The Case for Remaining in the Loop

[scandinavian13 5,192]

The following post has no images for now. I'm hoping to remedy this when I get home and can hopefully pull excerpts from the original flight release to give more information.

 

 

Today is my final day at my current company before I return to the aviation realm full time. That being said, Now that I've handed off all of my work tasks to other members of the team, I finally have some time to devote to sharing a situation I'd wanted to bring up for quite some time.

 

About a month ago, I was planning a flight out of VHHH to OMDB. Since FlightAware doesn't have full coverage for that area, I couldn't steal a flight plan from the real world flight. I'll admit it. When it comes to flights in the United States (and those originating from it), I'm pretty lazy when it comes to route selection. Luckily, I have a copy of PFPX to help me route myself around in the rest of the cases.

 

If you've watched the video I put up a long time ago about how to manually route yourself using some pretty simple resources, you'd know that my first step is to draw a Great Circle route between the two airports. From there, once I find out the SID exit and STAR entry points, I usually work in reverse - from destination back to the origin. One of the drawbacks to this is that the method isn't necessarily optimized. PFPX, though, offers an option for it to calculate a wind optimized route, which usually results in a lower fuel burn, despite the occasional increase in miles flown across the ground.

 

Utilizing this feature for the VHHH-OMDB leg, I got a route that did not deviate too far from the Great Circle direct routing, while still remaining on the airways. Trusting that it was giving me the best wind route it could, I had it compute the flight. The charts on the release showed the route was more into the wind than I would've preferred, especially given the fact that the wind was much calmer to the south. Seeing this, I went over to SkyVector (its map pans/draws a little faster than PFPX simply because it doesn't have all of the extra features), selected the World HI charts, turned on the wind barbs, and dragged the magenta line across a route south of the computed route and into calmer wind. After getting the route to where I was satisfied with it, I moved it into PFPX (manually typed the airways/points in - you're not missing any automation, don't worry).

 

Computing that updated route gave me a shorter time en route, and a lower total fuel burn. One thing you may not know about PFPX is that, for each run of the computations, it creates a new record on the results tab. This way, if you're presented with a scenario much like this one, you can see the each option side by side to choose the one that best fits your needs and release the flight with those parameters, whether it's a different route, passenger load, fuel load, and so on. While my intention for this post is not to be an ad for PFPX, I did need to point out that the program does allow you to run multiple scenarios to help determine the best fit for your operation, much like you'd have if you plotted and computed everything by hand on separate pieces of paper.

 

This would also be very useful in the event your normal route passes right through a large frontal cell, but it has two equidistant gaps on the front. Comparing the north route against the south route might help determine which one would be better. While the north route might be shorter in overall distance, it might have a higher tailwind, or lower headwind.

 

 

 

So why is this here???

 

 

 

One of the things I've noticed in the flight sim hobby is an increased reliance on automation over the years. Occasionally, I'll come across posts where people indicate that they've never truly learned how to land one of our planes, preferring to let it autoland itself. To me, stuff like that blows my mind. When I started to use the sim, stuff like that wasn't even an option. It was very much a "keep trying until you finally succeed" kind of thing. [insert witty "well back in my day, sonny" joke in an addled nonagenarian voice.]

 

Of course, that's not to cast aspersions on the sim community. The real world aviation community has been discussing over reliance on automation for years - decades, in fact. Remember that "Children of the Magenta" video? It's two years short of being 20 years old. That's older than, or about as old as, a decent number of people in this forum.

 

Regardless, our 777 offers a very high level of automation. I found it ironic that, for a flight in an aircraft that nearly bores me to tears in most phases of flight with all of its automation, I had to step in and manually adjust my routing to get a better end result. Automation is great. When things are "behaving," it decreases workload significantly. Occasionally, though, things get just outside of the standard tolerances enough to throw the automation off. In those cases, there just wasn't anything that could have been done to anticipate the circumstance, or it would've been too code-intensive to implement a solution for it.

 

Using my frontal storm case above, it's possible to program in a function to avoid radar returns, captalize on tailwind (or decrease exposure to headwind), and take into account a host of other factors. The issue, though, is that programming a flight planner requires setting certain boundaries in order to avoid the planner attempting to evaluate every single route combination around the world. Sometimes, a more efficient route lies just outside of these bounds and it usually takes a little bit of human intuition to recognize that.

 

If you're flying around and get the feeling that there's a better way to do something, all it takes is a little thought and messing about with different functions to find a potentially better solution. When flying a departure out of an airport that relies on a DME distance from a particular fix, many would try to force the point onto the LEGS page. Others would try to put it onto the FIX page, and use a /[Distance] entry. For those flying with the NGX and have the HUD, it might be best to use the NAV radio, as it would be displayed at the bottom of the HUD, avoiding the requirement to look down at the ND. Are any of those wrong? Not at all. Do they have varying levels of convenience? Absolutely. Find the level of automation that best suits your operation. Sometimes, that requires you to re-insert yourself into the loop and take over the situation yourself.

 

 

 

TL;DR: Automation is great. Occasionally, though, you can be a lot greater than the automation just by using your own internal CPU.

[MarkW 1,491]

A couple of comments.

 

First on the flightplan, I really like to "steal" the days plan from Flightaware when possible.  I have not needed to do a complete plan on my own for a while but have had to build part of the plan when flying from Europe to NA and Flightaware cuts off the flightplan somewhere over Ireland.

 

Second, on the automated landing.  I was flying on Vatsim last night KDFW-KATL in the -800 (using the lastes FSX-SE release).  There was another pilot I heard over the radio flying into Nashville that was instructed by the Controller to "expect visual runway XX".  The pilot sheepishly replied...."I need the ILS, I don't know how to land manually".  I was thinking how sad it was that this pilot was missing out on such an enjoyable (and learnable) part of simming. 

[Budbud 227]

Hi,

 

I fully agree with you post Kyle. Never rely entirely on the automation (but don't get paranoid either! ). Working with automation always requires monitoring and double check to ensure nothing gets wrong in the hidden.

 

As per PFPX and route calculation, sometimes it doesn't give the best route even when selected the wind optimized option. 

One example is the flight I'm doing now (KLAX-SBGR): PFPX gave me strange route going south over Mexico, then getting around the Yucatan and almost straight to Sao Paulo over Colombia and Brazil.

After checking the winds, I had a silly idea to test a route much more to the East (over Florida as there is a powerfull Eastward jetstream over South of USA).

So I drew a new route going straight East from LAX to the Florida, then South East to fly over Venezuela and Brazil.

 

I then checked at Flightaware and found out that the current flight was actually doing this as well! (Quite pride of myself on this ).

 

Conclusion, that route over Florida was a bit over 200nm longer than the straighter route (over 5700nm) but it saves 5min of flight and approx 1000kg of fuel! 

 

So as good as PFPX is (and my point is not to criticize it really), it shows the limits of the automation against the human capacities to find solutions out of the obvious. The automation is and will probably be for long bound to parameters programmed and has very low capacities if any to improvise or "think" differently.

[Don t moderate me this way 882]

There was a T7 pilot in the real world who became nervous when learning that the G/S at his destination was INOP ...

 

The great thing about flightsimming is you can learn, try and test everything you want. (Even autolands!    )

I don't do autolands to avoid flying myself, but to explore and appreciate the efforts that have gone into those specific bits and bytes of code.

I fly an Airbus today, and a Boeing tomorrow, just to learn about differences in system concepts and system details.

 

Put bad pilots in a fully-functioning Airbus and the automation might save the day.

Put two very proficient pilots in a malfunctioning Airbus and they might be having a tough time to keep the Airbus from killing all people onboard. (That's "normal law" in Airbus terms. In order to regain control, Airbus pilots have to switch certain 'puters off. That's "alternate" or "direct law". But "normal law" rules in Airbus flying. So the 'puters rule, not the Airbus pilots.)

Wish the pilot from my opening statement would have flown an Airbus. Maybe three more people would still be alive today. So it's all about clever automation, isn't it?

The Airbus company wishes pilots would do more hand-flying. So do I. So it's about not so much automation, but more proficiency, isn't it?

 

To me it looks like the demand for pilots in RW aviation exceeds the supply of proficient pilots. Instead of increasing proficiency among pilots, we put them on automation. Problem solved - or ... isn't it??? :unsure:

[robert young 2,897]

A recent report by an Airbus exec implored Airbus pilots and airlines to step up training for manual control and encourages Airbus pilots to routinely do more manual flying.

 

I find it incredible that any commercial pilot would feel uncomfortable if a glide slope is out of action at an airfield presenting easy VFR condtions, but it appears to be increasingly the case.

 

Something has gone wrong with ATP training for the last 20 years, but it seems to affect some nations more than others. I've shared a cockpit with a number of European (particularly UK) commercial pilots and all of them were exemplary in their control of almost anything they fly, including gliders and GA types. But I'm absolutely certain that simulator procedures in many countries simply do not substitute well for stick and rudder skills which seem to be in serious decline.

 

When I learned to fly I was put in a tail dragger and my first cross country was without a radio, and no navigation aids apart from a map and a £5 manual flight computer to plot the route. While no-one would suggest returning to radio-less, GPS-free training, there is no doubt that contemporary commercial pilots in some regions are woefully inadequate when it comes to basic aviation skills. At least five major disasters have occured recently through simply failing to assess what is happening with the aircraft when an auto-failure occurs, or when common sense screams at you that something is not quite right.

 

Many sim pilots running FSX and P3d are very good at procedure, but looking at many youtube examples, hand flying doesn't seem to be anywhere near as accomplished. The growing demand among simmers for ever more gizmos and sophisticated Flight Management functions serves to demonstrate that many see "authenticity" merely as packages of auto-sophistication rather than their aviation skills and stick and rudder technique, along with sim aircraft responses to them, which to my mind are much more relevant indicators of authenticity.

[Budbud 227]

I watched recently a documentary about air safety and aircraft accidents. I don't how accurate the documentary was but they said that the increased automation in aircrafts has several side effects:

_ It would lead pilots to have less vigilance/awareness due to the boredom and over-confidence in the automation

_ It would increase the work load in that it would require to have more knowledge to fully understand the way the automation works (?)

_ It would lead to descrease pilot skills in aircraft manual handling because autopilot is used in most stages of the flight and therefore pilots would have fewer opportunities to train and keep their flying skills.

 

I'm not a rw pilot but I can easily understand that it could be easy to get lazy and let the autopilot do the most work instead of actively flying the aircraft. I personally only perform autoland when performing very low visibility approach when I think it would be done for real. 

 

In an other field, we have that kind of downsides in my job with automation. I'm seaman on offshore vessels and with have now a lot of automation for positioning the vessels called Dynamic positioning system which is able to maintain a fixed position of the vessel by the means of calculators and thrusters. This system is a requirement for most of the works we perform (work on subsea installations with divers and ROV).

The main downsides of that system are that the operators on board are mainly in a position of monitoring the system and only interact with it through buttons and screens (just like autopilot and FMS). However some of the operators have never had the opportunity to manoeuvre the vessel manually and wouldn't be able to maintain the position manually if the system fails.

Also, some of them tend to distract themselves with electronic pads and cellphones while staying hours in front of the screen monitoring. It is something we don't allow for obvious reason, but I notice that it is less and less obvious and sensible for people now to refrain themselves from doing this (even when lives are at stake, divers in water) as the boredom potentially takes the lead in their mind over the conscience and because they often are over-confident about the system.

Why would the system fail when it's keeping the vessel position for a week without any issue?

But they also fail to consider external parameters such as the weather conditions and that lack of attention can lead to dramatic issues.

 

It seems that such issues occur more and more in the aviation too with pilots failing to catch change in situations while in autopilot (plane losing few thousand feet without any pilot noticing it, pilots missing the destination airport because their attention was focused on ipad...)

In the documentary stated above, they said they did an experiment with pilots sitting for hours on a cruise stages of a flight. These pilots were required to count the number of aircrafts they would sight during the test in a normal flight condition. Most pilots counted only half the numbers of aircrafts that were actually visible. The results were explained by the fact that the boredom leaded to a lack of attention of the pilots who would have a tendency to miss information they should normally catch. 

[downscc 1,327]

 

 

In an other field, we have that kind of downsides in my job with automation.

 

Reminds me of an event that happened a long time ago at a local refinery (name withheld) after we (the engineering company I worked for) upgraded process unit controls from old pneumatic to then-new computer-based electronics (DCS).  Instead of watching a board full of instruments with live data, the unit operators now sat in front of a computer screen and 'monitored.'  Events such as alarms were annunciated by exception much like the center screen MFD in an aircraft.  One day there was a problem, the screens froze and nobody noticed it for 24 hours... funny now but it could have been real bad.  Of course now there are activity monitors such as the rotating circle in Windows to indicate things are not frozen.  The operators were complacent.  Like a fire station, their job involved exercise and cooking until there was an alarm (unless in training of course).

[kevinh 409]

 

 

Put two very proficient pilots in a malfunctioning Airbus and they might be having a tough time to keep the Airbus from killing all people onboard. (That's "normal law" in Airbus terms. In order to regain control, Airbus pilots have to switch certain 'puters off. That's "alternate" or "direct law". But "normal law" rules in Airbus flying. So the 'puters rule, not the Airbus pilots.)

This is a bit misleading. In normal law the Airbus pilot is still in full control, within envelope limits of course. It's only under specific malfunction conditions where garbage in = garbage out comes into play that they must force the aircraft into alternate law. For example unreliable airspeed. But this is a killer situation for the unwary in any aircraft including Boeings. So any aircraft left in automatic control can kill you in malfunction circumstances.

[voske 36]

I really enjoy creating my own routes with PFPX. I can spend way too much time selecting, for example, the NAT, the North American Route, and the UK standard route for a transatlantic flight.

 

When it comes to flying, I rely on automation a lot, especially during the approach. Main issue for me on approach was and is aiming for the runway. When I look out the window, I have no clue whether I am heading for the runway or to the left or right of the runway. Perhaps it's because of my lack of depth perception, I don't know. I've tried sticking a pen or a pencil on my monitor, just to create an aiming point, but that didn't really help. I came to prefer to autoland, because that would make sure that I wouldn't ruin my flight by botching the landing. For me, it didn't matter what it looked like out the window: if flying manually, I had to rely on the FD to get me down to the runway.

The HUD on the 737 was a major development for me, because for the first time I could see what I was doing.  Nowadays I'm comfortable manually flying any type of approach with the 737, including visual approaches. On other types, I'll stil follow the FD down to the runway but almost always manually from about 1000 feet above field elevation. I still botch landings (often because of bad flaring), but not as much as a few years ago.

[Don t moderate me this way 882]

(...). So any aircraft left in automatic control can kill you in malfunction circumstances.

 

 

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.

[77west 139]

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

[scandinavian13 5,192]

 

 

Thanks, Kyle, for starting it.

 

Welcome. I'm actually really impressed with the responses/contributions that have come of it.

 

I figured most people would look at it and think "what's he on about today? Eh, too many words..." *close*

[Cai Zicheng 171]

By mentioning "at my current company", you do not mean PMDG right?

[LAdamson 240]

Since "Children of the Magenta Line" was mentioned.............................I must admit, I can't stand it's relevance to today's modern navigation, when it's brought up in reference to use of glass panels. It still often is, and yes, it's getting close to 20 years old!

I think of it, as information and automation. "Children of the Magenta Line" was produced just a few years after an American Airlines 757 flew into a mountain.  In this case, there was too much reliance on automation, but the systems were primitive by todays standards, and didn't offer a portion of the information, that can now be obtained with just a portable GPS.   The pilots of the AA 757 never had what I call the big picture. We can now see where we are exactly within a few feet, see the terrain mapping for hundreds of miles in all directions, and have a constantly updated weather map overlayed at the same time. The 757's radar altimeter was only effective for around 14 seconds, before impact. The modern GPS will be showing rising terrain, and potential conflicts far in advance. Synthetic vision makes it all the better!

 

Over the years, I've done a lot of research into many flight into terrain accidents. In some cases, there wasn't an automation problem. It was navigation mistakes, in reading the sectional/approach/departure charts. Usually someone who wasn't too familiar with mountain flight. The Navy C-130 crash out of Jackson Hole, Wyoming, is a good example.  A few years back, a multiple engine fire-fighting plane was lost on the other side of the mountain from where I live. They were even carrying a small Garmin portable, that was turning the terrain depiction red with X's. The first officer flying, mentioned this to the captain, but the captain didn't seem to hear, and was giving a geography lesson of the local area. They were basically scud running in IMC. The plane hit the mountain at 240 knots.

 

For me, I do believe that pilots need to remember to hand fly, and why. I just don't agree with anyone, who suggests that the GPS should be left home once and a while, for they they call a return to basic airmanship. It's not basic at all. The VOR/DME system just replaced an inferior navigation system before it.  IMO, there is no reason that airplanes have to continue to fly into mountains because of human mistakes.  The modern glass sytems will give pilots plenty of advance warnings before the inevitable. 

[~Craig~ 10]

Great post Kyle.  Scenarios like that do indeed make the simmer think about how much we rely on the complex tools and systems available to us, and how we can be left up s*** creek without a paddle, if those systems 'fail' (not always mechanically!), if we don't have a 'back up plan' that involves using our own grey matter. 

 

 

I figured most people would look at it and think "what's he on about today? Eh, too many words..." *close*

 

Seriously, definitely not! ..... You know I've been a bit critical of you in the past, but this was a brilliant post, containing pearls of wisdom that both makes us think, but also is a call to action about how we can use the sim more to challenge ourselves, and our responses to the unexpected; the unplanned...      

 

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.