> The only victim of the accident was Wim Delaere, a computer science student reported to have been either 18 or 19 years old.[4][5][1] He was sleeping alone after celebrating the end of his university exams the previous day when the MiG crashed and killed him at 10:30 am. His mother and brother were shopping for groceries in Kortrijk, and his father was working in Ypres.[4]
From the linked Wikipedia article on one of the answers.
I recently learned about the Green Ramp disaster, where the crew ejected from an F-16 under full afterburner, and the jet continued on to collide with several parked airplanes, resulting in 24 fatalities.
"As of 2025, this incident has the largest number of ground fatalities for an accidental crash of an aircraft on U.S. soil. It was also the worst peacetime loss of life suffered by the division since the end of World War II."
It seems like the sensible thing to do would be to fry / erase any IFF and encryption related stuff, but otherwise continue as before.
E.g, if it's already been programmed to fly straight and level, continue to do that. If it's deactivated, stay deactivated.
Just seems like a whole 'nother set of characteristics to test otherwise, as well as adding extra unpredictability. The aircraft is probably damaged / on fire, so its flight characteristics are already going to be extremely different to normal. The best thing in the moment may be to let the aircraft lawn-dart in a field, rather than attempt to get straight and level, and in the process potentially fly over inhabited area or towards a friendly set of aircraft / buildings / vehicles.
If the autopilot is engaged, the pilot won't be ejecting, because the aircraft will be in some kind of controlled flight. Autopilots will be disengaging and lighting up a big red light in the cockpit well before the aircraft gets to the point where the pilot would consider ejecting. Remember that ejecting is an absolute last resort, since the pilot is quite likely to be injured and runs a significant risk of being killed in the process of ejecting.
But more than that result in injury. The possible injuries are pretty severe.
> Compare to 20 in 20 jet airplane crashes resulting in death
Crashes if the plane is totally uncontrollable, probably yes.
But there's a lot of gray area in between "totally uncontrollable" and "controllable enough that an autopilot can fly the plane". There are plenty of cases where a pilot was able to make a controlled enough crash that they walked away from it, even though the plane itself was totalled.
And once we get to the point of "controllable enough that an autopilot can fly the plane", the pilot would have no reason to eject--because the plane is controllable enough that the autopilot can fly it. Which means whatever problems exist can't be very severe--or the autopilot would be disengaging, because it needs things to be working pretty well to fly the plane at all. That was the point of my response in the GP to this post.
A pilot would only eject if the aircraft was uncontrollable with no reasonable hope for recovery. Unlikely the autopilot can do anything deliberate at that point.
At the very least, something to keep the flight path predictable would make sense, to give the pilot a chance to point the plane at a "safe" area to crash in before pulling the lever. I remember reading several stories of pilots taking "where is the plane going to crash" into account for their ejection decision or last moves before ejection.
There were also several incidents where a pilot ejected because the plane was somewhat controllable but it was clear it couldn't be landed safely. At least one of them where they had tens of minutes of controlled flight before ejecting (they flew it over the ocean to minimize the risk of collateral damage).
Unlikely, you say? That's why it's being discussed. Like, if a pilot ejected because the engines stop, but the control surfaces still work... Maybe the plane avoids a kindergarten on its dive to the earth-sky interface.
Huh? If I'm the human pilot, I can pull the ejection lever for multiple reasons, including my just being an idiot. The plane, after I eject, should do something reasonable. Maybe it
* starts broadcasting a mayday?
* crashes into the nearest large body of water?
* attempts to fly itself back to base (we have the technology)?
I mean, it has to do something and flying straight and level until it runs out of fuel is unlikely to be the optimal value of "something"
Why would it be controversial to say "Look, guys, we should decide what the plane does after the pilot ejects. Maybe the best policy is just flying same course and speed until fuel exhaustion, but we should choose this policy, not default into it without consideration."
> The plane, after I eject, should do something reasonable.
There’s a good chance that it can't, and its not impossible that trying to do something reasonable combined with damage that led to and/or resulted from ejection could make things worse.
> starts broadcasting a mayday?
Great idea for peacetime over the homeland, maybe a very bad idea for military operations over contested or enemy territory.
> crashes into the nearest large body of water?
> attempts to fly itself back to base (we have the technology)?
If either of these are useful in a nontrivial share of ejections (except perhaps the former in conditions where it takes no special effort), then there is a serious problem with the training of the people pulling ejection handles and that needs to be fixed, rendering the action not valuable.
> Why would it be controversial to say "Look, guys, we should decide what the plane does after the pilot ejects. Maybe the best policy is just flying same course and speed until fuel exhaustion, but we should choose this policy, not default into it without consideration."
Because ejection is an action chosen when you can no longer meaningfully say what the plane does in any significant way. That’s the whole purpose. If it it is useful to address this question then you have a bigger problem that you need to urgently fix first.
Ejecting for for no reason would end the pilot's flying career. Ejecting for any reason will result in an investigation, at minimum. Pilots are expected to fly the airplane until the last extremity.
So while yes it's possible, it's unlikely, and the return on investment of making the plane able to do something like "return to base" in that circumstance would be a large negative number.
Even ejecting with good reason is enough to end a fighter pilot’s career. The rates for significant back injury are between 1 in 3 and 1 in 2 depending on the design.
Yes, but flying aircraft with ejection seats is demanding work and few humans are capable or qualified to do it. Most people can barely manage to drive cars safely.
I dunno, the current approach seems quite reasonable. In the grand scheme of things the overwhelming majority of the Earth's surface is empty space where a plane crash is unlikely to cause much damage.
You also have the complication that military pilots usually try to make sure their plane will crash in a "safe" area before they eject - many have died because they waited too long to eject trying to avoid a populated area. Giving the plan a mind of its own after they pull the handle would be unlikely to go over very well. I believe the scenario of a pilot ejecting from a perfectly good plane that keeps flying for more than a few seconds has only happened perhaps a dozen times in the entire history of aviation? Not really worth worrying about.
I'll do my best.
So you want to dedicate probably the rest of your career to automated diagnosis and recovery from crash conditions after ejection? Just so we can say we did a reasonable thing? Oh just the one case where the pilot rejects during level controlled flight you're saying we should be careful to let it continue on same course and speed? And if it's slightly changing course speed or altitude? Did we want to level out or continue the climb and turn? Do we attempt to maintain rate of climb even if it means throttling up? Descent?
The whole thing is so wildly ambiguous and niche that it's a black hole. When a pilot ejects the controller is gone. The controls are slack and it's just physics until fire.
If we had the technology to fly the aircraft back to base then there wouldn't be an ejection handle, there would be a "fly the aircraft back to base" button.
remember that part of optimization is the amount of resources spent developing a solution for a problem that just doesn't come up that often. in the microcosm of a single ejection there's probably a better way to handle it than to just let the plane continue on its course. in the macrocosm, there's probably better problems to deal with than the one that results from the relatively rare situation in the military and unheard of in the civilian sphere. it's also worth noting that ejector seats are explosive-assisted and any plane that's been ejected from is rendered structurally unreliable, and usually is so close to crashing that nothing can be done to save it even if saving it is viable. So most of what you do "in response" to an ejection isn't actually in response, it's about planning ahead. outside of a wartime situation where factors beyond your control tell you where you'll be flying, don't be in a place where it would be dangerous to bail if you think you might have to bail.
Even in that case this is a military aircraft, one of the most highly prized secrets many nations hold. Probably the most reasonable response to an ejection the aircraft could take is to nose dive into the ground and slam the throttle to afterburners, which isn’t far of what will happen naturally as soon as the pilot isn’t asking the plane to stay straight and level.
This is silly. And not true. There is no “would” other than your own prediction. What if the pilot deliberately wanted to crash the plane but not do it intentionally?
I was astonished at the claim that pilots ejecting can lose half an inch of height due to spinal compression induced by the G-forces. Claims seem to be borne out:
If it's a controlled ejection scenario, you try to fly to a specific location, airspeed, heading, and altitude, then pull. It will be in your local-area in-flight guide. The intent is, the plane ends up somewhere away from civilization. This if, of course, only suitable for scenarios where you have this luxury.
The analysis and conclusions of the responders here (2018) seem pretty invalidated by the 2024 F-35B ejection incident. Maybe more thought should be put into what the autopilot should do?
I think we should make an API call to an LLM with the current GPS location to decide what to do, bonus points if we can mount a forward facing camera and upload the picture as well
Thinking about this one (https://theaviationist.com/2025/02/12/ea-18g-growler-crashes...), it seems like after ejection you'd want the plane to lawn dart whenever possible. It allows the pilot to know if it is an okay place to ditch and it minimizes the reverse engineering risk.
I imagine there is a good reason this isn't the way things are though.
I think the most proper thing for the jet should be to destroy itself. In a war enviornment I would not like my enemy to gain intel about my military jets.
I think its high-speed collision with the ground or ocean generally takes care of destroying it. Especially with no pilot attempting to keep it level and slow it down and minimize damage.
Not my wheelhouse, but doesn't that involve packing the aircraft with explosives, and wouldn't that involve risk of blowing up if someone else shoots you? Or is there some better way to self-destruct?
One could calculate the amount of energy in a given amount of explosive and smashing into the ground at high velocity. I'm too lazy to do so, but I can tell you they're on the same scale.
A heavy object moving fast has a shocking amount of energy. When such an object impacts the ground, all that energy has to go somewhere.
The signal can originate on the plane itself due to a software glitch, stray currents in the grounding skin or whatever. If anything, jammers tend to interfere with the carrier's electronics itself.
> The only victim of the accident was Wim Delaere, a computer science student reported to have been either 18 or 19 years old.[4][5][1] He was sleeping alone after celebrating the end of his university exams the previous day when the MiG crashed and killed him at 10:30 am. His mother and brother were shopping for groceries in Kortrijk, and his father was working in Ypres.[4]
From the linked Wikipedia article on one of the answers.
What an unlucky kid.
https://en.wikipedia.org/wiki/1989_Belgium_MiG-23_crash
I recently learned about the Green Ramp disaster, where the crew ejected from an F-16 under full afterburner, and the jet continued on to collide with several parked airplanes, resulting in 24 fatalities.
"As of 2025, this incident has the largest number of ground fatalities for an accidental crash of an aircraft on U.S. soil. It was also the worst peacetime loss of life suffered by the division since the end of World War II."
https://en.wikipedia.org/wiki/Green_Ramp_disaster
It seems like the sensible thing to do would be to fry / erase any IFF and encryption related stuff, but otherwise continue as before.
E.g, if it's already been programmed to fly straight and level, continue to do that. If it's deactivated, stay deactivated.
Just seems like a whole 'nother set of characteristics to test otherwise, as well as adding extra unpredictability. The aircraft is probably damaged / on fire, so its flight characteristics are already going to be extremely different to normal. The best thing in the moment may be to let the aircraft lawn-dart in a field, rather than attempt to get straight and level, and in the process potentially fly over inhabited area or towards a friendly set of aircraft / buildings / vehicles.
If the autopilot is engaged, the pilot won't be ejecting, because the aircraft will be in some kind of controlled flight. Autopilots will be disengaging and lighting up a big red light in the cockpit well before the aircraft gets to the point where the pilot would consider ejecting. Remember that ejecting is an absolute last resort, since the pilot is quite likely to be injured and runs a significant risk of being killed in the process of ejecting.
About one in 20 ejections results in death, usually due to low altitude, or being hit/crushed by the seat.
Compare to 20 in 20 jet airplane crashes resulting in death and suddenly pulling that lever might seem a worthwhile risk to take
> About one in 20 ejections results in death
But more than that result in injury. The possible injuries are pretty severe.
> Compare to 20 in 20 jet airplane crashes resulting in death
Crashes if the plane is totally uncontrollable, probably yes.
But there's a lot of gray area in between "totally uncontrollable" and "controllable enough that an autopilot can fly the plane". There are plenty of cases where a pilot was able to make a controlled enough crash that they walked away from it, even though the plane itself was totalled.
And once we get to the point of "controllable enough that an autopilot can fly the plane", the pilot would have no reason to eject--because the plane is controllable enough that the autopilot can fly it. Which means whatever problems exist can't be very severe--or the autopilot would be disengaging, because it needs things to be working pretty well to fly the plane at all. That was the point of my response in the GP to this post.
This thread is a great example of how engineers by nature are tempted to add complexity to nearly any scenario.
A pilot would only eject if the aircraft was uncontrollable with no reasonable hope for recovery. Unlikely the autopilot can do anything deliberate at that point.
Planes that keep flying after an ejection do happen
https://en.wikipedia.org/wiki/Cornfield_Bomber
https://www.marinecorpstimes.com/news/your-marine-corps/2024...
These are not equivalent.
The first one, the Airplane was in an uncontrolled spin, the ejection happened to fix it.
At the very least, something to keep the flight path predictable would make sense, to give the pilot a chance to point the plane at a "safe" area to crash in before pulling the lever. I remember reading several stories of pilots taking "where is the plane going to crash" into account for their ejection decision or last moves before ejection.
There were also several incidents where a pilot ejected because the plane was somewhat controllable but it was clear it couldn't be landed safely. At least one of them where they had tens of minutes of controlled flight before ejecting (they flew it over the ocean to minimize the risk of collateral damage).
Unlikely, you say? That's why it's being discussed. Like, if a pilot ejected because the engines stop, but the control surfaces still work... Maybe the plane avoids a kindergarten on its dive to the earth-sky interface.
Not necessarily. For example, you'd eject if unable to get as safe landing configuration.
Yes, but the autopilot should have some kind of contingency programming in case the pilot is mistaken about the aircraft being unflyable.
If this was possible it would just be part of regular flight control laws and would be used to avoid becoming uncontrollable in the first place.
Huh? If I'm the human pilot, I can pull the ejection lever for multiple reasons, including my just being an idiot. The plane, after I eject, should do something reasonable. Maybe it
* starts broadcasting a mayday?
* crashes into the nearest large body of water?
* attempts to fly itself back to base (we have the technology)?
I mean, it has to do something and flying straight and level until it runs out of fuel is unlikely to be the optimal value of "something"
Why would it be controversial to say "Look, guys, we should decide what the plane does after the pilot ejects. Maybe the best policy is just flying same course and speed until fuel exhaustion, but we should choose this policy, not default into it without consideration."
> The plane, after I eject, should do something reasonable.
There’s a good chance that it can't, and its not impossible that trying to do something reasonable combined with damage that led to and/or resulted from ejection could make things worse.
> starts broadcasting a mayday?
Great idea for peacetime over the homeland, maybe a very bad idea for military operations over contested or enemy territory.
> crashes into the nearest large body of water?
> attempts to fly itself back to base (we have the technology)?
If either of these are useful in a nontrivial share of ejections (except perhaps the former in conditions where it takes no special effort), then there is a serious problem with the training of the people pulling ejection handles and that needs to be fixed, rendering the action not valuable.
> Why would it be controversial to say "Look, guys, we should decide what the plane does after the pilot ejects. Maybe the best policy is just flying same course and speed until fuel exhaustion, but we should choose this policy, not default into it without consideration."
Because ejection is an action chosen when you can no longer meaningfully say what the plane does in any significant way. That’s the whole purpose. If it it is useful to address this question then you have a bigger problem that you need to urgently fix first.
Ejecting for for no reason would end the pilot's flying career. Ejecting for any reason will result in an investigation, at minimum. Pilots are expected to fly the airplane until the last extremity.
So while yes it's possible, it's unlikely, and the return on investment of making the plane able to do something like "return to base" in that circumstance would be a large negative number.
Even ejecting with good reason is enough to end a fighter pilot’s career. The rates for significant back injury are between 1 in 3 and 1 in 2 depending on the design.
Do you realise we're dealing with humans? Humans who make decisions based on a multitude of factors, or sometimes none at all?
Yes, but flying aircraft with ejection seats is demanding work and few humans are capable or qualified to do it. Most people can barely manage to drive cars safely.
I dunno, the current approach seems quite reasonable. In the grand scheme of things the overwhelming majority of the Earth's surface is empty space where a plane crash is unlikely to cause much damage. You also have the complication that military pilots usually try to make sure their plane will crash in a "safe" area before they eject - many have died because they waited too long to eject trying to avoid a populated area. Giving the plan a mind of its own after they pull the handle would be unlikely to go over very well. I believe the scenario of a pilot ejecting from a perfectly good plane that keeps flying for more than a few seconds has only happened perhaps a dozen times in the entire history of aviation? Not really worth worrying about.
I'll do my best. So you want to dedicate probably the rest of your career to automated diagnosis and recovery from crash conditions after ejection? Just so we can say we did a reasonable thing? Oh just the one case where the pilot rejects during level controlled flight you're saying we should be careful to let it continue on same course and speed? And if it's slightly changing course speed or altitude? Did we want to level out or continue the climb and turn? Do we attempt to maintain rate of climb even if it means throttling up? Descent?
The whole thing is so wildly ambiguous and niche that it's a black hole. When a pilot ejects the controller is gone. The controls are slack and it's just physics until fire.
If we had the technology to fly the aircraft back to base then there wouldn't be an ejection handle, there would be a "fly the aircraft back to base" button.
remember that part of optimization is the amount of resources spent developing a solution for a problem that just doesn't come up that often. in the microcosm of a single ejection there's probably a better way to handle it than to just let the plane continue on its course. in the macrocosm, there's probably better problems to deal with than the one that results from the relatively rare situation in the military and unheard of in the civilian sphere. it's also worth noting that ejector seats are explosive-assisted and any plane that's been ejected from is rendered structurally unreliable, and usually is so close to crashing that nothing can be done to save it even if saving it is viable. So most of what you do "in response" to an ejection isn't actually in response, it's about planning ahead. outside of a wartime situation where factors beyond your control tell you where you'll be flying, don't be in a place where it would be dangerous to bail if you think you might have to bail.
Even in that case this is a military aircraft, one of the most highly prized secrets many nations hold. Probably the most reasonable response to an ejection the aircraft could take is to nose dive into the ground and slam the throttle to afterburners, which isn’t far of what will happen naturally as soon as the pilot isn’t asking the plane to stay straight and level.
This is silly. And not true. There is no “would” other than your own prediction. What if the pilot deliberately wanted to crash the plane but not do it intentionally?
I was astonished at the claim that pilots ejecting can lose half an inch of height due to spinal compression induced by the G-forces. Claims seem to be borne out:
https://www.forcesnews.com/news/can-ejecting-aircraft-make-p... https://www.quora.com/Do-pilots-lose-height-when-they-eject-...
Making sure that the consequences are well-publicized also has the useful side-effect of making sure that ejection is regarded as a last resort.
If it's a controlled ejection scenario, you try to fly to a specific location, airspeed, heading, and altitude, then pull. It will be in your local-area in-flight guide. The intent is, the plane ends up somewhere away from civilization. This if, of course, only suitable for scenarios where you have this luxury.
The analysis and conclusions of the responders here (2018) seem pretty invalidated by the 2024 F-35B ejection incident. Maybe more thought should be put into what the autopilot should do?
https://www.marinecorpstimes.com/news/your-marine-corps/2024...
I think we should make an API call to an LLM with the current GPS location to decide what to do, bonus points if we can mount a forward facing camera and upload the picture as well
Thinking about this one (https://theaviationist.com/2025/02/12/ea-18g-growler-crashes...), it seems like after ejection you'd want the plane to lawn dart whenever possible. It allows the pilot to know if it is an okay place to ditch and it minimizes the reverse engineering risk.
I imagine there is a good reason this isn't the way things are though.
I think the most proper thing for the jet should be to destroy itself. In a war enviornment I would not like my enemy to gain intel about my military jets.
I think its high-speed collision with the ground or ocean generally takes care of destroying it. Especially with no pilot attempting to keep it level and slow it down and minimize damage.
Not my wheelhouse, but doesn't that involve packing the aircraft with explosives, and wouldn't that involve risk of blowing up if someone else shoots you? Or is there some better way to self-destruct?
Assuming there is an autopilot present, nose diving at max afterburner straight down is probably going to yield a similar result
One could calculate the amount of energy in a given amount of explosive and smashing into the ground at high velocity. I'm too lazy to do so, but I can tell you they're on the same scale.
A heavy object moving fast has a shocking amount of energy. When such an object impacts the ground, all that energy has to go somewhere.
There are probably a number of clever and failsafe ways to divert jet fuel somewhere that would destroy the plane on command.
Impacting the ground usually does it.
Jets are usually full of jet fuel that can blow up if someone else shoots you, or they run into the ground.
Would you go near a plane that's an electronic signal away from blowing itself up?
Even if mechanical, warplanes get combat damage, and having a system like that could make a difference between survivable and sure death.
Maybe bring some strong jamming equipment to prevent the electronic signal from being received?
The signal can originate on the plane itself due to a software glitch, stray currents in the grounding skin or whatever. If anything, jammers tend to interfere with the carrier's electronics itself.
TLDR: It does nothing and it should never do anything