Comment by NikolaNovak | Hacker News Reader

Comment by NikolaNovak | Hacker News Reader

Parent Full thread

People's instinct and intuition are EXTREMELY incorrect in situation of your sight obscured while flying an airplane (and untrained and even trained personnel are extremely ill-equipped to judge their spatial awareness and capabilities in blind situations).

People end up trusting their body reflexes that have evolved for very different situation and absolutely lie to you. When an airplane turns, the actual G forces upon your body are very different than, for example, in a car. So the body feeling we have that we are turning or going straight in a car, misleads us when we are flying.

When you learn to fly, typically instructor will have a lesson where they obscure your sight (usually something called a "hood"), and let you fly by the seat of your pants for a minute or two. Typically, after a little while, when asked, student will be certain they are flying straight and level. Instructor will then remove the hood and demonstrate they are e.g. in a steep turning dive - something that would result in collision if allowed to continue.

Airplanes themselves do not crash after 180s (in fact, most small GA airplanes are designed to level themselves if left unattended). People crash them after a couple of minutes with sight obstructed :<.

(the Instrument Flight Rating is similar in USA & Canada as well, FWIW - you need to obtain that additional certification on top of your private license, to fly in clouds. Otherwise you are formally and legally constrained to VFR - Visual Flight Rules. There are extremely specific details on how close you are allowed to get to a cloud in every dimension, when you are and aren't allowed to fly, and what you must do if you cannot obey. These are drilled, tested, examined, and taken seriously)

Probably a stupid question, but why can't these lost pilots just maintain their attitude indicator in the middle, fly straight, grit their teeth, and wait to get out of the cloud?

Not at all stupid! Very real question that's been asked by aviation safety agencies around the world for decades!

Sometimes there are genuine issues with instruments. As well, note that most GA airplanes operate their crucial instruments and gyros via vacuum pump and they need to be re-calibrated after a while even in regular flight.

But I believe much more frequently, inexperienced humans distrust their instruments and trust their instincts/guts/feelings/body more. It's very very unnerving for inexperienced pilot to lost their sense of sight, ALL of their body is telling them they're turning and need to correct it, and some 1972 vacuum gyro is telling them they're going straight and don't touch anything.

Other times, things just accumulate, little drifts and changes and influences. Let's say that you're in cloud and trust your instruments, but after a while for whatever reason your VSI (Vertical Speed Indicator) is showing that you're descending 500ft/minute. What control inputs do you provide to correct that, without any visual references? There are right and wrong answers and integrating the basic six instruments, building correct mental picture of your position and angle and direction from these very limited and separate instruments, without visual reference, for long periods of time, is more tricky for untrained personnel than may be immediately obvious :-/

Not a real pilot, have never piloted a real plane, but I love my sims and have put considerable resources into building my aluminium profile simrig.

I was practicing increasingly denser fog/low visibility ILS landings of the F-16 in DCS, on my TrackIR'd 32"x3 triple screen sim rig (with transducers for buffeting sensations etc).

With triples (and much more so with VR, I'll transition to a Crystal "soon") you do develop a semi/meta(?)-"physical" connection with the sim after a while.

At a point visibility got so reduced I would completely experience vertigo. Even though I told myself it's only a sim, just look at the instruments (and the F-16 has so many, including HUD with the ILS glidescope), my head would spin out and I'd completely bork the landing.

Only after several more hours was I able to do it (with visibility so bad that the runway would only pop out when right on top of it metres from the ground). It would take a lot more to get comfortable with it, but by that point I had enough and - I still have nightmares about it. :)

Can completely understand how cloud based vertigo in a real plane with your life on the line has killed and will continue to kill so many. Mad respect to the real pilots here who have mastered this.

The thing that a sim never can simulate (not even a full-motion sim) are the various ways the vestibular system can get confused by e.g. the coriolis force when you move your head while in a turn that makes your feel like you're turning around another axis.

Most sim pilots can fly on instruments without a problem, because it's precisely the conflicting information between the instruments and your vestibular system that makes it easy to get disoriented. Flying a sim, you'll never have your ear telling you that you're leaning on your side even though you're flying straight and level, or vice versa, so you'll never be tempted to side with your vestibular system.

In a real plane, just turning a tight turn and suddenly looking down on the floor because you dropped something can induce conflicting vestibular information that's quite disconcerting even in day VFR.

It's not really a stupid question. In theory, you could do that, but it's not that easy in practice. First of all, the attitude indicator alone is not sufficient -- you need to pay attention to the entire instrument panel, including airspeed, vertical speed and turn coordinator, and mentally synthesize a complete picture of how the aircraft is moving.

Moreover, anybody who can fly a plane should intellectually know how to read an attitude indicator. But it takes a lot of training to be able to trust that indicator, when you have powerful sensory evidence that makes you instinctively feel the aircraft is doing something that it's not.

See also: https://en.wikipedia.org/wiki/Sensory_illusions_in_aviation and note the long list of accidents at the bottom.

Quoting from the Wikipedia article:

[Coriolis illusion] can produce an overpowering sensation that the aircraft is rolling, pitching, and yawing all at the same time, which can be compared with the sensation of rolling down a hillside.

Oh, wow, ok, that sounds strong indeed... I'm starting to understand it would be super hard, while vividly feeling like I'm rolling down a hillside, to try and tell myself: "oh, I just need to focus on the instruments, and trust them..."

Exactly! Precisely it.

It's hard to imagine just how strong the feelings are from our current position of at home and typing on phone or computer, but it really genuinely is your entire body completely convinced one thing is happening, while in fact something else is happening.

(Note this is part of the reason why even with full visibility so many people get nauseous - subconsciously what your body thinks is happening and what your eyes tell you is aftually happening, disagree - sometimes strongly)

Not all airplanes are required to have attitude indicators. If so, the pilot is expected to just look outside . . . and also to be responsible enough not to fly into instrument conditions.

You can do that. It takes a ton of discipline and practice to actually fly off the instruments and not feelings. It's really disorienting to fly instruments only.

Mind you that level flight through clouds only works if there are no buildings or mountains ahead, and keeping a straight flight path is even harder.

As a glider pilot, artificial horizont is very rare. All you got is speed, vertical speed, altimeter.

Once blind, your first prio is to keep the speed and vertical speed indicator constant. The problem is, without artificial horizont you cant really tell your bank angle, and your feelings just straight up lie to you. So you think you going straight, but suddenly your speed starts increasing. So you think "I am prolly nose diving a bit", and pull to compensate, only for speed to not decrease as much as youd like (or not at all), and suddenly you feel you are pulling 3g, and then you notice your altitude is dropping like crazy.

Thats how you get into diving spiral, and worst thing is you still have no idea if you are spiraling left or right. Keep it up for few more seconds and you can easily exceed safe speed, which given the g-forces is much lower than actual maximum speed in clean air. And a second or two later your first wing decides to go different direction than the rest of the plane.

I tested this with my instructor once, where we flew into a cloud just slightly above its bottom edge, with the goal of "try to fly trough it", only to find myself spinning out of the bottom with exactly this situation. We were prepared with brakes, lots of space beneath us and cleared general area of other planes, and checked beforehand theres no themral beneath to suck us higher, but that memory clearly showed how unintuitive and dangerous flying trough clouds often is.

Are the clouds ground-level? I'm just wondering what happens between exiting the cloud and impacting the ground.

Sometimes they can be. A lot of times people without IFR enter clouds because situation is worsening - so after a few minutes, cloud/fog might reach ground, or you might fly to a place where clouds are lower.

Other times, you get your airplane in a situation / position where, when you do exit a cloud, you don't have enough altitude to correct it. Some airplanes have various types of spins you can enter that you cannot exit or cannot exit quickly. As well, people can remain very disoriented when they exit the cloud if they are in a spin, or they can panic and provide incorrect control input.

Think about cars - people can accidentally induce understeer or oversteer when they encounter unexpected lose surface, and unless they've had training and practice, they frequently will do the wrong thing even when they're aware of situation and problem. Similarly, people can exit the clouds in a stall or spin, and due to panic or incorrect thinking or inexperience or even lack of oxygen, they can continue producing incorrect input. For example, if you're in a simple stall and losing altitude - the absolutely correct thing to do is to point your stick/airplane DOWN, gain speed, and then regain control. But it takes practice and conviction to do that when you're close to ground and falling. Spin recovery can be more intricate (depending on the airplane), with stick, rudder, and throttle inputs that your primitive brain may not register are conducive to immediate goal of "don't hit ground". It's one thing to practice things 4,000ft in the air on a clear day with instructor with you; and another when you're 600 ft from the ground, your airplane is spinning, you can barely ascertain which direction it's spinning, the sound is overwhelming, passengers are screaming, and you're panicking :<

Even extremely qualified airline pilots have numerous times provided textbook incorrect control input in critical situations :(

Even with all these explanations I'm still totally unclear on how flying in a cloud gets you an 86% fatality rate. It's difficult and counterintuitive, things can go wrong with instruments etc.., but what exactly makes a crash a near-certainty in this situation?

Hmm, if you've already read the previous and especially parallel posts in this thread, I'm not sure what other examples or analogies to use, so somebody else would need to try another tack. Basically - people's sense of direction actively lies to them, extremely strongly, so they do the wrong thing, and they crash the airplane, virtually without fail.

As per my previous analogy - people without explicit driving instructions and practice (you must build actual reflexes, not just theoretical knowledge!), will overwhelmingly do the wrong thing when they understeer or oversteer a car - a machine they may operate daily otherwise and still have full sensory inputs. Flying in clouds / entering a spin is exponentially more complex and counterintuitive and punishing :<

If that still feels incredibly unlikely, and please take this in the kindest possible way as an illustration and not personal :), that's exactly what kills people! The profound and pervasive unawareness of how dangerous the situation is, and just how much your body senses will lie to you and how much following them will kill you.

(fwiw, note it's not "sum total of all people entering the clouds results in 86% fatality rate", of course - airlines go through clouds hundreds if not thousands of times around the world daily! But for untrained people to enter a cloud, especially inadvertently / unplanned, IS deadly)

Reply on news.ycombinator.com