pb1dg&index=14
Sorry, don't know yet how to embed the veiwer...
Join Date: Jan 2001
Location: Home
Posts: 3,399
Likes: 0
Received 0 Likes
on
0 Posts
Humans are almost exactly what you don't want in a modern cockpit.
We can be very good at things if we have enough practise.
In a modern airliner we so rarely ever get to practise it might as well be never.
We respond badly to dull repetitive tasks.
Airline flying is dull and repetitive.
We respond badly to time zone changes and unsociable hours..
Airline flying is all about the unsociable.
We work well in well rehearsed teams.
All the good SOP stuff can't avoid the fact that we have rarely even met the other guy.
We have a limited capacity for concurrent tasks before we max out.
A modern airliner is spectacularly complex.
Machines don't get tired.
They don't need practise
The love dull
They care not about unsociable.
They don't need to rehearse.
They excel at complex concurrent tasks
I say again.
I don't expect them to be perfect. There will be smoking holes.
They just have to be better than humans, which isn't hard.
People keep mentioning how often they have to make an input.
The first point I will make yet again is that the current crop are designed that way. They expect a human to be there and they use you to make that input.
The second point is a question. How many times is that input just turn it on and off again.
In an Airbus, nearly everything is solved by "leave it alone, it's just having a bad morning" or off then on.
The third of course is why are you judging the future on a really old piece of archaic junk like an Airbus?
Glueball
Read back a few pages.
There is a strong suggestion that the river was not the best option. If that is the case, that is the sort of thing a computer is good at judging. It is just a matter of geometry and speed distance time glide angle calculations. That is where a computer has us totally beaten.
Uplinker.
I don't think you have thought this through.
What do you do in an engine/generator/gear failure/flap jam failure that involves thought? You follow the ECAM. You are not supposed to think. All you currently are is an error waiting to happen as you don't follow the instruction properly or most likely you make a tiny error in the landing distance chart and totally mess up.
The ECAM tells you what to do.
You do it. well done, you must be a pilot.
Yes there are non ECAM procedures, but they are usually just awaiting an ECAM change or outside the current scope of the very old systems to check and monitor.
I'm guessing the fuel imbalance checklist is like the Airbus one and for similar reasons. Don't try to tell me you think that is outside the scope of a computer to run.
Aluminium shuffler
You talk as if I think that current airliners should be left to get airborne by themselves. They shouldn't, they are not designed to.
A line engineer working on 30yr old computer tech will of course say that. Look at your 30 yr old phone/telly/car. It's sh1te.
The world changes fast.
Many flight engineers thought it couldn't happen to them. Where are they now?
Can we please put to bed this "we can't even do driverless trains!" thing.
List of automated urban metro subway systems - Wikipedia, the free encyclopedia
Is that enough trains running by themselves? I'm not sure that it matters, because it is vastly simpler, practically one dimensional, but here they are..
Cars are inbound at great speed.
Autonomous car - Wikipedia, the free encyclopedia
We can be very good at things if we have enough practise.
In a modern airliner we so rarely ever get to practise it might as well be never.
We respond badly to dull repetitive tasks.
Airline flying is dull and repetitive.
We respond badly to time zone changes and unsociable hours..
Airline flying is all about the unsociable.
We work well in well rehearsed teams.
All the good SOP stuff can't avoid the fact that we have rarely even met the other guy.
We have a limited capacity for concurrent tasks before we max out.
A modern airliner is spectacularly complex.
Machines don't get tired.
They don't need practise
The love dull
They care not about unsociable.
They don't need to rehearse.
They excel at complex concurrent tasks
I say again.
I don't expect them to be perfect. There will be smoking holes.
They just have to be better than humans, which isn't hard.
People keep mentioning how often they have to make an input.
The first point I will make yet again is that the current crop are designed that way. They expect a human to be there and they use you to make that input.
The second point is a question. How many times is that input just turn it on and off again.
In an Airbus, nearly everything is solved by "leave it alone, it's just having a bad morning" or off then on.
The third of course is why are you judging the future on a really old piece of archaic junk like an Airbus?
Glueball
Read back a few pages.
There is a strong suggestion that the river was not the best option. If that is the case, that is the sort of thing a computer is good at judging. It is just a matter of geometry and speed distance time glide angle calculations. That is where a computer has us totally beaten.
Uplinker.
I don't think you have thought this through.
What do you do in an engine/generator/gear failure/flap jam failure that involves thought? You follow the ECAM. You are not supposed to think. All you currently are is an error waiting to happen as you don't follow the instruction properly or most likely you make a tiny error in the landing distance chart and totally mess up.
The ECAM tells you what to do.
You do it. well done, you must be a pilot.
Yes there are non ECAM procedures, but they are usually just awaiting an ECAM change or outside the current scope of the very old systems to check and monitor.
I'm guessing the fuel imbalance checklist is like the Airbus one and for similar reasons. Don't try to tell me you think that is outside the scope of a computer to run.
Aluminium shuffler
You talk as if I think that current airliners should be left to get airborne by themselves. They shouldn't, they are not designed to.
A line engineer working on 30yr old computer tech will of course say that. Look at your 30 yr old phone/telly/car. It's sh1te.
The world changes fast.
Many flight engineers thought it couldn't happen to them. Where are they now?
Can we please put to bed this "we can't even do driverless trains!" thing.
List of automated urban metro subway systems - Wikipedia, the free encyclopedia
Is that enough trains running by themselves? I'm not sure that it matters, because it is vastly simpler, practically one dimensional, but here they are..
Cars are inbound at great speed.
Autonomous car - Wikipedia, the free encyclopedia
Where's Mayday38 when you need him?
He had about 10 seconds to decide what to do. Maybe a bit less.
What he did was not what the 'puter would have done. 'Puter, (programmed by humans) would have done something else. Even more quickly than Mayday. ...........and got it wrong.
Mayday38 got it right.
He had about 10 seconds to decide what to do. Maybe a bit less.
What he did was not what the 'puter would have done. 'Puter, (programmed by humans) would have done something else. Even more quickly than Mayday. ...........and got it wrong.
Mayday38 got it right.