ual777
Posts: 1641
Joined: Thu Aug 14, 2003 6:18 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 12:49 am

An automated airliner would work fine in a vacuum with ILS's. However in the real world, there are countless threats that an airliner cannot see or respond to. Controller error, birds, objects on the runway, windshear, etc etc. Im not going to get dragged into this, but having a single pilot airliner is a STUPID idea. Having an unmanned airliner is even worse.
It is always darkest before the sun comes up.
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 1:37 am

Quoting ual777 (Reply 50):
However in the real world, there are countless threats that an airliner cannot see or respond to.

Because they haven't been designed to deal with those "countless threats." Other than perhaps cargo flights, there won't be any SP ops on any currently flying airliner.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
Mir
Posts: 19491
Joined: Mon Jan 05, 2004 3:55 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:06 am

Quoting DocLightning (Reply 49):
There was a mechanical failure that was completely recoverable.

Yet if that failure never happens, the accident never happens. While it's very attractive and very easy to focus on the last link in the chain to break, all of them are important.

-Mir
7 billion, one nation, imagination...it's a beautiful day
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:18 am

Quoting Mir (Reply 52):
While it's very attractive and very easy to focus on the last link in the chain to break, all of them are important.

And the first and the last links didn't have to "break".

Quoting DocLightning (Reply 46):
Then again, ten years ago, the idea of a computer driving a car was absurd. I was just behind one today. It drove quite well.

As noted earlier, Tesala will probably be the first to offer a self-driving option in a couple of years.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
Mir
Posts: 19491
Joined: Mon Jan 05, 2004 3:55 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:20 am

Quoting planemaker (Reply 53):
And the first and the last links didn't have to "break".

No, they didn't, but no system is perfect. It's just unfortunate that both the machines and humans failed in that situation.

-Mir
7 billion, one nation, imagination...it's a beautiful day
 
User avatar
N62NA
Posts: 4413
Joined: Sun Aug 10, 2003 1:05 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:29 am

Quoting spink (Reply 10):
From an infrastructure perspective, a lot more work is required on. The ATC and aviation infrastructure for the most part is still trapped back in the 60s and 70s.

And the aircraft too!

The 737MAX, A320NEO, 777, 787 and A350 will still be the predominant aircraft in the skies 20 / 30 / 40 years from now, so any new "pilotless" commercial aircraft won't be seen by most of us on this website - we'll all be dead by then!
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:47 am

Quoting Mir (Reply 54):
No, they didn't, but no system is perfect. It's just unfortunate that both the machines and humans failed in that situation.

In the "first" instance, the fault was known but the "fix" wasn't carried out as other aircraft in the fleet were "fixed" first. And, in the "last" instance, there were several solutions but I'll quote this one since Sully is so respected...

Quote:
"We have to infer angle of attack indirectly by referencing speed. That makes stall recognition and recovery that much more difficult. For more than half a century, we've had the capability to display AoA (in the cockpits of most jet transports), one of the most critical parameters, yet we choose not to do it."

As I have mentioned several times in past threads of this nature, aviation could be even safer but airlines only see the bottom line and literally gamble with peoples lives.

Quoting N62NA (Reply 55):
Quoting spink (Reply 10):From an infrastructure perspective, a lot more work is required on. The ATC and aviation infrastructure for the most part is still trapped back in the 60s and 70s.
And the aircraft too!

The 737MAX, A320NEO, 777, 787 and A350 will still be the predominant aircraft in the skies 20 / 30 / 40 years from now, so any new "pilotless" commercial aircraft won't be seen by most of us on this website - we'll all be dead by then!

First of all, ATC in NOT "trapped" back in the 60's and 70's. And there is no reason why updated those A/C won't be SP optional at some point in the next 20 years.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
User avatar
7BOEING7
Posts: 3039
Joined: Fri Oct 12, 2012 5:28 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 3:07 am

Quoting planemaker (Reply 56):
First of all, ATC in NOT "trapped" back in the 60's and 70's. And there is no reason why updated those A/C won't be SP optional at some point in the next 20 years.


  

The studies going on now by Airbus, Boeing, Fedex, UPS etc, are not based on an airplane that doesn't even exist yet, but on what's available today and will be in the freighter inventory 10-15 years from now. If it was as impossible as several people here suggest, I'm sure the $$$ being spent on these studies would be used elsewhere. These companies don't spend money without the expectation of some return/benefit.
 
spacecadet
Posts: 3484
Joined: Thu Sep 20, 2001 3:36 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 5:02 am

Quoting garpd (Reply 23):
Until a fully developed true to life Artificial Intelligence is ready to go, the whole idea is a very, very bad one.

Some argue that we're not really that far from this point: http://www.technewsdaily.com/17898-t...ogical-singularity-definition.html

According to them, we've got about 25 to 40 years...
I'm tired of being a wanna-be league bowler. I wanna be a league bowler!
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 5:28 am

Quoting spacecadet (Reply 58):
Quoting garpd (Reply 23):Until a fully developed true to life Artificial Intelligence is ready to go, the whole idea is a very, very bad one.

We don't need anywhere near to AGI for SP.

Quoting spacecadet (Reply 58):
According to them, we've got about 25 to 40 years...

And according to the most visible person involved with the singularity it will be achieved by the end of the 2020's.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
User avatar
flyingturtle
Posts: 5682
Joined: Mon Oct 31, 2011 1:39 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 9:45 am

Quoting hivue (Reply 44):
I think it can be argued that the AF447 crew did not earn their $$$ (or €€€).
Quoting Mir (Reply 48):
This is somewhat disingenuous. Yes there was pilot error, but there was also mechanical error as well. Pinning that one solely on the pilots isn't right.

AF447 is a very peculiar example. One case of unreliable airspeed resulted in hundreds of deaths, while about 30 other cases of unreliable airspeed were survived without superior piloting skills. In most of the 30 cases, the crew did not even apply the correct QRH procedure. No. Nothing.

An unreliable airspeed warning during stable cruise might even be a case of the computer doing more harm than a human pilot. The computer will pick up the UAS warning, act on it... always, flawlessly. But if reacting to it isn't even needed (as the 30 incidences show), the computer might do more harm.

Likewise, one could delegate medical diagnoses to a computer. The computer would flawlessly pick up raised creatinine and ALAT/ASAT levels... but it is needed? After strenuous bodily exercise, I can readily achieve a "Winner! You're sick!" prize in liver function tests.


David
Reading accident reports is what calms me down
 
User avatar
garpd
Posts: 2507
Joined: Thu Aug 11, 2005 9:29 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 10:25 am

Quoting flyingturtle (Reply 60):

Are you arguing then, that common sense is needed in a lot of situations, not just knowledge and the ability to diagnose and react? If so, I completely agree.

I do not think common sense can be programmed. Either you have it, or you don't. (As evident in a lot of humans!)
A powerful A.I. might be able to react quickly and compute what's needed to be done. But can we teach it the common sense it will need to read any given situation correctly? I'm sceptical.

AF447 is a weird case indeed, but if anything, it makes an argument for replacing pilots with computers, not the other way round.

IMO, the primary cause for that crash was the junior pilot pulling back on the stick for almost the entire time during the incident. IIRC, upon getting the error about unreliable air speed, the crew should have advanced the throttles to a known safe setting to maintain safe flying speed for the altitude and kept the wings and nose level then fly out of the problem. A computer would undoubtedly have followed that procedure.
arpdesign.wordpress.com
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 1:13 pm

What I notice is that nobody on the **Let's get rid of the pilots** crowd has answered any of my comments.
Are they so disturbing to you guys ?

To compare antilock for an old lady, automated cars is really - but really disingenuous : after all, automated subway trains have been in service in Paris, Lille, Singapore and elsewhere for nearly twenty yeras, without any mishap I could remember... on the other hand, I've seen networks standing still while human engineers correct the system fault (s). Like in Singapore, it could come from a lost chewing gum, which triggered their ban   
The same type of faults would have much more dramatic consequences on an airplane : it doesn't stand still in the air when the system sorts itself out.
By the same token, it is flagrant that the AF447 crew lost situational awareness. I posit that they weren't helped by the system indications : a stall warning sounding off when sticks are being pushed forward, a flight director still showing, way past the stall an order to pitch up... The system, even with the careful wording of the BEA report, is far from being blameless.

THE FUTURE ENVIRONMENT

This aspect is - of course - totally disregarded by my adversaries :
At the actual rate of growth, the traffic doubles every ten years or so, meaning that in thirty years'time, it will be some eight times busier than it is now. At half that rate, the increase would still reach nearly 300%.
Apart from the fact that these numbers make some good sense of the A380 sales prospects, we will be operating in an environment that is rather saturated : At the present time, a TCAS generated warning involves an average of three aircraft. Most of the times, it's due to conflicts that are well separated but with one departing airplane with a high rate of climb getting into the warning envelope of arriving flights. I wonder what will happen in a denser traffic...
In these conditions, an UAV, cargo or not will be part of a potentially dangerous system involving multiple conflicts.
One has to be very certain the system will cope.

Another aspect is that ALL arrivals will be done in a minimum-less operation, akin to a Cat IIIc autoland, this with every wind / gust value realistically forecasted. These approaches will be done using GPS-generated guidance. As for now, autolands have a limit of around 15 kt crosswinds, 30 kt headwinds and a GPS approach is regarded unsafe below 500 ft.
And then, on the ground, there would arise the need of resolving taxi conflicts : are you to go on automated systems or are you going for ATC ground guidance... where will be the transition between the two requirements and are you certain that your radio remote control would be totally safe, free of any interferences or criminal takeover ?

THE COSTS

The sort of airborne equipment you are considering hasn't really been considered in this thread.
Are you considering double, triple, quadruple... redundancy ? It's a very important factor as simple failures mean a degraded performance aircraft plying the crowded skies we talked about on the previous paragraph. Think about air traffic problems that will cause.
These costs would double / triple the avionics prices... on top of which you'd have to add maintenance, validation, certification-keeping costs. When you think about just how much keeping the ETOPS certificate is worth, I really do not see the economics of automatons vs flight deck salaries being so in favour of pilotless airliners.

PRESENT TRENDS

Right now, we can observe two trends, which are not mutually exclusive :
1/- More automation : that include new features of modern Airbus planes : the automatic *Brake to Vacate* which led to ROW / ROP ( runway overrun warning or prevention ), the automated TCAS resolution manoeuvre which includes Flight Director guidance instead of the normal TCAS red / green vertical speed advisory, the auttomated terrain avoidance...the windshear protection adopted on the 787... there are quite a few examples of this sort of solutions which IMHO improve air safety.
2/- And then you have a definite trend away from automatics : witness the use of HUDs, sometimes with enhanced vision, may be associated with a future virtual runway image-generation... leading to a more widespread use of manual Cat III landings, already the feature of some US airlines with an intelligent vision of the future, at a fraction of the costs of keeping a Cat III b auto-pilot current.

AND FINALLY,

This pilot-less argument strikes me as a series of badly-reasoned judgements of value, the main one being " the pilot is the weak link in a system of flight safety. Let's get rid of him. This furiously resembles the sharia law :* this hand has comitted theft, cut it off *.
Weak link ? Could be. However, I would posit that in most cases a factor has been about badly presented information or un-understanding of the situation presented. Just need to read any QRH to see that an announced failure means in reality problems which do not seem to be related to the presented warning.
And sometimes, engineers and test pilots forget that aging airplanes do not really behave like the tests have demonstrated or develop with age ailments that were not really understood :

In this respect, I would like to give two instances of unforeseen problems I had to deal with:
- The first one was the loss on a 737 of a generator : The check-list involved re-balancing the electrical loads and start the APU to use its generator if needed. That accomplished and with some time to think of what happened, I realised that, the electrical problem solved, the mechanical part - i.e the generator drive - had been left untouched... I disconnected the drive from the engine and had back at base a shouting match with the supervising engineer on duty: He claimed that the disconnection wasn't part of the check-list and did I know in may feeble pilot mind how long it will take to reconnect it ?... Three hours later, he called me on the phone to apologise as they discovered that the IDG was " a salad of rods and wheels and bearings" and a fire situation could well have happened. Three weeks later, Boeing amended the loss of one generator check-listto include a mandatory disconnection of the drive.
- Thje second one was on an A320 takeoff. passing 100 ft, a tune fork-shaped arc spead along the diagonal of my windscreen. No ECAM message, but a disturbing vision of bubvbling glass along the branches of the arc. I asked the then F/O to shut down the heater... no result and I shouted he cut the electrical power to my winscreen heater, saw his hand going to the right switch... no result, and I could see an inch wide bubling glass around the arcing.
I probably broke a world record of RTB, to realise after landing that the arcing had stopped... The F/O and I reviewed the whole event and we discovered that the switchlight involved was an ON / AUTO command. No OFF position.
A month later, Airbus issued a note including a *Electric Arcing On A Cockpit Window* check-list in which one has to remove electrical power to the affected window by pulling their respective cb, identified on their respective panel behind the pilots'seats.

There are countless of occurrences of this type... and then you would like us to blindly trust engineers' God-like infallability ?

Think again.

[Edited 2013-05-30 06:14:56]

[Edited 2013-05-30 06:24:53]
Contrail designer
 
AngMoh
Posts: 981
Joined: Fri Nov 04, 2011 5:03 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 1:50 pm

Quoting Pihero (Reply 62):

What I notice is that nobody on the **Let's get rid of the pilots** crowd has answered any of my comments.

I think you take this a bit personal. The questions you ask are based on todays technology, and today it is not possible.

I believe we see pilotless planes. Nothing to do if I think that is good or bad. I just believe that with the way technology is developing it is going to happen and it will happen sooner rather than later.

Today the technology is not available. I think that someone will fly a transatlantic flight automatically within the next 10 years. It will be a pure research event. There will be 2 pilots ready to grab the stick if it goes wrong. And I think it will be a converted Gulfstream or something similar. I am pretty sure that from first test flight to first commercial flight with passengers there will be at least 20 years.

And last and not least: I think the main driver for going for automated pilotless flights is going to be the density of air traffic. Nothing to do with safety and nothing to do with cost reduction or "get rid of the pilot". This density is going to drive the desire to take variation out of the process and that is going to drive the removal of humans. The same way as driverless trains are closer spaced than trains with drivers, the automated flights will be closer spaced than non-automated flights. Same way it has gone in manufacturing. Take out the variation and humans cause variation. And to get that done, the whole control and associated technology will be fundamentally different from what is being done today. It will be scary, but I believe it will happen, whether I like it or not.
727 732 733 734 735 73G 738 739 739ER 742 743 744 752 753 762 772 77E 773 77W 788 A300 A310 A319 A320 A321 A332 A333 A343 A345 A346 A359 A35K A388 DC-9 DC-10 MD11 MD81 MD82 MD87 F70 ERJ145 E170 E175 E190 E195 ATR72 Q400 CRJ200 CRJ700 CRJ900 BAE146 RJ85
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:09 pm

Quoting AngMoh (Reply 63):
I think you take this a bit personal. The questions you ask are based on todays technology, and today it is not possible.

No. The question I raised as an example has nothing to do with the state of technology, today's or tomorrow's. It's about a logic that , if not understood, can't be programmed, therefore soved.

Quoting AngMoh (Reply 63):
Take out the variation and humans cause variation. And to get that done, the whole control and associated technology will be fundamentally different from what is being done

I disagree. With or without pilots, today's FBW will require an even greater level of redundancy. All the laws and protections will be ssentially the same.

Quoting AngMoh (Reply 63):
It will be scary, but I believe it will happen, whether I like it or not.

If it is scary, therefore it won't be done : Scared people don't make good customers.
Contrail designer
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:10 pm

Quoting AngMoh (Reply 63):
I think you take this a bit personal. The questions you ask are based on todays technology, and today it is not possible.

No. The question I raised as an example has nothing to do with the state of technology, today's or tomorrow's. It's about a logic that , if not understood, can't be programmed, therefore solved.

Quoting AngMoh (Reply 63):
Take out the variation and humans cause variation. And to get that done, the whole control and associated technology will be fundamentally different from what is being done

I disagree. With or without pilots, today's FBW will require an even greater level of redundancy. All the laws and protections will be essentially the same.

Quoting AngMoh (Reply 63):
It will be scary, but I believe it will happen, whether I like it or not.

If it is scary, therefore it won't be done : Scared people don't make good customers.
Contrail designer
 
AngMoh
Posts: 981
Joined: Fri Nov 04, 2011 5:03 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 2:29 pm

Quoting Pihero (Reply 65):
It's about a logic that , if not understood, can't be programmed, therefore solved.

It will be understood. You probably don't like it and can't accept it, but people who will decide will do it.

Quoting Pihero (Reply 65):
I disagree. With or without pilots, today's FBW will require an even greater level of redundancy. All the laws and protections will be essentially the same.

The redundancy will increase and take out the pilots.

Quoting Pihero (Reply 65):
If it is scary, therefore it won't be done : Scared people don't make good customers.

The whole point is that the passengers won't be scared. They won't even know. In the beginning there will be an automated system and the pilots will be there just looking at the screens and they will hope things go wrong to prove that they are still needed. Those people implementing and developing it will be scared. That is the part you don't understand about engineers. They are awake at night wondering what they could have overlooked if they put a system like this in production. When I was still a software engineer I did worry every time I release a safety critical software to production about what I could have missed. And the whole team I worked with was the same. You just don't have a clue what is engineering all about. Refer to your quote...

Quoting Pihero (Reply 62):
here are countless of occurrences of this type... and then you would like us to blindly trust engineers' God-like infallability ?

Engineers are the one least likely to believe that they are infallible. It is not engineers who are going to drive this. It is going to be cold hard businessmen who think they can increase profit margins. Some MBA will come up with a business case and the engineering team will be whacked into delivering it - even if the engineers think the technology is not ready.
727 732 733 734 735 73G 738 739 739ER 742 743 744 752 753 762 772 77E 773 77W 788 A300 A310 A319 A320 A321 A332 A333 A343 A345 A346 A359 A35K A388 DC-9 DC-10 MD11 MD81 MD82 MD87 F70 ERJ145 E170 E175 E190 E195 ATR72 Q400 CRJ200 CRJ700 CRJ900 BAE146 RJ85
 
User avatar
seat55a
Posts: 220
Joined: Thu Jan 17, 2013 7:18 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 3:56 pm

The OP started with the recent BA cowl departure event. This seems likely to be the result of an avoidable human failure in maintenance procedures.

For this thread let's assume that anyway, since such failures do occur.

My question is, when will maintenance crew be replaced by automation that won't forget the obvious?

Surely robot eyes are better than maintenance tech eyes? And once those robots are fixing the planes then some of the issues of "unknown unknowns" in the air will go away also.
 
ual777
Posts: 1641
Joined: Thu Aug 14, 2003 6:18 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 3:58 pm

Quoting AngMoh (Reply 66):

It will be understood. You probably don't like it and can't accept it, but people who will decide will do it.
Quoting AngMoh (Reply 66):

Engineers are the one least likely to believe that they are infallible. It is not engineers who are going to drive this. It is going to be cold hard businessmen who think they can increase profit margins. Some MBA will come up with a business case and the engineering team will be whacked into delivering it - even if the engineers think the technology is not ready.

The problem will be getting regulatory approval.

The primary reason flying is so safe these days is solid CRM training and situational AQP training. Having TWO sets of eyes, TWO brains, and TWO well-trained pilots communicating with each other in the cockpit is the biggest and best safety measure in a flight operations environment. The whole concept is about communicating what you see, sharing ideas between the two pilots, and trapping errors....and it WORKS. If you look at almost every crash attributed to pilot error, you will see that poor CRM was the underlying cause.

I have literally been receiving vectors on approach to a busy airport at 5:30am, with an MEL'd autopilot, and the EGPWS went absolutely nuts. I was hand flying the aircraft and was receiving "Landing Gear, Landing Gear", "Too Low Terrain", "Pull Up, Pull Up", on a segment of the flight with no auto pilot that was VERY busy. The fun part is that the warnings would not shut off. Having the other pilot was CRUCIAL to workload management while I focused on flying the aircraft. The captain was running the QRH, yanking circut breakers, running the approach checklists, and communicating with ATC all while I was flying.

I can relate story after story that proves that two pilots are needed. Single/no pilot ops are a BAD idea, and if implemented people will die.

I won't breakout tons of other reasons or stories simply because this thread is like herpes, in that it comes back twice a year, but many of you on the pro automation side are simply looking at things from the wrong perspective.
It is always darkest before the sun comes up.
 
User avatar
7BOEING7
Posts: 3039
Joined: Fri Oct 12, 2012 5:28 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 4:17 pm

Quoting Pihero (Reply 62):
The first one was the loss on a 737 of a generator : The check-list involved re-balancing the electrical loads and start the APU to use its generator if needed. That accomplished and with some time to think of what happened, I realised that, the electrical problem solved, the mechanical part - i.e the generator drive - had been left untouched... I disconnected the drive from the engine and had back at base a shouting match with the supervising engineer on duty: He claimed that the disconnection wasn't part of the check-list and did I know in may feeble pilot mind how long it will take to reconnect it ?... Three hours later, he called me on the phone to apologise as they discovered that the IDG was " a salad of rods and wheels and bearings" and a fire situation could well have happened. Three weeks later, Boeing amended the loss of one generator check-listto include a mandatory disconnection of the drive

All I can assume was this was prior to 1978 (near the dawn of the jet age) because I don't remember ever having to rebalance loads or disconnect the CSD (unless I had a CSD indication of some sort). As an "ancient aviator" myself I understand your reluctance to change and neither of us will live long enough to get on a pilotless commercial airplane but I'm betting we'll both see SPCO (single pilot cargo ops). (My apologies if the 737 incident happened since my retirement about 2 years ago.)

Quoting AngMoh (Reply 63):
And last and not least: I think the main driver for going for automated pilotless flights is going to be the density of air traffic. Nothing to do with safety and nothing to do with cost reduction or "get rid of the pilot". This density is going to drive the desire to take variation out of the process and that is going to drive the removal of humans. The same way as driverless trains are closer spaced than trains with drivers, the automated flights will be closer spaced than non-automated flights. Same way it has gone in manufacturing. Take out the variation and humans cause variation. And to get that done, the whole control and associated technology will be fundamentally different from what is being done today. It will be scary, but I believe it will happen, whether I like it or not.

  
 
User avatar
7BOEING7
Posts: 3039
Joined: Fri Oct 12, 2012 5:28 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 4:29 pm

Quoting ual777 (Reply 68):
I can relate story after story that proves that two pilots are needed. Single/no pilot ops are a BAD idea, and if implemented people will die.

People are already dying and the future won't be pretty -- more airplanes, more congestion and less experienced pilots.

Quoting ual777 (Reply 68):
I won't breakout tons of other reasons or stories simply because this thread is like herpes, in that it comes back twice a year, but many of you on the pro automation side are simply looking at things from the wrong perspective.

That's what the fireman on the train and the flight engineer you no longer fly with said. If you're in cargo, say goodbye to your copilot -- it will happen on your watch.
 
hivue
Posts: 1957
Joined: Tue Feb 26, 2013 2:26 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 4:41 pm

Quoting garpd (Reply 61):
I do not think common sense can be programmed. Either you have it, or you don't. (As evident in a lot of humans!)

Common sense is the least common sense. 
Quoting Pihero (Reply 62):
At the actual rate of growth, the traffic doubles every ten years or so, meaning that in thirty years'time, it will be some eight times busier than it is now. At half that rate, the increase would still reach nearly 300%.

Hasn't a lot of aircraft automation been driven by precisely this? I believe in most places in the world A/P use is mandated above a certain FL since it can do a better job at maintaining RVSM over long flights than a pilot.
"You're sitting. In a chair. In the SKY!!" ~ Louis C.K.
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 6:41 pm

Compare this :

Quoting AngMoh (Reply 66):
Engineers are the one least likely to believe that they are infallible.

and this :

Quoting AngMoh (Reply 66):
It will be understood. You probably don't like it and can't accept it, but people who will decide will do it.

... and tell me that I'm wrong. That kind of self-confidence, arrogant self-assuredness in one's belief that they can solve anything is, as a matter of fact the most frightening aspect of this whole thread.
"it-will-be-understood"... WOOOOW !

Quoting seat55a (Reply 67):
My question is, when will maintenance crew be replaced by automation that won't forget the obvious?

Answer is : " When the obvious will not be dismissed by engineers"

Quoting 7BOEING7 (Reply 69):
All I can assume was this was prior to 1978 (near the dawn of the jet age

My bad : t'was *load shedding* iso *balancing*. As you seem to have access to the manufacturer's documents, see when the QRH was amended on this item. Should be end of 1982.

Quoting hivue (Reply 71):
Hasn't a lot of aircraft automation been driven by precisely this? I believe in most places in the world A/P use is mandated above a certain FL since it can do a better job at maintaining RVSM over long flights than a pilot.

Yes and no : altitude-keeping autopilots have been there since the - as 7BOEING7 says - "the dawn of the jet age". We went into a higher degree of precision which already happened before we even considered reduced vertical separation . And the pilots have acknowledged that in cruise, George is better - lots better - at keeping the passengers'dinner in their stomachs and not on their neighbours' lap.
Another aspect is navigation precision which allows flights tighter to their ideal path, therefore reducing flight times ---> therefore reducing costs. It's mainly an after effect of economics-led progress.
Someone is of course going to say that TCAS allows safer flights in the same airspace ; I posit that TCAS resolution manoeuvres have reached a point that they could be bringing more confusion into the general picture : Imagine a tight bunch of aircraft piled on top of each other and one reacting to a resolution warning - spurious or not - the cascade effect could be quite spectacular.
Contrail designer
 
spink
Posts: 317
Joined: Tue Aug 09, 2005 5:58 pm

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 7:09 pm

Quoting seat55a (Reply 67):
My question is, when will maintenance crew be replaced by automation that won't forget the obvious?

Surely robot eyes are better than maintenance tech eyes? And once those robots are fixing the planes then some of the issues of "unknown unknowns" in the air will go away also.

This is largely in progress. Each new airframe model has significantly increased self detection and reporting and even within an airframe model they continuously add more self detection and reporting.

There are some industries where they are adding permanent visual inspection capabilities into designs via small cameras so that they can continuously compare visual status of equipment and catch faults and faulty repairs.

The actual human maintenance crew is unlikely be replaced by robots due to cost issues any time soon (actual robots to crawl around the planes would be rather expensive compared to human labor to replace/install parts) but like with the pilots, they will generally be doing significantly less work and the automated and remote systems will be taking up a large bulk of the actual checking.
 
hivue
Posts: 1957
Joined: Tue Feb 26, 2013 2:26 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 7:19 pm

Quoting Pihero (Reply 72):
I posit that TCAS resolution manoeuvres have reached a point that they could be bringing more confusion into the general picture : Imagine a tight bunch of aircraft piled on top of each other and one reacting to a resolution warning - spurious or not - the cascade effect could be quite spectacular.

As I understand it TCAS merely makes suggestions (although forceful ones). It's up to pilots to take action. If we do see computers displacing pilots, the first sign may well be for flight in congested terminal areas if the certificating authorities become convinced that computers can handle complex conflict situations much better than humans.
"You're sitting. In a chair. In the SKY!!" ~ Louis C.K.
 
User avatar
garpd
Posts: 2507
Joined: Thu Aug 11, 2005 9:29 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 8:22 pm

Quoting hivue (Reply 71):
Common sense is the least common sense.

Touché    
arpdesign.wordpress.com
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 8:30 pm

Quoting hivue (Reply 74):
the first sign may well be for flight in congested terminal areas if the certificating authorities become convinced that computers can handle complex conflict situations much better than humans.

If you're ready to trust your family's and friends' lives on that assumption, you're a far more open-minded man than I am.
Unfortunately, I see how the system work everyday and in these congested airspaces, I'd rather work with a human controller than an automated system. I've seen transponder - on which TCAS is based - failures and the possibility increases with a growing number of aircraft in the same airspace, to be rather doubtful of the total safety of an automated system. ( By the way, how does the automaton keep separation, resolve diversion in this case ? )

Quoting hivue (Reply 74):
As I understand it TCAS merely makes suggestions (although forceful ones). It's up to pilots to take action.

Without the prior notice of a switched-on controller you obey a TCAS resolution . Your life and your passengers' depend on it ; it would be foolish to second-guess the system with the limited vision we have, especially at night.
Contrail designer
 
hivue
Posts: 1957
Joined: Tue Feb 26, 2013 2:26 am

RE: Planes Can Fly Themselves. BS!

Thu May 30, 2013 9:02 pm

Quoting Pihero (Reply 76):
If you're ready to trust your family's and friends' lives on that assumption, you're a far more open-minded man than I am.

No, I am not ready for that. But then I was not ready for airbags and automatic external defibrillators either (if you had told me 30 years ago, when I worked with the the computerized EKG monitoring of that time, that some day computers would be allowed to decide whether or not to shock a heart attack victim I would have told you that you were nuts). The human vs automation debate is very interesting to me, though.



Quoting Pihero (Reply 76):
I've seen transponder - on which TCAS is based - failures and the possibility increases with a growing number of aircraft in the same airspace, to be rather doubtful of the total safety of an automated system. ( By the way, how does the automaton keep separation, resolve diversion in this case ? )

I agree. The potential for hardware failures may be what keeps humans in the cockpit (and control centers and towers) no matter how automated things get. Humans may do weird things but, unless they die, they don't just up and quit.
"You're sitting. In a chair. In the SKY!!" ~ Louis C.K.
 
2175301
Posts: 1576
Joined: Wed May 16, 2007 11:19 am

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 3:37 am

Quoting Pihero (Reply 62):
What I notice is that nobody on the **Let's get rid of the pilots** crowd has answered any of my comments.
Are they so disturbing to you guys ?

I did not previously respond because I believe you miss my points just as much as you mistakenly believe I miss yours.

I am the one who posted about the antilock brake analogy to demonstrate that this IS NOT - and in the business world will never be a debate about how sometimes pilots can save a plane that the automated control cannot. This real issue is when will automated controls on average produce less crashes than pilots.

I agree that the technological infrastructure does not yet exist for these systems, and that triple redundant systems with new aircraft will be needed - but I am convinced they are coming. This has been the pattern of technology for over 100 years. Everything from elevator operators to new nuclear plant operators have been displaced by technological solutions that worked better than humans (and don't try to tell me that nuclear power plants are not very complicated and without hundreds (if not thousands) of problem scenarios [Note that the operator staffing levels at the new generation nuclear power plants is minimal compared to my plant designed in the late 1960's - and want to start a new AP-1000 from cold shutdown once the mechanics release it to the operators: Just press the start button and stand back and watch and monitor for a few days while the plant starts system by system and comes online and ramps up to full power (it takes a few days to start up any major sized steam cycle power plant). For every failure that occurs on start-up a redundant system takes over; and virtually all problems can be fixed while the plant is starting up or running].

Too complicated you say, Too costly, Only "Pilots" can save the day, etc. Those are all the refrains (just change the job title) from hundreds of other "Professional Operators" of one sort or another that have been displaced by technology in the last 50+ years. Please study history - and then come back and tell me why - really why based on real world technology improvement history - that Pilots will be exempted from this historical process baring major earth catastrophe that stops technological innovation.

I have lived some of that history - and in about 1990 was one of about 50,000 degreed engineers who was cut out of the Power Generation industry within the US alone in about a 5 year period due to technological change in control and instrument technologies. I remember well how we (and I) used to say you cannot run a power plant without us. I learned otherwise - the hard way. A decade later I was very fortunate to get back into the industry. The modern power plant control room and staffing look nothing like it did even in the 1980's. A whole series of new fossil fueled power plants don't even have anyone normally staffing the control room. They are normally started, operated at various power levels, and stopped remotely. Once a day someone will typically check on the plant in person. If the plant is large enough it will have a day shift maintenance staff. But smaller plants don't even have any assigned staff maintenance - just roving maintenance teams that show up when needed.


I agree that flying a plane involves many variables - and that sometimes the automated controls will even make the situation worse. But, that does not mean that the automated controls are not coming - and that they will at some point perform better on average than a pilot. There is a lot of money being spent on development of those controls because of the potential savings - and they will succeed at some point.

Have a great day,
 
User avatar
mayor
Posts: 6218
Joined: Sun Mar 09, 2008 3:58 pm

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 4:19 am

Remember the old joke about the "cockpit of the future"? It would have a dog and a pilot in it. The pilot is there to keep an eye on things and WHY is the dog there? To bite the pilot if he tried to touch anything.  
"A committee is a group of the unprepared, appointed by the unwilling, to do the unnecessary"----Fred Allen
 
User avatar
flyingturtle
Posts: 5682
Joined: Mon Oct 31, 2011 1:39 pm

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 8:49 am

Quoting 2175301 (Reply 78):

I'm no nuclear power engineer, and so I can't really say anything. But I have the faint feeling that aircraft systems are much more complex, many more variables have to be dealt with. After the fact we can program a computer to deal with previously "unforeseen" stuff like QF32 with the uncontained engine failure.

Or something like a spurious radar altimeter and a malfunctioning EGPWS.

And perhaps we should ask how many of the "planes can fly themselves" fanboys are actually Boeing fans, and poke fun at how computerized the Airbus is.

Perhaps the best reasons to have a human pilot in the front office are:
1. A pilot can remember accident reports. He's completely free to spend his whole leisure reading them, he can continuoulsy update his expertise. If any system of the aircraft goes drunk, he can remember them and act appropriately. With a computer, we would have to wait for a) programmers to devise ways to deal with certain scenarios, and b) regulators to certify these changes.

In countless things of your own lifes, you're doing the very same. Constantly updating what you know, and be prepared better than any computer can.

Plain old experience is a sharper knife than a committee of programmers who quarrel about which failure modes to integrate in their computer pilot.

2. Human pilots fight for their lives, which automatically translates into fighting for their passengers' lives.

With a computer flying the aircraft, it is just a cold machine that reacts to inputs. It's up to you who you would want to trust.

3. Very far-fetched: Maliciously, one could even program such a computer to kill all the passengers. With human pilots, this has happened in the Egypt Air suicide flight over the Atlantic before, but with computers, there won't be anybody in the cockpit trying to stop a "suicidal" computer.


David
Reading accident reports is what calms me down
 
User avatar
airmagnac
Posts: 451
Joined: Wed Apr 18, 2012 10:24 pm

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 9:32 am

Having just gone through the 80 posts of the discussion, I find it interesting that the various posters are discussing many completely different things. Some people are discussing single-pilot ops, some talk about remotely-piloted aircraft, others completely pilotless machines. There is talk about FBW, auto-pilots or long-term decision-making, examples based on driverless trains or power plants vs aircraft. Some are adamant the technology “is here”, others say it’s not ; there is an argument that machines should be intelligent, which is countered by “artificial intelligence exists”...

But no one addresses what seems to me to be the first order of business. Before discussing about moving or replacing pilots by automation, shouldn’t we start by agreeing on what exactly is a pilot ? And what are his roles on an airliner ?
Then we could move on to investigating the pros and cons of human operators, those of automatic controllers, what type of “intelligence” is required to drive an airliner, and finally discuss whether or not automatics might be advantageous over humans in the previously discussed roles, in different use cases :
-single pilot ops : 1 human augmented by advanced avionics.
- remote-pilot ops : aka “drones”. A human is still required, he is just not physically in the aircraft.
- full automatic ops : no human pilot anywhere.

If no common frame is set first, the discussion just goes all over the place.

I have my own detailed idea of what a pilot is, but to leave the discussion open (and as I have little time  ) I’ll just say that IMO the pilot is in charge of making decision at various scales of time & space. From the very localized and very short-term (moving the control surfaces) to long term decisions regarding navigation over hundreds of kilometres (or miles, my American friends) over several hours.
The smaller the scale, the more the decisions are simple “reactions” to observed events, for which dedicated computers are well suited.
The bigger the scale, the more “prediction capabilities” are required, as is the capacity to combine different types of data to reach a decision (see Pihero’s example in post 42). This is where inductive reasoning capabilities (let’s call it imagination) is useful, and where human brains are better.
Sooooo :
-single pilot ops ? Maybe... Human factor and human/machine interface considerations will be important, and increasingly so the longer the flight is (see Pihero’s remark on fatigue). So depending on our knowledge on human factors, I could see SP ops building up from small planes to RJs to short-haul. Maybe eventually to long-haul. But it’ll take time.
- remote-pilot ops ? The reliability of the comlink will be key, as it will have to be considered as a critical failure, and so have a failure rate of less than 10^-5, maybe even 10^-7 or 10^-9. Think we can achieve that ? More precisely : can we achieve that for a cost low enough to justify removing the pilot from the airplane ?
- full automatic ops ? As I said before, not possible until machines have some form of “imagination”

Quoting AngMoh (Reply 63):
This density is going to drive the desire to take variation out of the process and that is going to drive the removal of humans. The same way as driverless trains are closer spaced than trains with drivers, the automated flights will be closer spaced than non-automated flights. Same way it has gone in manufacturing

There is a difference. If the external operating conditions do not vary much, your system does not need to vary much either. Hence variability of the controller is not a valuable characteristic, and it can be relaxed to achieve better efficiency.

All the driverless-trains that I know of are restricted to airport movers or urban metro systems. They evolve in an environment that does not vary a whole lot, apart from temperature and humidity, but those would not have a great impact if the system is well designed. And if the external conditions very too much, you just shut everything down and wait for normal operating conditions. Same for manufacturing, it usually takes place in non-variable environments.

But do I have to explain that airplanes evolve in highly variable environments ? And it’s not just the weather. Think about medical emergencies, ground infrastructure issues, comm blackouts, etc... Capability/Incapability of on-board systems to reliably detect and report problems is also a big source of variability ; this was a major issue for AF447, and that accident shows that there is still much progress to be made. Last point : you can't just shut down your airplane and stop if you don't know what to do...So in the case of aircraft, variability of the controller (aka flexibility or adaptability) is useful. So the human element is desirable.
My goal as an engineer is to fill my soul with coffee and become immortal
 
hivue
Posts: 1957
Joined: Tue Feb 26, 2013 2:26 am

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 2:22 pm

Quoting airmagnac (Reply 81):
- remote-pilot ops : aka “drones”. A human is still required, he is just not physically in the aircraft.
- full automatic ops : no human pilot anywhere.

I think this is a false dichotomy. My suspicion is that military UAV missions are flown entirely by computer with systems engineer "pilots" monitoring. If something starts to go a little bit wrong, it's probably rectified with a keyboard rather than a flight sim joystick and rudder pedals. If we ever see "full automatic commercial ops" there will still be these human systems engineer "pilots" monitoring.

Quoting airmagnac (Reply 81):
Capability/Incapability of on-board systems to reliably detect and report problems is also a big source of variability ; this was a major issue for AF447, and that accident shows that there is still much progress to be made

In the case of AF447 I don't think there was any significant issue with the automated systems detecting problems and acting appropriately. The issue was with the crew being unable (for whatever reason) to follow correct SOPs. This is the sort of situation that, if it were to become common, could result in no more pilots on the flight deck.

[Edited 2013-05-31 07:29:45]
"You're sitting. In a chair. In the SKY!!" ~ Louis C.K.
 
User avatar
flyingturtle
Posts: 5682
Joined: Mon Oct 31, 2011 1:39 pm

RE: Planes Can Fly Themselves. BS!

Fri May 31, 2013 5:36 pm

Quoting hivue (Reply 82):
If we ever see "full automatic commercial ops" there will still be these human systems engineer "pilots" monitoring.

I have serious problems with that approach. Because then we will definitively have "office" pilots that are called to duty only when things go wrong. And I fear that they won't get a feeling for the seriousness of the situation soon enough when they could chatter and sip a coffee just three minutes before...

David
Reading accident reports is what calms me down
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 12:44 am

Quoting flyingturtle (Reply 80):
1. A pilot can remember accident reports. He's completely free to spend his whole leisure reading them, he can continuoulsy update his expertise. If any system of the aircraft goes drunk, he can remember them and act appropriately. With a computer, we would have to wait for a) programmers to devise ways to deal with certain scenarios, and b) regulators to certify these changes.

Absolutely no insult intended, as you have contributed thoughtful viewpoints but, which... unfortunately, are erroneous.

There are several examples but "Watson's" a$$kicking of the top two Jeopardy! winners is the best known (but isn't really isn't known... and much less understood and appreciated or there wouldn't be anyone on here thinking that we might have SP Ops in 50 years... or, worse, never!)

Since Watson's debut on Jeopardy! it has gone to work as an oncologist... and has already achieved a 90% success rate in lung cancer diagnosis compared to 50% for human oncologists. That is TODAY! Not next year, let alone 10 years from now. And that kind of processing power will be available on your Google Glass or Apple "iWrist" in a few years.

How did Watson do this? By reading over 2 million pages from medical journals, +600,000 pieces of medical evidence, all national guidelines, hospital best practises, clinical trials and text books, and searching up to 1.5 million patient records, for information.

There will sooner, rather than later, be a "Watson 2.0 " for aviation.... obviously - and it will replace the 2nd crewmember.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
Pihero
Posts: 4318
Joined: Mon Jan 31, 2005 5:11 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 12:50 am

Quoting hivue (Reply 82):
My suspicion is that military UAV missions are flown entirely by computer with systems engineer "pilots" monitoring. If something starts to go a little bit wrong, it's probably rectified with a keyboard rather than a flight sim joystick and rudder pedals.

That is not correct. There could be phases of automatic pilot but the job is basically not different from piloting an airfcraft.
And they use a joystick as this article and pic show :
Drone Pilots

Quoting hivue (Reply 82):
If we ever see "full automatic commercial ops" there will still be these human systems engineer "pilots" monitoring.

... which brings us back the comlink problems which Airmagnac and I were referring to...

Quoting hivue (Reply 82):
In the case of AF447 I don't think there was any significant issue with the automated systems detecting problems and acting appropriately

No ? They just packed up when they lost data inputs...

Quoting 2175301 (Reply 78):
I have lived some of that history - and in about 1990 was one of about 50,000 degreed engineers who was cut out of the Power Generation industry within the US alone in about a 5 year period due to technological change in control and instrument technologies. I remember well how we (and I) used to say you cannot run a power plant without us. I learned otherwise - the hard way. A decade later I was very fortunate to get back into the industry. The modern power plant control room and staffing look nothing like it did even in the 1980's.

... and yet, we have the Fukushima accident, in which engineers couldn't forecast a series of uncontained failures which released quite an amount of radioactive mlaterial into the atmosphere and the sea.
And I daresay that the level of complexity in that system is in no way comparable to an airliner's.

Quoting 2175301 (Reply 78):
Too complicated you say, Too costly, Only "Pilots" can save the day, etc. Those are all the refrains (just change the job title) from hundreds of other "Professional Operators" of one sort or another that have been displaced by technology in the last 50 years. Please study history - and then come back and tell me why - really why based on real world technology improvement history - that Pilots will be exempted from this historical process baring major earth catastrophe that stops technological innovation.

I have been for years studying aviation history and the technology progress made.
Let's share a moment what progress did inside an aircraft :
First, we have to define the different functions on a flight deck:
They are :

1/- Command

2/- Piloting, aka short term trajectory control

3/- Systems management

4/- Navigation

5:- Communications

I didn't invent anything ; that's the way we present our functions and duties.
Now let's take them one by one :

5/- Communications : with the avent of generalised VHF stations and relays / SatCom and ACARS / ADS-B there is no need for a radio-operator any more. The function is now fulfilled by the pilots

4/- A navigator was a highly specialised officer but at the dawn of jet airliners navigation went slowly to "Skipper, this is where we were half an hour ago and with this new heading I'm giving you, this is where we should be in half an hour's time" to "We are 37.831 Nm to the runway 13". Exit the navigator . The function - mainly one of monitoring - is now a pilot's one.

3/- Systems management : the functions of a flight engineer have been in most of ways automated... Sometimes some alert crew - pilots - could act to rectify a situation, unseen by the automaton which presents grave danger : an example is a 744 taking off from Narita which encountered of transfer pumps from the tail tank. The system - therefore the engineers whop designed it - didn't see any proiblem, until a human realised that with emptying wing tanks, the CoG would be out of the aft range with the result of an unpilotable airplane. The consequence was RTB, overweight landing...
BA 038 is another instance of an event that neither the system - or the engineers who designed it - could forecast, therefore plan for. It needed some really fancy waltzing around the aerodynamics of the T7 to save the day... You have countless examples of that sort of event... which make seriously distrust the engineers' assertion that they will know all and plan for all of them.
Once again as these two instances show, that systems management function - although most of the time on monitoring - is now on the pilots'shoulders

2/- Piloting : As defined above, it involves short term alterations to the flight path... and yes, computers can do that ;
which brings us to :

1/- Command : It involves a set of - generally - medium to long term assessment of the flight progress and the *strategic* decisions which result. They include results of the flight planning ( Level changes, modification of the planned cruise schedule,... etc... ) and in normal situations, it can be done by the FMS. No argument here.
My first concern is that this aircraft is not alone in the sky. What happens when it reaches TOD ? and who, apart from a ground controller / operator will confirm the descent clearance : back once again to the comlink problem.
Now consider any event outside the technical field : medical emergency / bomb on board / unruly passengers and fisticuffs all abnormal situations for which the computer has no sens (ing). I'm curious as to the solutions in a pilotless aircraft.
Now consider technical faults, of which electrical power is one of the weakest : No power, no automatic captain ; Is the solution more redundancy to the system, i.e. returning power to a shorted equipment ? Good luck !
The there are *minor* failures that don't even translate into an ECAM message : failure of the toilets system flushing when you've departed for a 12 hr flight... It could be an interesting re-definition of being in deep matter   

Sorry for this long post. I hope I've managed to pass my concerns.
One last comment : progress in airborne automatons hasn't been parallel to ATC improvements, contrarily to some assertions here. Yesterday, a colleague of mine was descending ionto CDG for a 26L downwind.... Traffic Traffic, followed by "Climb ! Climb !". As per procedures, the F/O announced " De Gaulle, XXX TCAS clmimb ", to which the ATCO urgently said " XXX, descend immediately to flight level 70 !"... quite shaken, our colleague told us that in his 25 tears of flying, that was the first time he heard "immediately" ( the only instruction we have to obey without delay ). He told us he never saw what triggered the resolution or which traffic he came into conflict during his climb.. Quite a dilemma for George, hey ? Plus the fact that I don't see how quicker he would have responded to the ATCO's instruction... Food for thought, anyone ?

[Edited 2013-05-31 18:41:23]
Contrail designer
 
AngMoh
Posts: 981
Joined: Fri Nov 04, 2011 5:03 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 1:07 am

Quoting Pihero (Reply 85):
I didn't invent anything ; that's the way we present our functions and duties.

And the key thing is that are the duties today.

Quoting Pihero (Reply 85):
5/- Communications : with the avent of generalised VHF stations and relays / SatCom and ACARS / ADS-B there is no need for a radio-operator any more. The function is now fulfilled by the pilots

And you give a good example where duties change over time.

The pilots duties will change. How, I don't know. Maybe the first step to automated flights will be pilots serving drinks at the meal service and only be on active duty during take-off, landing and emergencies
  
727 732 733 734 735 73G 738 739 739ER 742 743 744 752 753 762 772 77E 773 77W 788 A300 A310 A319 A320 A321 A332 A333 A343 A345 A346 A359 A35K A388 DC-9 DC-10 MD11 MD81 MD82 MD87 F70 ERJ145 E170 E175 E190 E195 ATR72 Q400 CRJ200 CRJ700 CRJ900 BAE146 RJ85
 
spink
Posts: 317
Joined: Tue Aug 09, 2005 5:58 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 2:03 am

Quoting Pihero (Reply 85):
That is not correct. There could be phases of automatic pilot but the job is basically not different from piloting an airfcraft.
And they use a joystick as this article and pic show :
Drone Pilots

Both the X-47 and X-45 operate primarily autonomously. They have fallback remote pilot capabilities but that is only used as a backup. Normal flight ops are done via navigational instructions to the airframe which actually pilots the plane including TO/landing and formation flying.
 
hivue
Posts: 1957
Joined: Tue Feb 26, 2013 2:26 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 2:07 am

Quoting Pihero (Reply 85):
Quoting hivue (Reply 82):In the case of AF447 I don't think there was any significant issue with the automated systems detecting problems and acting appropriately
No ? They just packed up when they lost data inputs...

Yes. That was the appropriate action. It's what the autmated systems were programmed to do. It's interesting that the flight was progressing well under the control of the automated systems. Then an air data device invented before the Wright Brothers flew failed (due to external causes) and the humans were not able to cope with the resulting (confusing) situation. And I think one of the main reasons they failed to cope was because up to that point they had almost nothing to do but monitor the systems. If pilots ever are eliminated I don't think it will be due to a single decision. They will get elbowed out by "creeping automation."
"You're sitting. In a chair. In the SKY!!" ~ Louis C.K.
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 2:32 am

Quoting Pihero (Reply 85):
... and yet, we have the Fukushima accident, in which engineers couldn't forecast a series of uncontained failures which released quite an amount of radioactive mlaterial into the atmosphere and the sea.

Yes, an accident that occurred in a reactor that was designed over 50 YEARS AGO ago and built 45 YEARS AGO - and the very first built by TEPCO... when SP ops are something that will start to happen 7-10 years in the FUTURE.

Quoting Pihero (Reply 85):
... which brings us back the comlink problems which Airmagnac and I were referring to...

Aside from the fact that comlinks will be better (and, obviously, with backups)... if there were ever to be a total comlink "problem" the aircraft would simply continue to the destination airport.

Quoting Pihero (Reply 85):
I have been for years studying aviation history and the technology progress made.
...

My first concern is that this aircraft is not alone in the sky.

Then surely you would be aware of just some of the many projects by OEM's and regulatory agencies. Some of which were even reported not just in the aviation press.

Quoting Pihero (Reply 85):
Now consider technical faults, of which electrical power is one of the weakest : No power, no automatic captain ; Is the solution more redundancy to the system, i.e. returning power to a shorted equipment ? Good luck !

APU, batteries, RAT, etc, and that is now.

Quoting Pihero (Reply 85):
Food for thought, anyone ?

Yes, it is called 4D flight trajectories and sense & avoid.
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
Mir
Posts: 19491
Joined: Mon Jan 05, 2004 3:55 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 3:23 am

Quoting AngMoh (Reply 86):
And the key thing is that are the duties today.

Those will always be the duties. Those things must get done in any flight, whether directly by the crew or via automation (for which the crew is still responsible).

The issue that will have to be overcome in order to have a safe single-pilot cockpit is not the automation of the flying, but rather the automation of all the other tasks - the running of checklists, the coordination with the FAs, dispatchers and ATC, and the decision-making process. And those are a lot more difficult than just automating the flying of the airplane.

-Mir
7 billion, one nation, imagination...it's a beautiful day
 
User avatar
7BOEING7
Posts: 3039
Joined: Fri Oct 12, 2012 5:28 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 4:08 am

Quoting planemaker (Reply 89):
Yes, it is called 4D flight trajectories and sense & avoid.

   To all you said.

Quoting Mir (Reply 90):
Those will always be the duties. Those things must get done in any flight, whether directly by the crew or via automation (for which the crew is still responsible).

The issue that will have to be overcome in order to have a safe single-pilot cockpit is not the automation of the flying, but rather the automation of all the other tasks - the running of checklists, the coordination with the FAs, dispatchers and ATC, and the decision-making process. And those are a lot more difficult than just automating the flying of the airplane.

Crew of one.

Actually the automation be it "fail operational" on the airplane or through a "ground pilot" will probably be the most difficult -- engineering for total incapacitation of the pilot on board. The checklists on the 777 (and 787) have been partially automated for years -- the systems have become more autonomous such as the electronic load shed and recovery. Since it will begin with cargo, no FA's to worry about and dispatch will work just as well with one pilot as it does with two. ATC is already interfacing with the airplane through the navigation system, all the pilot has to do is push the button to accept the request. It's coming sooner than you think.
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 4:45 am

Quoting 7BOEING7 (Reply 91):
The checklists on the 777 (and 787) have been partially automated for years

That trend (for example, EMB ended up eliminating 8 or 9 out every 10 checklist items on the Phenom) was partially the rational for EMB to "come out" of the commercial SP "closet" and say that they are looking to provide single-pilot commercial aircraft capability in the 2020-25 timeframe.


Quoting 7BOEING7 (Reply 91):
ATC is already interfacing with the airplane through the navigation system, all the pilot has to do is push the button to accept the request.

You obviously already know quite a bit about InFlight Optimization Services  
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
2175301
Posts: 1576
Joined: Wed May 16, 2007 11:19 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 4:46 am

Quoting planemaker (Reply 89):


Quoting Pihero (Reply 85):
... and yet, we have the Fukushima accident, in which engineers couldn't forecast a series of uncontained failures which released quite an amount of radioactive mlaterial into the atmosphere and the sea.


Quoting Planemaker:
Yes, an accident that occurred in a reactor that was designed over 50 YEARS AGO ago and built 45 YEARS AGO - and the very first built by TEPCO... when SP ops are something that will start to happen 7-10 years in the FUTURE.

The comparison is worse than that - the Fukushima Nuclear Power Plant was built without regulations that required the siting and design of the plant to handle the worst natural disasters known in the area. It was known that a Tsunami hit the area on average ever 90-100 years that would be higher than the plant - but they only had to design for what was most likely in the next 50 years or so and the historically routine Tsunami did not make the list. In the USA nuclear plants have always had to design for the worst known natural disaster identified in the geological record for the local area (actually 1.5 times the worst identified event). Some of the European countries as well (but not all - and I cannot speak to the other countries).

Have a great day,
 
spacecadet
Posts: 3484
Joined: Thu Sep 20, 2001 3:36 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 5:01 am

Quoting planemaker (Reply 84):
How did Watson do this? By reading over 2 million pages from medical journals, +600,000 pieces of medical evidence, all national guidelines, hospital best practises, clinical trials and text books, and searching up to 1.5 million patient records, for information.

These aren't even the real accomplishments of Watson (which I almost brought up in my earlier reply, but I deleted about 70% of what I'd written for brevity's sake).

The real accomplishments of Watson are a) that the machine has learned natural language, and b) that the machine has learned how to use the data it's ingested to solve the problems it's presented with. It's one thing for a computer to have a vast database of information; it's another thing for it to actually know what that information *means*. That's really a watershed moment, and something we'd previously thought of as a uniquely human trait. This is a new way of thinking about AI that I think some in this thread may just not be aware of. It's no longer about brute force and programming every single potential fault and response. This AI learns in a way similar to how we do, but it does so much faster.

Watson's programming is, according to the engineers who programmed it, relatively simple. It was programmed to learn and then given vast processing power, huge storage capacity and the entire internet plus various reference texts as a knowledge database. The programmers themselves have been surprised by what it's been able to do; there was a NOVA episode on it a while back where the shock on their faces is obvious when it demonstrates an understanding of linguistic concepts that it was never programmed to understand and had never understood before.

This may not be getting close to the technological singularity yet, but it's certainly a step closer.

I see no reason why Watson or a Watson-like system couldn't be flying planes within the next few decades.
I'm tired of being a wanna-be league bowler. I wanna be a league bowler!
 
planemaker
Posts: 5411
Joined: Mon Aug 25, 2003 12:53 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 5:25 am

Quoting 2175301 (Reply 93):
The comparison is worse than that

Thanks for that additional info.

Quoting spacecadet (Reply 94):
These aren't even the real accomplishments of Watson (which I almost brought up in my earlier reply, but I deleted about 70% of what I'd written for brevity's sake).

I agree with you! I was hoping that when I said that Watson "read" that people would understand that it was "read" and not "inputted" (someone actually entering the data into the Watson database). I didn't want to get into your level of detail as it unfortunately seems that most tech detractors are simply not interested in knowing what is actually already out there... let alone in the pipeline. For example, it will be very interesting to see what they come up with at the NASA D-Wave quantum computer that Google just bought for them (I believe that Google will have 40% use of it).
Nationalism is an infantile disease. It is the measles of mankind. - A. Einstein
 
aeroflop
Posts: 53
Joined: Thu Mar 29, 2012 7:12 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 5:37 am

Ctrl+F "Planemaker"....... Yup. Of course.
 
spink
Posts: 317
Joined: Tue Aug 09, 2005 5:58 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 6:31 am

Quoting 2175301 (Reply 93):
The comparison is worse than that - the Fukushima Nuclear Power Plant was built without regulations that required the siting and design of the plant to handle the worst natural disasters known in the area. It was known that a Tsunami hit the area on average ever 90-100 years that would be higher than the plant - but they only had to design for what was most likely in the next 50 years or so and the historically routine Tsunami did not make the list. In the USA nuclear plants have always had to design for the worst known natural disaster identified in the geological record for the local area (actually 1.5 times the worst identified event). Some of the European countries as well (but not all - and I cannot speak to the other countries).

And the requirements are even beyond that. For the new reactors that TVA is planning for, they have to model and plan for a 100 year flood with cascading dam failures all along the Tennessee river. AKA, they are planning and designing for water levels well beyond what has ever been recorded and well beyond what is probable.
 
User avatar
airmagnac
Posts: 451
Joined: Wed Apr 18, 2012 10:24 pm

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 10:10 am

Quoting hivue (Reply 82):
think this is a false dichotomy. My suspicion is that military UAV missions are flown entirely by computer with systems engineer "pilots" monitoring. If something starts to go a little bit wrong, it's probably rectified with a keyboard rather than a flight sim joystick and rudder pedals. If we ever see "full automatic commercial ops" there will still be these human systems engineer "pilots" monitoring.
Quoting spink (Reply 87):
Both the X-47 and X-45 operate primarily autonomously. They have fallback remote pilot capabilities but that is only used as a backup. Normal flight ops are done via navigational instructions to the airframe which actually pilots the plane including TO/landing and formation flying.

So we are discussing different things. You seem to define a "pilot" as a guy moving a stick to turn to the left and right or climb and descend, ie making short term decisions regarding the flight path. As you wish, as long as it is clearly stated !
But those functions are already in great part automated, and are not the issue. As I have hinted, and as Pihero explained in more detail, the duties of a pilot on board an aircraft also consist in much longer-term decision-making, involving many factors which may be difficult to link to one another. The fact that they are "long-term" makes them abstract, hence difficult to explain and understand (especially on a text-based forum), yet those functions are there and have to be carried out.

Quoting AngMoh (Reply 86):
And the key thing is that are the duties today.

The functions involved in piloting an airplane have been the same since the Wrights (or Weisskopf ??) flew 110 years ago, and will remain the same. What changes is the concrete technological solutions that are used to perform these functions. These consist in increasingly complex automatic systems, which have slowly taken over the short-term decision loops : flight control (control surface deflections), flight guidance (attitude) and flight management (trajectory).
So far we have achieved full automation for only one of these functions, the flight control loop (through the use of FBW). There is still a long way to go before automatics permanently take over the rest of these functions. Let alone take over the still higher level functions described by Pihero.



Quoting hivue (Reply 82):

In the case of AF447 I don't think there was any significant issue with the automated systems detecting problems and acting appropriately. The issue was with the crew being unable (for whatever reason) to follow correct SOPs
Quoting hivue (Reply 88):
an air data device invented before the Wright Brothers flew failed (due to external causes) and the humans were not able to cope with the resulting (confusing) situation

We are discussing automatics taking over human decision-making activites, we can't just shrug off a case of automatics giving up as irrelevant
The crew f***d up, big time. No denial. But they were given the opportunity to f*** up in the first place because the automatics gave up. All the sophisticated gizmos are able to understand that the airspeed data was weird, but unable to detect why it is weird, and shut themselves down as a result, leaving the crews to deal with the confusion.

As you say, pitot tubes have been used for hundreds of years, and yet we are still incapable of reliably detecting and reporting when they fail. And some here are saying that it's just a few years till we have solved all the problems...

Detecting and reporting failures is important in automatics, because as automatics do not have inductive reasoning capabilities, hence no adaptability, it is essential that any data they use is validated. This is an immensly difficult problem, which only becomes more difficult as you head to the abstract long-term levels of decision-making.

Quoting planemaker (Reply 89):
Yes, an accident that occurred in a reactor that was designed over 50 YEARS AGO ago and built 45 YEARS AGO - and the very first built by TEPCO... when SP ops are something that will start to happen 7-10 years in the FUTURE.

So what ? 50 years ago, engineers screwed up, but 7-10 years from now engineers will be perfect, thanks to technological progress ?? That makes no sense.
50 years ago, overconfident people took shortcuts to build an nuclear powerplant before necessary level of knowledge, understanding, technologies and regulations were in place. If overconfident people take short cuts to implement SP ops before the necessary knowledge and technological solutions are available, then we'll obtain the same results.
Same inputs, same results, be it 1960, 2013, or 2050.

Quoting spacecadet (Reply 94):
the real accomplishments of Watson
Quoting spacecadet (Reply 94):
a) that the machine has learned natural language, and b) that the machine has learned how to use the data it's ingested to solve the problems it's presented with

Indeed this is a great achievment. As I mentioned before the ability to link together several pieces of data stored in several different formats/languages, regarding seemingly different and unrelated types of information, is a crucial thing in decision-making. And I do agree that such technologies may assist the pilots and reduce their workload to the point that 1 human can cope in normal operations. But that does not take away the issues of human factors and human/machine interfacing that must be solved for SP ops.
And Watson remains a deductive machine, not an inductive one. It can deduce conclusions by linking together an impossible amount of data, but only if it evolves in an environment it understands. Give it a situation it does not recognize, and it will be lost. But usually unprecedented situations correspond to crisis times. So the computer would pack in when you need it the most. Until machines achieve inductive reasoning capabilty, we'll need humans somewhere.

Quoting planemaker (Reply 95):
tech detractors are simply not interested in knowing what is actually already out there... let alone in the pipeline

I see no tech detractors here. What I see are a few people trying to understand what piloting is, what are the benefits/drawbacks of automating those activities, what are the problems to be solved to automate those tasks, and whether or not the current tech or near-future tech is good enough to solve these problems.
And on the other hand I see some talk about how great technology is, how it will solve all problems (but which problems ??), how it is so much better than humans (according to which criteria ??), and them some mentions of various techy solutions like Watson, 4D trajectories, or a Boeing data-crunching software, but no real explanation of how they actually address the problems (again, which problems ??).
My goal as an engineer is to fill my soul with coffee and become immortal
 
Mir
Posts: 19491
Joined: Mon Jan 05, 2004 3:55 am

RE: Planes Can Fly Themselves. BS!

Sat Jun 01, 2013 12:28 pm

Quoting 7BOEING7 (Reply 91):
dispatch will work just as well with one pilot as it does with two.

Not if the pilot has to keep an eye on how the airplane is flying as well.

-Mir
7 billion, one nation, imagination...it's a beautiful day

Popular Searches On Airliners.net

Top Photos of Last:   24 Hours  •  48 Hours  •  7 Days  •  30 Days  •  180 Days  •  365 Days  •  All Time

Military Aircraft Every type from fighters to helicopters from air forces around the globe

Classic Airliners Props and jets from the good old days

Flight Decks Views from inside the cockpit

Aircraft Cabins Passenger cabin shots showing seat arrangements as well as cargo aircraft interior

Cargo Aircraft Pictures of great freighter aircraft

Government Aircraft Aircraft flying government officials

Helicopters Our large helicopter section. Both military and civil versions

Blimps / Airships Everything from the Goodyear blimp to the Zeppelin

Night Photos Beautiful shots taken while the sun is below the horizon

Accidents Accident, incident and crash related photos

Air to Air Photos taken by airborne photographers of airborne aircraft

Special Paint Schemes Aircraft painted in beautiful and original liveries

Airport Overviews Airport overviews from the air or ground

Tails and Winglets Tail and Winglet closeups with beautiful airline logos