A surfeit of "Research" papers telling us what we already know
Thread Starter
A surfeit of "Research" papers telling us what we already know
Yet another research paper telling us that automation is the cause of degradation of manual flying skills. Ah - but this is a new paper! The writer reports; "To overcome the erosion of cognitive skills, Casner’s team recommends additional practice, during actual flights or on simulators. Another possibility is teaching active monitoring of cockpit automation during long flights. Indeed, some airlines now encourage their pilots to turn off the autopilot and other automation systems occasionally. No ****, Sherlock...
But do operators learn from history and change their policies after all this research indicates a problem with automation addiction? No way. With some airlines in China banning first officers from doing take off and landings period, then no wonder automation addiction is rapidly approach the same level as smoking and alcohol addiction.
How Cockpit Computers Prejudice Pilots? Performance | Inside Science
But do operators learn from history and change their policies after all this research indicates a problem with automation addiction? No way. With some airlines in China banning first officers from doing take off and landings period, then no wonder automation addiction is rapidly approach the same level as smoking and alcohol addiction.
How Cockpit Computers Prejudice Pilots? Performance | Inside Science
Per Ardua ad Astraeus
Join Date: Mar 2000
Location: UK
Posts: 18,579
Likes: 0
Received 0 Likes
on
0 Posts
That last action required flight crews to keep in their heads the picture of where they were, where they were going, and where they should be going.
Chrimble going well down there?
Join Date: Apr 2013
Location: everywhere
Posts: 128
Likes: 0
Received 0 Likes
on
0 Posts
Centaurus,
You are correct that there are a lot of papers telling us what we already know. On the other hand perhaps, it's good that someone takes their time to quantify what has been a conventional wisdom for long. Also, if losing skills due to automation dependency had been so obvious for all, why we have manufacturers/operators/instructors pushing for using it all time? Perhaps this paper is for their benefit?
Coming back to the article itself,
What surprised me was the conclusion that basic "stick and rudder" skills, instrument scan etc. do not go away with excessive use of automation. I'm not sure if that's true - mine have definitely gone south since I started flying Airbus planes
As for the second conclusion regarding reduced situational awareness due to automation - well, that's what I've been pondering a lot. We don't plan our flights ourselves and I sometimes even have difficulty determining over which country I'm flying, let alone city. You just know if you're on the "magenta line", or not... You don't know what track you should be flying now, or after next waypoint. The ETA's on the flight log don't mean much etc. All in all, automation prevents building the "big picture"...
I'd be interested in reading the original paper - can anyone post a link?
You are correct that there are a lot of papers telling us what we already know. On the other hand perhaps, it's good that someone takes their time to quantify what has been a conventional wisdom for long. Also, if losing skills due to automation dependency had been so obvious for all, why we have manufacturers/operators/instructors pushing for using it all time? Perhaps this paper is for their benefit?
Coming back to the article itself,
What surprised me was the conclusion that basic "stick and rudder" skills, instrument scan etc. do not go away with excessive use of automation. I'm not sure if that's true - mine have definitely gone south since I started flying Airbus planes
As for the second conclusion regarding reduced situational awareness due to automation - well, that's what I've been pondering a lot. We don't plan our flights ourselves and I sometimes even have difficulty determining over which country I'm flying, let alone city. You just know if you're on the "magenta line", or not... You don't know what track you should be flying now, or after next waypoint. The ETA's on the flight log don't mean much etc. All in all, automation prevents building the "big picture"...
I'd be interested in reading the original paper - can anyone post a link?
Join Date: Sep 2011
Location: Vienna
Age: 50
Posts: 359
Likes: 0
Received 0 Likes
on
0 Posts
Seems to be freely accessible in fulltext: The Retention of Manual Flying Skills in the Automated Cockpit (if it does not work, send me a PM, I might be able to get the PDF via the online library at our university).
From the perspective of a humble low-hour-PPL but almost 15-year researcher in the social sciences, the study is niftily designed but has some methodological issues nonetheless. First, the sample size is very low, but this is more of a statement than a critique, as a big sample would be almost impossible/unaffordable to get in this setting. On the other hand, the fact that they report statistically significant findings with this small sample suggests that for the 16 pilots (or 48 sim flights, could not find out by a quick scan which number the calculations are finally based on) suggests that the effects were considerable (read: not so small as to make no practical difference). Second, it is not quite clear to me to which extent the "currency" of the participating pilots in manual flying was accounted for in the analysis, and this could affect the results.
As for the conclusion that basic manual skills do not erode as fast and/or dramatically as the cognitive skills of "staying in the loop", this does not really surprise me. Frankly, having followed the debate on skill erosion owing to automation (just out of interest), I have always been skeptical about the sole emphasis on manual skills (as opposed to situation awareness and mentally "being on top of the game") for improving safety. (But then again, my frontline experience is arguably nil compared to almost all other folks here.)
From the perspective of a humble low-hour-PPL but almost 15-year researcher in the social sciences, the study is niftily designed but has some methodological issues nonetheless. First, the sample size is very low, but this is more of a statement than a critique, as a big sample would be almost impossible/unaffordable to get in this setting. On the other hand, the fact that they report statistically significant findings with this small sample suggests that for the 16 pilots (or 48 sim flights, could not find out by a quick scan which number the calculations are finally based on) suggests that the effects were considerable (read: not so small as to make no practical difference). Second, it is not quite clear to me to which extent the "currency" of the participating pilots in manual flying was accounted for in the analysis, and this could affect the results.
As for the conclusion that basic manual skills do not erode as fast and/or dramatically as the cognitive skills of "staying in the loop", this does not really surprise me. Frankly, having followed the debate on skill erosion owing to automation (just out of interest), I have always been skeptical about the sole emphasis on manual skills (as opposed to situation awareness and mentally "being on top of the game") for improving safety. (But then again, my frontline experience is arguably nil compared to almost all other folks here.)
Join Date: Jun 2000
Location: last time I looked I was still here.
Posts: 4,507
Likes: 0
Received 0 Likes
on
0 Posts
All in all, automation prevents building the "big picture"...
I disagree with this simple statement. I've found that automation, and I assume we are talking about modern LNAV (map) VNAV a/c has broadened my SA enormously because there is so much more clear information displayed; and the automatics and auto-throttle allow my mind time to scan and digest it all in a relaxed manner. However, I qualify this with the fact that I was brought up on needles & dials and the most basic of autopilots where CWS in climb & descent was the norm. i.e. you had to control attitude and there was no auto-throttle. Navigation was via needles of VOR & NDB, with a few DME's. Many airfields did not have an ILS or DME on the field, and if they did have an ILS, it often had no DME. SA was the name of the game and there was much visual flying. You had to have a mental picture of where you were at all times and manage not only the navigation, but also the energy of the beast. The modern displays are a doddle, but I still maintain the basic scan of needles & DME's etc. to see that Windows 95 is not leading me up or down the garden path.
As for the conclusion that basic manual skills do not erode as fast and/or dramatically as the cognitive skills of "staying in the loop",
I'm in disagreement with this one as well. Reason? In todays fast track cadet into jet RHS with <200hours they never have learnt the basic manual flying skills of the beast they are trying to control in the first place . IR skills in a Seneca in no way transfer into a jet, neither do the basic visual circuit techniques. It requires much more finesse and alertness and to be way ahead of the a/c. Thus, if you never had them it's difficult to say that they are slow to erode. I'm curious where this conclusion came from. If you are discussing experienced crews, then perhaps, but the old farts are slipping out; the modern airlines dissuade manual approaches, so in a generation's time there will be no old farts only modern ones whose wings have gone the same way as the Dodo through lack of use.
I disagree with this simple statement. I've found that automation, and I assume we are talking about modern LNAV (map) VNAV a/c has broadened my SA enormously because there is so much more clear information displayed; and the automatics and auto-throttle allow my mind time to scan and digest it all in a relaxed manner. However, I qualify this with the fact that I was brought up on needles & dials and the most basic of autopilots where CWS in climb & descent was the norm. i.e. you had to control attitude and there was no auto-throttle. Navigation was via needles of VOR & NDB, with a few DME's. Many airfields did not have an ILS or DME on the field, and if they did have an ILS, it often had no DME. SA was the name of the game and there was much visual flying. You had to have a mental picture of where you were at all times and manage not only the navigation, but also the energy of the beast. The modern displays are a doddle, but I still maintain the basic scan of needles & DME's etc. to see that Windows 95 is not leading me up or down the garden path.
As for the conclusion that basic manual skills do not erode as fast and/or dramatically as the cognitive skills of "staying in the loop",
I'm in disagreement with this one as well. Reason? In todays fast track cadet into jet RHS with <200hours they never have learnt the basic manual flying skills of the beast they are trying to control in the first place . IR skills in a Seneca in no way transfer into a jet, neither do the basic visual circuit techniques. It requires much more finesse and alertness and to be way ahead of the a/c. Thus, if you never had them it's difficult to say that they are slow to erode. I'm curious where this conclusion came from. If you are discussing experienced crews, then perhaps, but the old farts are slipping out; the modern airlines dissuade manual approaches, so in a generation's time there will be no old farts only modern ones whose wings have gone the same way as the Dodo through lack of use.
Join Date: Sep 2011
Location: Vienna
Age: 50
Posts: 359
Likes: 0
Received 0 Likes
on
0 Posts
As for the conclusion that basic manual skills do not erode as fast and/or dramatically as the cognitive skills of "staying in the loop",
This (IMHO) partly comes back to the apparent lack of explicitly controlling for "IFR manual flying currency" in the analyses. A longitudinal multi-cohort study which repeatedly examines the effects presented in the study both for old farts and fast track cadets would certainly be a lot more revealing, but I would not want to be the one implementing it in practice, even ignoring budget restrictions for a moment
.
Moderator
I've found that automation ... has broadened my SA enormously because there is so much more clear information displayed ... allow my mind time to scan and digest it all in a relaxed manner
I doubt that anyone here seriously is denying that observation .. when the gee-whizz stuff works well, it works really well.
Certainly, Centaurus (who has been a close colleague over a great many decades) isn't anti-automatics, per se.
Indeed, philosophically, the "ideal" goal might be total automation on the pilot/dog principle when considering the greater range of circumstances where the automatics function satisfactorily and reliably ?
I suggest that Centaurus's (and many others') concern is with what too high a reliance on the automatics might do to the (pilot's) ability - over a period of time - to rise to the occasion when they (the automatics) degrade significantly, or disappear totally, in high workload situations where the pilot (crew) is the last line of defence against disaster.
This assumes that the pilot's original training was sufficient to get to a highly competent standard in the first place. As an aside, lest some think me to be an anti-automatics heretic, I have trained many low time/ab initio airline pilots to a high standard of manipulative capability (albeit without much cognitive depth at such an early stage of their careers) so it can be done. The ongoing concern is the level of retained skill in the face of significantly reduced habituation by exposure and routine practice due to SOPs which might over-emphasise automatic flight.
I was fortunate that my early jet flying was during a period when one had to be proficient at both but could exercise personal judgement as to the amount of line flying on nil automatics/autopilot/autothrottle/FD/etc. The boss's requirement was that we should just make sure we got it right. For myself (domestic sectors) I routinely chose to hand fly all non-cruise parts of the flight and, occasionally, extended parts of the cruise. I don't ever recall any complaints from the cabin crew so I guess I wasn't too rough with my stick and rudder work.
I suggest further that there are two opposing camps of thought in this debate ..
(a) we accept a low (automatics hardware/software) risk of failure and have the occasional prang which the traditional stick and rudder pilot would have saved whilst drinking a coffee and chatting amiably to the other crew.
The benefit (for the majority of flights) is/should be an enhanced risk profile for the typical routine flight.
(b) we work to maintain those traditional stick and rudder/cognitive skills so that the isolated simple mishap might be avoided.
The downside is that the overall risk for routine flights will rise.
Philosophically, which is the preferable option ?
For consideration, on the one hand we have examples such as AF447 while on the other UA232.
A difficult call, I guess, which will engender debate for a considerable time to come.
I doubt that anyone here seriously is denying that observation .. when the gee-whizz stuff works well, it works really well.
Certainly, Centaurus (who has been a close colleague over a great many decades) isn't anti-automatics, per se.
Indeed, philosophically, the "ideal" goal might be total automation on the pilot/dog principle when considering the greater range of circumstances where the automatics function satisfactorily and reliably ?
I suggest that Centaurus's (and many others') concern is with what too high a reliance on the automatics might do to the (pilot's) ability - over a period of time - to rise to the occasion when they (the automatics) degrade significantly, or disappear totally, in high workload situations where the pilot (crew) is the last line of defence against disaster.
This assumes that the pilot's original training was sufficient to get to a highly competent standard in the first place. As an aside, lest some think me to be an anti-automatics heretic, I have trained many low time/ab initio airline pilots to a high standard of manipulative capability (albeit without much cognitive depth at such an early stage of their careers) so it can be done. The ongoing concern is the level of retained skill in the face of significantly reduced habituation by exposure and routine practice due to SOPs which might over-emphasise automatic flight.
I was fortunate that my early jet flying was during a period when one had to be proficient at both but could exercise personal judgement as to the amount of line flying on nil automatics/autopilot/autothrottle/FD/etc. The boss's requirement was that we should just make sure we got it right. For myself (domestic sectors) I routinely chose to hand fly all non-cruise parts of the flight and, occasionally, extended parts of the cruise. I don't ever recall any complaints from the cabin crew so I guess I wasn't too rough with my stick and rudder work.
I suggest further that there are two opposing camps of thought in this debate ..
(a) we accept a low (automatics hardware/software) risk of failure and have the occasional prang which the traditional stick and rudder pilot would have saved whilst drinking a coffee and chatting amiably to the other crew.
The benefit (for the majority of flights) is/should be an enhanced risk profile for the typical routine flight.
(b) we work to maintain those traditional stick and rudder/cognitive skills so that the isolated simple mishap might be avoided.
The downside is that the overall risk for routine flights will rise.
Philosophically, which is the preferable option ?
For consideration, on the one hand we have examples such as AF447 while on the other UA232.
A difficult call, I guess, which will engender debate for a considerable time to come.
Join Date: Jun 2000
Location: last time I looked I was still here.
Posts: 4,507
Likes: 0
Received 0 Likes
on
0 Posts
I teach according to a syllabus designed by others. It covers all the requirements in quick time. It is heavily orientated towards line operation SOP's. There is a token gesture towards manual flying, but only that required by the LST items; e.g. turns (even steep), stalls, raw data ILS and the SE ILS's & G/A's. This is not teaching the student to manoeuvre the a/c manually in a variety of real situations. Further, there is circuit training. Then they go on the line where manual approaches are discouraged and the type of circuit they flew in base training is not allowed. What chance have they got to develop manual skills. In the TQ syllabus I see far too often that the first reaction of a student is to engage the A/P and then, when misunderstanding why the a/c is not climbing, descending or whatever it is they want it to do, selecting VNAV in the hope that Captain FMC will sort it out.
So the problem is three-fold: lack of manual skills, lack of scanning basic performance parameters to ensure the a/c is doing what YOU want, and lack of a full understanding of what the AFDS system is capable of, and what it is not.
All in all, therefore, there is a lack of deep knowledge and skill in manipulating the a/c both manually and via the automatics. The SOP's are orientated heavily towards use of automatics and every scenario, (nearly) has been written with instructions. On an ideal day in and ideal environment you can get away with this. Then ATC turn you in short and you are high; you get this and a tailwind; you have to deviate from the ideal route/path for various reasons; ATC ask you to speed up/slow down; a whole variety of items that cause you to deviate from the ideal SOP profile. I see this big question mark rising out of the student's (and even line F/O's) head. Eyes spinning and the piano playing starts.
The easiest solution, even to the automatic orientated operators, is to increase the training & understanding of those automatics. That can be done in a sim. The deeper manual skills need to be on the line and that needs a cultural shift of Flt Ops management. In my B732 days manual approaches were the norm, necessary and encouraged. As an apprentice F/O you watched and learnt and practiced. Not anymore. There is no demonstration to watch from LHS. And with captains sprouting wings at 3000hrs, who've never had the manual skills this downward spiral is firmly spinning in motion.
I hear one large airline, after B777 SAN, is starting the TQ course with 2 sessions of basic flying with no FD or automatics; then add the AFDS and A/P and then add the NNC's. Building blocks. An enlightened Flt OPs dept and training dept. This will never happen across the board without XAA intervention. Perhaps they should be held more accountable when pilot error handling incidents occur.
Is it an adequate defence to say that all legal training and checking requirements were carried out. Bare minimums are often not enough. As has been said, the crew are the last insurance policy and line of defence and the pax expect better.
Interestingly, and perhaps adding weight to the argument, it that John T has noticed the trend in teaching in USA s I have teaching in EU. This is an industry-wide issue not a national one.
So the problem is three-fold: lack of manual skills, lack of scanning basic performance parameters to ensure the a/c is doing what YOU want, and lack of a full understanding of what the AFDS system is capable of, and what it is not.
All in all, therefore, there is a lack of deep knowledge and skill in manipulating the a/c both manually and via the automatics. The SOP's are orientated heavily towards use of automatics and every scenario, (nearly) has been written with instructions. On an ideal day in and ideal environment you can get away with this. Then ATC turn you in short and you are high; you get this and a tailwind; you have to deviate from the ideal route/path for various reasons; ATC ask you to speed up/slow down; a whole variety of items that cause you to deviate from the ideal SOP profile. I see this big question mark rising out of the student's (and even line F/O's) head. Eyes spinning and the piano playing starts.
The easiest solution, even to the automatic orientated operators, is to increase the training & understanding of those automatics. That can be done in a sim. The deeper manual skills need to be on the line and that needs a cultural shift of Flt Ops management. In my B732 days manual approaches were the norm, necessary and encouraged. As an apprentice F/O you watched and learnt and practiced. Not anymore. There is no demonstration to watch from LHS. And with captains sprouting wings at 3000hrs, who've never had the manual skills this downward spiral is firmly spinning in motion.
I hear one large airline, after B777 SAN, is starting the TQ course with 2 sessions of basic flying with no FD or automatics; then add the AFDS and A/P and then add the NNC's. Building blocks. An enlightened Flt OPs dept and training dept. This will never happen across the board without XAA intervention. Perhaps they should be held more accountable when pilot error handling incidents occur.
Is it an adequate defence to say that all legal training and checking requirements were carried out. Bare minimums are often not enough. As has been said, the crew are the last insurance policy and line of defence and the pax expect better.
Interestingly, and perhaps adding weight to the argument, it that John T has noticed the trend in teaching in USA s I have teaching in EU. This is an industry-wide issue not a national one.
Last edited by RAT 5; 26th Dec 2014 at 13:23.
Moderator
I have to concur with RAT 5 albeit that I have been out of the up front flying game for some time now.
Specifically,
I see far too often that the first reaction of a student is to engage the A/P and then, when misunderstanding ... selecting VNAV in the hope that Captain FMC will sort it out.
And, like all computers .. it can only do what it is told/programmed to do. If the pilot has little clue as to the ins and outs .. sooner or later a foregone conclusion will occur. So far as Captain FMC's capability is concerned, in the early days, a common flightdeck cry was along the lines of "what's it doing now .. ? ". I guess that still happens even with more sophisticated and well-conditioned systems ?
Eyes spinning and the piano playing starts.
Indeed, whence the critical importance of finding time in the endorsement sim training to develop something in the way of out of the box thinking-on-the-feet skills .. for many, this is the last time the opportunity might present itself.
For instance, my preferred trick was to work the sim student up to a point where he/she (SP, raw data, hand flown, minimum helpful instruments) could do a min vis takeoff with a failure, and then find his/her way back to pick up the ILS for a 0/0 landing .. sweaty stuff but most enjoyed it immensely in a non-criticising environment. I fondly recall one command upgrade chap who started off grossly overconfident and moderately undercompetent. We both worked hard to rectify the deficiencies and, at the end of a gargantuan late night session (courtesy of a kind-hearted sim tech on duty who let us play on into the following empty session) he ran the above exercise to a very high standard and literally floated out of the sim on a high note .. quite knackered but supremely satisfied with himself.
Centaurus, for instance, much favours SP min air time circuit and landing for a similar confidence builder exercise.
Invariably, a number of ATC-style bunching-up-getting-higher-and-higher problems are thrown in during the endorsement .. the student ends up, necessarily, having to force him/herself to push the cognitive figuring well out in front of the aircraft .. whether on automatics or stick and rudder.
The deeper manual skills need to be on the line
.. but, if that doesn't present itself for whatever reason, the sim remains a very useful tool to achieve a good starting standard. However, this only works in a high task workload but nurturing sim endorsement environment (it may be a bit hackneyed .. but it seemed to work fine as I saw things). Otherwise, there just isn't time to fit everything in ... sim tech cajoling for extra time notwithstanding.
Bare minimums are often not enough
.. ever.
I was fortunate to be brought into airline flying within the XXX Airline environment where management philosophy was directed to gross overtraining of crews .. I guess that isn't seen these days ?. Admittedly, that was in a strictly regulated domestic environment where the cost pressures were far more benevolent than we have seen over recent decades.
The basic tenet was that one had to achieve a sufficiently above minimum standard at recurrent check signoff to allow for the weaker folk who might allow their standards to slip over the following six months .. The aim was to ensure, so far as reasonably practicable, that everyone was above the minimum standard for the intervening period. Idealistic, perhaps, but the line standards were high.
John T has noticed the trend in teaching in USA
Not USA, albeit several countries, and quite dated now.
Specifically,
I see far too often that the first reaction of a student is to engage the A/P and then, when misunderstanding ... selecting VNAV in the hope that Captain FMC will sort it out.
And, like all computers .. it can only do what it is told/programmed to do. If the pilot has little clue as to the ins and outs .. sooner or later a foregone conclusion will occur. So far as Captain FMC's capability is concerned, in the early days, a common flightdeck cry was along the lines of "what's it doing now .. ? ". I guess that still happens even with more sophisticated and well-conditioned systems ?
Eyes spinning and the piano playing starts.
Indeed, whence the critical importance of finding time in the endorsement sim training to develop something in the way of out of the box thinking-on-the-feet skills .. for many, this is the last time the opportunity might present itself.
For instance, my preferred trick was to work the sim student up to a point where he/she (SP, raw data, hand flown, minimum helpful instruments) could do a min vis takeoff with a failure, and then find his/her way back to pick up the ILS for a 0/0 landing .. sweaty stuff but most enjoyed it immensely in a non-criticising environment. I fondly recall one command upgrade chap who started off grossly overconfident and moderately undercompetent. We both worked hard to rectify the deficiencies and, at the end of a gargantuan late night session (courtesy of a kind-hearted sim tech on duty who let us play on into the following empty session) he ran the above exercise to a very high standard and literally floated out of the sim on a high note .. quite knackered but supremely satisfied with himself.
Centaurus, for instance, much favours SP min air time circuit and landing for a similar confidence builder exercise.
Invariably, a number of ATC-style bunching-up-getting-higher-and-higher problems are thrown in during the endorsement .. the student ends up, necessarily, having to force him/herself to push the cognitive figuring well out in front of the aircraft .. whether on automatics or stick and rudder.
The deeper manual skills need to be on the line
.. but, if that doesn't present itself for whatever reason, the sim remains a very useful tool to achieve a good starting standard. However, this only works in a high task workload but nurturing sim endorsement environment (it may be a bit hackneyed .. but it seemed to work fine as I saw things). Otherwise, there just isn't time to fit everything in ... sim tech cajoling for extra time notwithstanding.
Bare minimums are often not enough
.. ever.
I was fortunate to be brought into airline flying within the XXX Airline environment where management philosophy was directed to gross overtraining of crews .. I guess that isn't seen these days ?. Admittedly, that was in a strictly regulated domestic environment where the cost pressures were far more benevolent than we have seen over recent decades.
The basic tenet was that one had to achieve a sufficiently above minimum standard at recurrent check signoff to allow for the weaker folk who might allow their standards to slip over the following six months .. The aim was to ensure, so far as reasonably practicable, that everyone was above the minimum standard for the intervening period. Idealistic, perhaps, but the line standards were high.
John T has noticed the trend in teaching in USA
Not USA, albeit several countries, and quite dated now.
