Originally Posted by
vilas
In modern aircraft automation dependency is encouraged because that's the future of aviation. A350 ATTOL is an indication of that. A350 also automatically takes care of unreliable speed and loss of pressurization. The AP is also available with dual engine flameout. So in rare case of AP loss you are expected to keep your ability to fly a few type of approaches or ability to execute basic manoeuvres like climb, descent and visual circuits. You are not expected to handle everything and every situation manually. Besides no matter how proficient a human gets he can't get rid off human factors. Most accidents are due to human error and by experienced pilots at that. One Sully or Al Haynes don't make a summer. So technology will, if it can't replace(eventually that will happen) want to restrict humans to do limited tasks. So make the best of that is there and have fun while it lasts.
That's an interesting argument.
In the future, you can envision that aircraft will be able to manage every trajectory.
But, first, this does not guarantee at all that it will work. The MAX had a system that was so well designed to assist the crew that it sent two airplanes down in less than six months (with a fleet of 200 and a few aircraft).
Second, aircraft as of right now still need pilots, because LVPs which allow autolands seriously degrade aerodrome capacities.
Plus, even in the case of a perfectly functional aircraft, a car passing through the ILS critical area can send an airplane sideways.
And as of now and also in the future, we would still need pilots to monitor the aircraft in case it does wrong.
Since humans are better at doing than monitoring, and even more so when they practise frequently, some say that it would be more relevant to maximise the amount of manual flying, and assist the human with the machine, not the other way around.
Or, are you saying that in the future, pilots will "take over" in a different way ? Instead of going from AP ON to AP OFF, they would degrade the managed modes into selected modes ?
That could be a possibility, but if you have pilots that aren't trained for normal flight handling, in case the AP goes wrong (even in selected modes) than essentially anything can happen to your aircraft.
We are already seeing a new type of crash where an almost perfectly good aircraft goes down even in CAVOK conditions.
Turkish at EHAM, Asiana at KSFO... This fact is what pushes airlines into saying that basic stick and rudder skills should be worked on.
Originally Posted by
Check Airman
I’ll agree that a wider range of movement would be ideal- but you get used to it. Better to sort it out on a normal day, than try to learn it when it’s not performing satisfactorily, or is broken.
It looks like I agree with you once again.
At first I was a bit unprecise with these levers but it gets better the more I use it.
Originally Posted by
EI_DVM
I also think it's important that we don't get drawn into an "The automation can do it better" debate. I'm fully aware that it can most of the time, there are few things I can do better than an autopilot, save for perhaps thrust control on a stormy night where certainly the Airbus' A/T lets itself down frequently, but for the most part I know I am at an absolute best, as good as the autopilot, and in most cases not so. The importance comes in being good enough, being able to fly within the tolerances, being able to correct deviations before they develop etc.
It is obvious that automation can do better.
It has a far better processing power than the human pilot. Can you calculate with any given set of speed, heading, final course, and wind the anticipation distance or angle to start the final turn onto the ILS ? I can with the help of excel, sitting at my desk. In flight I can't. I try to (and could discuss some techniques here) but the computer will have much more precise results.
Can you look at all flight parameters at a fast refreshing rate ? The human pilot can only do so much, one refreshment of all parameters per second, or so, maximum. I don't know the figure for the computer but much more obviously. They also have the intelligence of many engineers encoded into them to know the ideal amount of correction related to an amount of deviation. Once again, I can try to have a method to calculate this sort of thing in flight, but in a much simplified version (I use : on final, 1° of pitch up or down = 3% variation in thrust. Speed trend at x kt per second : plus or minus x% thrust. It works quite well but it is a simplification of reality whereas the aircraft can use a model accounting for a wide variety of parameters, wind, mass, configuration, etc..)
But I can work without the radioaltimeter, and George can't.
And if George wants to capture a false glide with +22° positive pitch, it fortunately never happened to me but I hope I would refuse to pitch up (and to this extent) at a moment when I'm supposed to descend 3°.