PDA

View Full Version : The 'ol "automatics Vs pilots"


Gnirren
11th Mar 2006, 20:20
I think most of us are familiar with the idea that technology is overtaking the flight deck slowly and there are arguments that pilots could soon become obsolete. Automated flights with either empty flight decks, ground staff monitoring with the possibility of remote override controls or an onboard technician.
I thought about it the other day, what with all these accident stats and CRM courses and other sources of information that pilots digest. I don't have the percentages exactly but it's safe to say that out of the accidents that occur a good what, 80% are deemed to have been caused by "pilot error" of some sort?
Of course, us pilots like to pretend that we are invaluable to the industry and maybe we are. Thinking of how we drill and drill and basically "program" our brains with SOPs and procedures so we can perform you could say that we try to approximate what an onboard computer can do effortlessly and consistently. So now we have pilots playing catch-up with the systems of sorts.
Typically what the pilot community sais with regards to the pilot-less flight deck is "well what happens when the sh*t hits the fan and the systems fail?" and "well, no passenger is ever going to accept a plane without a pilot" Now think about it, reading accident reports did the plane crash because the aircraft broke down? How often does a modern 737-800 or A320 for example malfunction badly enough that an automation would be unable to handle it? There are retired captains out there with 10k+ hours who never once had an engine fail, a fire start, or even had to touch the QRH.
Typically it's the pilot(s) who end up doing or not doing something and that causes the accident to occur, and it can often be small things that a computer would have performed correctly but with our human fallability we miss. Helios missing the bleed configuration for example, or maybe it was the packs I'm not sure. Either way, would a computer have missed it? The midair collision where a pilot didn't obey the TCAS alert, would automation have done the same?
I guess my ultimate point is, what would the overall statistics be if you automated flight thus removing the accidents caused by pilot error such as poor CRM - shutting down the running engine, CFIT or what have you, and introducing the crashes introduced by automation breakdown and technical errors?
I have a sinking feeling that the result might point towards automation, after all how many situations arise where the "human qualities" of a pilot end up saving the day and if so, did he save the day from a situation that he or she created in the first place through his own error? It feels like the human qualities that we try to prove our case with are often the reason why there are incidents to begin with.
Sensitive topic I suppose, no pilot likes to hear something that rings closely to an accusation or a proposal that he's reduantant but I fly same as you and I feel it's an interesting topic. No I'm not trying to start some sort of flame war or whatever, just curious that's all.
Any thoughts?

Gingerbread Man
11th Mar 2006, 20:38
I really don't know if passengers would feel safe on a completely automated aircraft. After all, the technology exists to have automated cars, and this is far simpler than aircraft flight control, but I doubt many people would have a car that was computer driven.

Even trains have people driving them, and all they get is go/stop controls. I can see why you would think pilotless aircraft are not far away, but I think that there are simpler machines that society would want to see automated before airliners were.

Also, it would be far less fun for pilots :) .

GBM ;)

haughtney1
11th Mar 2006, 21:01
Typically what the pilot community sais with regards to the pilot-less flight deck is "well what happens when the sh*t hits the fan and the systems fail?" and "well, no passenger is ever going to accept a plane without a pilot" Now think about it, reading accident reports did the plane crash because the aircraft broke down? How often does a modern 737-800 or A320 for example malfunction badly enough that an automation would be unable to handle it? There are retired captains out there with 10k+ hours who never once had an engine fail, a fire start, or even had to touch the QRH.
Well Gnirren I dunno how much you know about modern cockpit automation..and really thats a little irrelevant. As an example of something that you seem to be implying wont happen that often..there was a certain BBJ airborne out of an Airport north of London (that boeing wont advertise) had exactly the type of failure you describe whilst in autoflight mode late last year....the aircraft flew out of control around the london TMA for about 7 minutes before the "Pilots" were able to regain control and fly it manually. If there is a possibility Murphy states its gonna happen.....automation is an aid, not a sloution:ok: Are you gonna put your wife and kids on that pilotless aircraft..I know I wouldnt..(even if I had a wife and kids:} )

Onan the Clumsy
11th Mar 2006, 21:12
I have a computer that's computer driven and I'll be damned if I'll get on an airplane that is too :*

VP959
11th Mar 2006, 21:12
Even trains have people driving them, and all they get is go/stop controls.
GBM ;)

Well, we already have driverless automatic trains, the Docklands Light Railway for one.

The defence view is that combat aircraft will probably be pilotless within the next 15 to 20 years. Most investment in aircarft technology is now going into UCAVs.

My guess is that the overall benefit of having a pilot on board a transport aircraft will gradually reduce to the point where human intervention errors cause more accidents than technical failures. When that point is reached I rather think that pilots will become redundant.

VP

Onan the Clumsy
11th Mar 2006, 21:26
pilotless combat aircraft are a different kettle of fish though as it allows for either deep penetration missions, or the ability to pull an unsurvivable amount of g.

Ropey Pilot
11th Mar 2006, 21:32
There are retired captains out there with 10k+ hours who never once had an engine fail, a fire start, or even had to touch the QRH. (My Bold)

Gnirren:
Can you back that statement up? (Not being aggressive - am curious and would be astounded if it was true). I fly in a modern glass cockpit jet and have probably had a master caution (and associated QRH 'touching') probably on average of once a month since I started flying it!

All v. minor and most cleared themselves but would still be very surprised if anyone could go 10k hours without one (the eng fail/fire I'll happily believe)

steinycans
11th Mar 2006, 21:46
It's a nice idea to have computers controlling the planes like that without pilots but the problem is that computers themselves are programmed by other human beings who are just as likely to make mistakes or leave out essential commands. For example if we look at the midair where the pilot disobeyed the TCAS, at the time (afaik) there was no universal set-in-stone convention on how to deal with it in relation to atc instructions. Whose to say the software writer for that plane wouldn't have written in the same error?
There will still be human error crashes because the only way to really test systems 100% is operationally and over a long period of time. The only difference is that the errors will be made by computer programmers.

Gnirren
11th Mar 2006, 21:54
Can you back that statement up? (Not being aggressive - am curious and would be astounded if it was true). I fly in a modern glass cockpit jet and have probably had a master caution (and associated QRH 'touching') probably on average of once a month since I started flying it!

All v. minor and most cleared themselves but would still be very surprised if anyone could go 10k hours without one (the eng fail/fire I'll happily believe)

Yes, well I've been told this straight from the horse's mouth so to speak, now if he was telling the truth or not is a different story. It could have been an attempt to impress on me how reliable the aircraft is, but that's what I have been told anyway.

Onan the Clumsy
11th Mar 2006, 22:13
The only difference is that the errors will be made by computer programmers.Not really. They'd probably be made by management. Some programmer would say "I just need a little longer because I was wondering what if blah blah" and management would say "We have a deadline. I don't care if passengers get killed"

soddim
11th Mar 2006, 22:14
Problem with computer piloted aircraft is still human error. Instead of human error on the flight deck it is human error in either the software or the engineering. In either case, if there is no human present to pick up the pieces and produce a flexible remedy the passengers die.

Welcome aboard the World's first fully automatic flight fully automatic flight fully automatic flight........................

tinpis
11th Mar 2006, 22:23
We have moved into computer controlled aircraft
Have done for what 25 -30 odd years?
Doesnt take a gimlet eyed hairy chested pilot type to operate em either.
Just a clerk really is all thats required now.

Gnirren
11th Mar 2006, 22:32
Another point of course is that IF aircraft manufacturers should ever venture into automated airliners then the systems would be built for this purpose. It's not like they would just take a stock plane and do some quick alterations to make it work. There would also have to write a complete set of regulations regarding their operations, so trying to answer the question "is it possible" by looking at existing aircraft and procedures probably won't work very well.

Personally I'm just happy that the CDU doesn't have a microsoft logo on it...

Gingerbread Man
12th Mar 2006, 00:02
VP959, I didn't realise that, but my point was that it will take a long time to become commonplace.

Personally I'm just happy that the CDU doesn't have a microsoft logo on it...

I'll never forget Jeremy Clarkson sitting on the deck of an A340 into Heathrow, watching the plane land automatically as the captain just monitored it, turning to the camera and saying "I wouldn't trust my computer to do this". Classic :ok: .

Ginger :D

Send Clowns
12th Mar 2006, 00:32
What about my friend who had to pull all the CBs on his autopilot, because it wanted to fly in completely the wrong direction, and wouldn't even disconnect on the sidestick button? An A320-load flying in circles over the North sea until they ran out of fuel?

The not obeying the TCAS would still have happened, as their SOPs were to obey ATC, so that was a classic case of where a computer would have done the wrong thing, a human had a chance to at least check. Unfortunately that did not help in this case. How many cases have we never heard about because they did check?

soddim is quite correct.

Jerricho
12th Mar 2006, 01:14
The concept of automation always invokes some interesting (and emotive) responses. IMO, there are too many variables, especially in a Terminal environment. I've often heard much speculation regarding the automation of ATC. Unfortunately, even with the best flow management practices in the world, an airport has a finite length of runway, a finite capacity and an infinite number of variables that can screw the whole thing up.

Also, how about the other sort of "CB's" out there - weather. Computer playing the "let's dodge the stuff on the Weather Radar"........lost count of the number of times I've had somebody request to go around something that a driver 10 miles ahead is happily flying through reporting "it's just a bit of rain".

Huck
12th Mar 2006, 01:50
I landed an MD11 last night in gusty crosswinds that exceeded the autoland limits of the autopilot. I looked a bit like a monkey humping a greased football, but got a nice touchdown out of it.

The difference between ham and eggs? The chicken's involved, but the pig's commited.

No ground-based "operator" is going to have anywhere near the motivation I had last night, putting myself down at 164 knots in the gusts.... He kills people, he may feel real bad, but he's still driving home to the kids that night.

Wino
12th Mar 2006, 02:24
All you need to know about this is the accident statistics for the UCAV and other unmanned aircraft the military is flying. They crash all the time, and I mean ALL THE TIME.

We have lost so many predators to accidents, and what not, (and those are the cheap ones.) we lost the entire fleet of the dark stars or whatever they were called.

Cheers
Wino

acbus1
12th Mar 2006, 09:15
This threads a wind up, innit!

Anyone who flies old, new, dial instruments, glass screen, fifty computered, no computered.......whatever flavour......aeroplanes knows only too well that, without surprisingly frequent human intervention, the accident rate would swing from "80% due to pilot error" to "99.9% due to aircraft failure/glitch/foible/design error/computer throwing a wobbly in need of CTRL-ALT-DEL/whatever."

And that the 99.9% would be of a much greater number of accidents!



Try "2001 - A Space Oddysey" for a refresher, if you're in any doubt. :}

E. Mach
12th Mar 2006, 11:06
Here a nice A-320 fully automated:} :} :} :} :} :} :
http://www.break.com/index/planemissesrunway.html
Will you still go on an aircraft with no pilots???:eek: :eek: :eek: :eek:
:eek: :eek:
Me not :} :} :}
E. MAch

Sleep Deprivation Chamber
12th Mar 2006, 12:02
. . .I looked a bit like a monkey humping a greased football, but got a nice touchdown out of it.
Never seen a "touchdown" like that before on the field or in a plane. Got pictures? ;)

Wino
12th Mar 2006, 12:56
Yeah, it is a wind up, or based on bad data.

80 percent of accidents are no longer because of pilot error anyway.

That was true after jets came onto the scene, and then CRM was progressively implemented and refined and the balance of a much smaller number of accidents is now mechanical failure...


So yeah its either a windup or a post by someone who is ignorant of the true statistics these days.

It has always been the dream of the ground pounders (especially british ones for some reason) to declare the pilot obsolete (how many times did the british military aquisitions establishment declare just that?) yet time and again they have been proven wrong.

Cheers
Wino

Non Normal
12th Mar 2006, 13:02
Will low cost carriers start selling flight deck seats as 'premium section' or whatever for x5 the normal fare, if they manage to get rid of the need for pilots onboard? :}

Jerricho
12th Mar 2006, 13:58
Ooooh, hadn't thought of that!

I bagsie shotgun.

BlooMoo
12th Mar 2006, 21:59
the technology exists to have automated cars, and this is far simpler than aircraft flight control

Literally true but I don't think necessarily fair. The variables involved in managing the control surfaces of an aircraft in 'normal' - say 90% - of conditions are complex yes, but manageable given the inherent safety envelope of being 'up in the air' in normal conditions with ATC management of traffic vs on the ground in same conditions and no ground based equivalent of ATC (eg A4). Deviate by 1ft 'on the ground' for 0.2 of a second (in a car) and you can forget about the latency of automation - you've hit the kerb(control inputs now off the scale)tree/pedestrian/oncoming car etc. and its game over. Same is not true in normal flight (or in fact almost all weather abnormal flight).

Automation of flight from a few hundred ft on climbout to a few hundred on approach is commonplace although currently and foreseeably pretty inadequate in situations of

1) weather (anything outside the 90% say)
2) admin - dodgy comms or dodgy ATC
3) any technical problem requiring change to flight plan - dodgy engine/control in/out,dodgy fms (not that you'd necessarily notice these days), fuel,
4) any human problem requiring change to flight plan (drunk/hijacker/medical/politics of airspace, etc)
5) and even whilst in 'normal flight' above a few hundred (assuming all ATZ's globally are the epitome of predictability) anything in the air that isn't also automated under the same MO - GA, gliders, volcanoes, etc
3) and even then, before you taxi and after you 'run to the end' anything on the ground that isn't also automated under the same MO (see above about cars)


...and there's many more - in short, my view is total flight automation is viable BUT we're a long, long way from that. I can understand how experiencing 14hrs in flight with only half an hour under 'manual' seems like there's only a short distance to cover to automate the whole thing but (given we require near 100% success ) the last 10% (whether on ground or air) is actually very, very, very difficult.

In the meantime I think i'd prefer a pilot...

BM:hmm:

G-CPTN
12th Mar 2006, 22:33
Perhaps it would be easier to automate the passengers?

Jerricho
12th Mar 2006, 23:00
2) admin -........ or dodgy ATC

*Gasp*

That would never happen ;)