The Case for Remaining in the Loop

[downscc 1,327]

 

 

Responding to the unexpected is a significant part of pilot skill within real aviation but it's probably a skill set we're lacking in within the sim world, because we can 'control' our entire environment and it's outcomes.

 

Understatement.  I was an hour from home on a two hour flight from the paint shop when all electric/electronics in the twin went dead.  There are two alternators so what's this about... hmm, can't wait for the power to come back on as if in front of the computer screen.  Tried turning off one alternator and click! Heard Houston Center is asking others to try to contact me and I was able to reconnect.  Turns out the battery was so weak that it could only handle one alternator's field coils not both.  I felt pretty good about that experience because I was sharing the cockpit with an AirTran B717 pilot and I figured it out first.

 

Bad weather, engine failures, instrument failures... you name it.  A general aviation pilot expects the unexpected.

[Artur Munteanu 1]

I manually fly on departure to at least 10,000ft and usually on approach the entire STAR and landing.

This is wrong. In the real world aviaton is strictly unrecomanded By FAA, NTSB, etc. doing this, aircraft incidents happend because the pilots didn't activate the autopilot after take-off at 400 Ft 1000 FT or 1500Ft. Just remember the case of Ethiopian Airlines B738 near Beirut on Jan 25th 2010, wich lost height after takeoff and impacted Mediterranean! A major cause was human errors made in cascade by the crew, combinated with the rain weather condition at night (desorientation, stress and chronic fatigue after 188 consecutive hours of flight in 51 days, little experience on that airport), etc. One of the conclution of Lebanon's Civil Aviation Authority was that "Therefore the captain's decision to fly manually was a major contributor towards the degradation of the situation. Technically, the autopilot could have been engaged after 400 feet, according to FCOM, but however, there was no recording of the autopilot being engaged."

Nowdays the aircraft are higly automated so there is no need to pilot them by hand (there studing even to eliminate the pilots from cockpit). The main reason is (seeing what happend recently on Germanwings) the safety, then of course another important factor is the human error. Basically even today is posible to fly an A380 from take-off to landing without a pilot. Off course, I personally don't want to be on a plane with no one in the cockpit from gate to arrival. The theory is that is more safe today to fly with the computers on board, because they are not subject at the human errors.

So what you are doing, in the real world never heppends (of course in normal fly conditions - all instrumentation on board working and no hurricane on landing or take-off).

[77west 139]

 

 

This is wrong. In the real world aviaton is strictly unrecomanded By FAA, NTSB, etc. doing this, aircraft incidents happend because the pilots didn't activate the autopilot after take-off at 400 Ft 1000 FT or 1500Ft.

 

I am afraid you are wrong my friend. If the "pilots" are unable to operate the airplane without using autopilot from 400ft to landing, they should not be in the cockpit. Asiana 214?

[kevinh 409]

I agree 95%, Kevin. The remaining difference between the manufacturers' designs is that in a Boeing it is fairly easy for the pilots to override malfunctioning automation; in an Airbus - in rather rare circumstances as you pointed out - it might be more tricky to achieve that.

 

Still this is a topic about automation and human factors associated with it, and I did not mean to turn it into a Boeing vs. Airbus discussion.

 

The big point still is that, no matter what the manufacturer or the A/C type is, automation provides the bitter with the sweet. Which makes this a very fascinating topic. Thanks, Kyle, for starting it.

Not that tricky. Two switches just above the Captain's head. I agree with the rest of your post 100%.

[James Bennett 52]

Eh, people can use it however they want. I know i for one don't have much time at all to sim, maybe once a month, and i'm not going to use that time to plan my own route out! Far from feeling like i had to have as much realism in the planning and execution as possible, i now realise that it's more about just having fun with the sim regardless of how that's achieved.

[Don t moderate me this way 882]

What I am curious about is, how do RW pilots actually fly these days? Do they maintain proficiency by doing a lot of manual flying, or are they rather afraid of making mistakes, which then can be tracked back via flight data recordings, noise measurement devices and the like?

 

I would assume a large no. of pilots decide to "be safe rather than proficient" and leave some tricky portions of their flights (like SIDs and STARs) up to the nifty autoflight systems.

[LAdamson 240]

What I am curious about is, how do RW pilots actually fly these days? Do they maintain proficiency by doing a lot of manual flying, or are they rather afraid of making mistakes, which then can be tracked back via flight data recordings, noise measurement devices and the like?

 

I would assume a large no. of pilots decide to "be safe rather than proficient" and leave some tricky portions of their flights (like SIDs and STARs) up to the nifty autoflight systems.

I'll comment on the GA side of things. My RW plane had a 2 axis auto-pilot, in which pitch could be operated without roll.  When flying 20-30 miles through Class B airport airspace in our local area..............I was more than happy to use the altitude function of the A/P. It could keep altitude within -/+10'. That's much better than a proficient pilot could do (without constant concentration), and it allowed me to scan for other aircraft that ATC was pointing out, as well flying the directions ATC specified. 

[scandinavian13 5,192]

This is wrong. In the real world aviaton is strictly unrecomanded By FAA, NTSB, etc. doing this, aircraft incidents happend because the pilots didn't activate the autopilot after take-off at 400 Ft 1000 FT or 1500Ft.

I'll respond more when I get home. Saw this on my phone and had to nip this in the bud, because it's false. Very false.

 

http://www.faa.gov/other_visit/aviation_industry/airline_operators/airline_safety/safo/all_safos/media/2013/SAFO13002.pdf

 

EDIT:

 

Home now.

 

I posted the FAA guidance earlier, so I'll address the rest of the issues here.

 

Citing the ETH 738 flight really doesn't address the point. Operating at too low a level of automation certainly contributed to the issue, but there's a lot more there, as you brought up: fatigue, stress, and a high workload. Not using the autopilot certainly let the other issues all degrade the safety of the flight, but this does not mean, in any way, shape, or form, that people should start advocating AP use at all times, or even all times that it's possible. That assertion, to be blunt, is absolutely insane.

 

As for the rest, this is where the simmer group gets on my bad side, and even real world pilots too if they try to over simplify. Stating that pilots aren't needed is a dangerous proposition. Suggesting it gives the general public this idea that a pilot's job is very simple, and easily rid of. This is absolutely not true. If you believe that it is, and wish to continue perpetuating this idea, then please go fail a bunch of things on the 777 and have it attempt to autoland somewhere. What would happen if the aircraft were assigned vectors? How could we get it to fit into the system at a busy airport? If you think we should command it from the ground, how can you design a system that is reasonably resistant to hacking, while still being quick enough to react quickly when necessary (increasing levels of encryption require increasing time to decrypt)?

 

I get that, to you, this all seems very simple, but it is not simple...at all. Assertions like yours spread like wildfire, and the fallout from it can reach the people who make the regulations and decisions for how the future of aviation will look. These people are not always the FAA, NTSB, and so on. Unfortunately, these people could be elected officials who only think they know enough about the industry to regulate it. If all they've ever heard was "oh, flying's easy, planes can fly themselves now-a-days," then their decisions on policies that affect aviation could have very detrimental effects.

 

Moreover, and to really drive this point home, you must remember that computers are made by people. While they are not subject to some of the slip ups that a pilot might make in the 90% scenario, they are not error proof. Database errors can and do occur. Ever wonder why you occasionally see a revision number above 1 for your nav data? That means either the source to the provider, or the provider itself, made an error. As much as I like to think we're perfect, the very fact that we release service packs goes to show that we, as humans, even when creating code to do things automatically, can and do err.

 

As I mentioned in the very first post: automation is great, but it is not perfect.

[skelsey 765]

Interesting read and interesting topic, Kyle -- as you say, PFPX is great but the automatic route finder has its limitations. I normally run through a series of canned company routes, and then let it auto-generate a wind-optimised route and compare them on the results tab. Probably around 60-70% of the time one of the coroutes is quicker/more efficient (or the auto-route fails CFMU validation and once fixed to comply is then no longer optimal). Another thing to watch out for is that I believe it picks SIDs/STARs in alphabetical order (or at least it used to) which can in some cases result in incorrect results. However, on the whole it's much better than anything that we've had previously!

 

Incidentally, this has also given me the opportunity to mention one of the things which grinds my gears -- which is the "Children of the Magenta" presentation being wheeled out every time there's an incident or a discussion about automation, usually accompanied by complaints about deteriorating handling skills.

 

"Children of the Magenta" is not and was not about saying that automation is bad. It is about knowing how the automation works and using the appropriate mode for the given situation. Sometimes that might be all the bells and whistles, sometimes it might be none at all and sometimes it might be somewhere in between. The point is that one should both know how to make full use of the automatics and use airmanship to select the appropriate level of automation for the situation at hand.

 

Just because the weather is CAVOK doesn't necessarily mean that it's a sensible idea to turn everything off and hand-fly: there may be other factors (for instance, if the airport you're approaching is using the good weather to pack in more aircraft, it might be better to free up capacity within the crew to be looking outside for traffic during the initial approach rather than having PF 'poling' around the sky whilst PM twiddles knobs on the MCP whilst also making sure PF's hand-flying is accurate and dealing with all his/her other duties.

 

Having said that, clearly it is important to maintain hand-flying proficiency. But ultimately the automation is (should be!) there as a tool to be used to reduce workload and increase efficiency, which is how most enlightened operators should (and do) view it. Of course, some carriers are less enlightened than others!

[scandinavian13 5,192]

 

 

Incidentally, this has also given me the opportunity to mention one of the things which grinds my gears -- which is the "Children of the Magenta" presentation being wheeled out every time there's an incident or a discussion about automation, usually accompanied by complaints about deteriorating handling skills.

 

I'm hoping the rant about the mention of the video wasn't aimed at me. I think it's pretty clear that my post made light of various levels of automation and that, occasionally, punching them all off in favor of "switching to guns," if you will, is the ideal solution.

What I am curious about is, how do RW pilots actually fly these days? Do they maintain proficiency by doing a lot of manual flying, or are they rather afraid of making mistakes, which then can be tracked back via flight data recordings, noise measurement devices and the like?

 

I would assume a large no. of pilots decide to "be safe rather than proficient" and leave some tricky portions of their flights (like SIDs and STARs) up to the nifty autoflight systems.

 

In my entire aviation career, I've used an autopilot twice. In fact, one weekend, I spent 15 hours in a Cessna 207 (equipped with an AP). While 5 of those hours were flown by my uncle, the other 10 were mine and I flew all of those 10 by hand. The end of that flight was a low visibility, low ceiling LOC approach into BCB, which is no joke. I'm still alive to talk about it.

 

As for AP use in the airlines, it depends on the crew, and it depends on the airline's culture. Asian carriers are very automation dependent, as we've seen increasingly over the years. Over here, it's a mix of both. I've seen pilots who will fly from the ground to RVSM airspace on their own; I've seen those who turn it on at the lowest altitude possible; and I've seen everything in between.

 

This notion of the lack of autopilot use making things more dangerous needs to go away. The use of the autopilot is required in certain conditions (RVSM airspace being one of them), and can help to reduce the crew workload in important situations, but apart from that, it's perfectly fine to fly by hand.

[Alan_A 1,650]

Great discussion.  Just to add a dimension to it - NASA's Ames Research Center conducted a study late last year (part of an ongoing project) that suggested that the problem with cockpit automation wasn't so much the degradation of hand-flying skills, but rather the degradation of "mind skills." When automation was in use, pilots seemed to be less aware of the such things as the aircraft's position or systems performance.  And of more concern, they had a hard time "re-engaging" when they had to intervene suddenly, for example to troubleshoot a problem.  According to the researchers, "while pilots’ instrument scanning and aircraft control skills are reasonably well retained when automation is used, the retention of cognitive skills needed for manual flying may depend on the degree to which pilots remain actively engaged in supervising the automation."

 

Here's an abstract of the original paper.  There's a short summary here at the Human Factors and Ergonomics Society website.

 

So, some additional stuff to think about...

[WR269 367]

Sadly the days of pilots learning "stick and rudder" before going to automation are long gone. I do recall in the 80's, the "thing" to do was to go to a God-foresaken part of the world and get your hours. Airlines would only look at you if you had command time in places like PNG, Vietnam, etc. real flying.

 

I spent some time in the British Aerospace college in Australia and saw students going from Ab-Initio to ATPL in just over a year, and then making the jump from a Seneca to the right hand seat in an A320.....scary.

 

Also recall meeting a delegation of pilots from Cathay Pacific in the 90s....we had to have a good chuckle....they said "You blokes in YBBN and the folk at YPPH are the ONLY ones in the entire world who offer us Visual approaches....you should see the faces of our new First Officers when we accept!" and this had a hidden message...the good old fashioned skills were being eroded even back then.

 

Today, I have good friends as senior Captains and Instructors who still love to hand fly the aircraft to cruise or from Top of Descent down, and they need to describe what is happening to their first officers, for obvious reasons. It is like Capt De Crespigny stated, if you know how your aircraft handles, you know when things are right and when they are not right. He took manual control of a severely damaged A380 and knew straight away everything would be ok, the aircraft responded to his commands the way he knew it would.

 

Is this a sign of times?  Probably, Human Factors as a science will have ICT added as the understanding of the interaction of control evolves into a complex MODULE, not just a single entity, this being either the Pilot or the Autopilot/Automation.

[Artur Munteanu 1]

 

 

automation is great, but it is not perfect.

I agree with you, that's why I even said that personally I never board in a plain without a pilot in a cockpit (if this will ever heppend). Of course again, the computers error occurs but in the numbers of airline company today at the end et the month all that matter is the $ symbol. The fuel cost and maintanence cost today is more higher then in the past (15-20 yrs ago). All this at the detriment at the passenger. The company will try to make they're fleet more and more automatic and will try try to eliminate little by little the crew. Just remember the third man crew on B767? It was eliminate from the cockpit, because now the crew engineer was no more necesary on a 777. So, as we see the direction for the tomorow aviation, as we like or not, it will depend by computers. Is very simple: The computer don't have to sleep (so the plain could fly allways), the computer don't have to eat or drink (so two mels on board and two coffe and water bottle economize on every flight means less cost for company), and most important the computer doesn't need the checkpay at he end of the month.

IMO it's a sad situation.

[avgaskoolaid 292]

This is a very good, very relevant topic. It always irks the hell out of me when people say that pilotless airplanes are just around the corner, and that the only real hurdle to overcome is making people comfortable with the idea. I'd say that 90% of the "normal" people (who aren't interested in aviation) I talk to legitimately think that pilots hop in the plane, take it off, then push a button marked 'fly', then another one marked 'land'. I'm sincerely afraid that one day the bean counters will have their way and work the pilot out of the cockpit, a decision that hundreds of people will pay for in blood so that the airline can save a few bucks. 

 

On the plus side, though, if we don't have pilots, we won't need a cockpit, and the airlines would be able to sell some seats with a pretty cool view up front. 

[~Craig~ 10]

A good example of hand flying in a large airliner can be seen on one of the Just Planes videos; Atlas Air.   It's a Boeing 747-200F flight and the co-pilot, who is PF, disconnects the autopilot and autothrottle at 24,000 feet (mid-descent) and hand flies the entire remaining descent, approach and landing, through busy US airspace.