We'll agree to disagree on that. I feel it would be a100% certainty that no one would survive.

A lot of pilots hate them, but I'm ok with red-eyes. You get direct routing from ATC (out of Las Vegas, we often get direct to Wilkes-Barre, PA after switching from Las Vegas Departure control to L.A. Center), there's a lot less annoying chatter on the radio, and you get in early and have the whole day off.

I am usually too jaded to laugh at anything on the internet anymore, but this morning W0X0F provided me with not one but TWO "lol"s.

I'm sure I'm not alone in thinking that this magnum opus of a thread is the absolute best piece of online writing on the web, bar none.

On the other hand (playin devil's advocate), that landing was considered by many a 'miracle' (mix of pilot skills, incredible 'sang froid' and lot of luck.
Wouldn't you say that the chance that would happen again without casualty is very slim ?

On the other hand, pilotless airplanes would (obviously) be safer in regards to the human-related incidents that seem to be the cause of most of the crashes lately ? (Rio / Paris, Germanwings, the last Airasia...)

It's a silly discussion because no pilotless plane exists. No one has even attempted to create an auto-pilot that would be able to deal with such a scenario so we simply can't say whether or not a pilotless plane would or would not be able to ditch successfully.

The question is whether hypothetically it is possible to create software that could fly a plane as well as a human pilot currently can in both normal circumstances and through malfunctions. I would guess that in theory it should be possible but we're just very,very far away from it.

I think you're overlooking the human error while making the 'autopilot' here. There also seems way more issues with broken electronics/malfunctioning electronics. Not to forget the safety issues with someone being able to take over a plane without even being in the cockpit, security measures would have to be insane and probably 100% closed circuit. One of the great things we do as humans is quickly being able to asses situations and I just think that we're going to have a huge issue with building those sensory observations into a plane. We probably will at some point, I just think it will become a proven concept in other much more mainstream areas before we try implanting it in airplanes./knowledgelessrant

We are actually not that far away from it. A self driven car is actually harder than a plane and Google already has one driving around. The reason a car is harder is because the biggest gap between computers and humans is the ability to process visual clues. Something that is much more important in cars.

The reason we are still a very long time away from pilotless planes is that it is almost impossible to convince people that the computer performs better than a human. Not just because people don't trust computers but how do you get the data to prove that the computer is better. You can get a driverless car with one backup driver to drive around LA for years to prove it works but I don't see anyone do the same with a plane.

Humans are actually damn slow in assessing situations compared to computers. Especially in planes where all the information is provided to you by instruments. So all the sensory observations are already build in every plane out there. Where we beat computers is anything to do with vision and coming up with completely new ideas. Even the landing on the Hudson was not a new idea. The amazing thing about it was executing it flawlessly with little time to think and under a lot of stress.

It's not really accurate to say that flying is an easier problem than driving. In some ways it is, but in other ways it definitely isn't.

Take the case of an unexpected hardware malfunction. In the case of a car you can default to stop the car and get off the road. Pretty easy.

In the case of a flying plane you still need to keep flying and eventually land the damn thing. So the error handling software is significantly costlier.

Flying also has a very different political and public relation problem. Things like getting licensed and dealing with the first accident that kills hundreds of people.

If a hardware failure interferes with the keep flying bit of a plane then it crashes with or without a pilot. So for a computer landing the plane after a mechanical failure is just as hard as parking the car in a safe spot at the side of the road. We just look at it differently because the impact of getting of the road in a ditch is a lot lower than dropping out of the sky.

So it is the public relations that make pilotless planes much less likely than driverless cars even though the car is currently harder.

For example the biggest issue right now is that a driverless car can't see the difference between an empty paper bag on the road and a rock. So it will try to avoid both while a human would just drive over the paper bag. Stuff like that doesn't happen in the air and planes can't avoid stuff on the runway so it doesn't need to know the difference.

The bolded is just wrong.

A simple example to make my point more concrete: Driverless car software can deal with a flat tire, overheating engine, loss of hydraulic fluid, etc. by stopping the car and calling for help. A pilotless plane needs to have software that deals with engine failures, loss of instruments, combinations of failures, etc. that don't involve just shutting the plane down.

I find this derail kind of annoying, so I guess I'll leave it at that.

Well if we are talking about technology available today, sure. But flying a plane is essentially a software problem. Current Google cars are really good, but only on roads that have been mapped out with lasers ahead of time. But that's changing. Fast.

This is pretty ridiculous. ATC has to be able to give directions to the plane so it doesn't run into other planes. That means radio com to the plane for directions, which can be easily interfered with deliberately or accidentally. A pilot that loses com can still fly the plane, whereas a computer controlled plane will be clueless.

I think a bigger fear of pilotless planes would be someone hacking the CPU system that's controlling them.

Could you imagine a terrorist group hacking and crashing every plane that's in the sky from a specific airline one day?

If a driverless car just stops the car in the middle of the highway when its engine overheats then it would be just as wrong as a pilotless plane dropping out of the sky if it doesn't know how to deal with any of the issues you mention. Neither one will be allowed in public. Pilots and drivers both follow standard procedures when hardware failures happen and it is already possible for computers to do the same no matter how complex we might think it is.

As for the computer needing com to fly a plane. No idea why that would be the case. It has the same information a pilot has and is perfectly capable to avoid other planes and land without communications just like a pilot would. What procedures do you think a pilot without com takes that a computer can't.

Anyway I will also stop the derail. If people really are interested they should check what the Google car is already capable of and other fault tolerant AI systems.

Might a pilot lose their license for forgetting to deploy landing gear?
http://www.liveleak.com/view?i=b41_1429465294
Some of the comments give more data.

We may be slow but we're good at correctly assessing a ton of different situations when compared to current available technologies for these kind of things. The question will also rise whether its cost effective. Im 100% sure that it will happen in the future but Ill be dead by then. Even 'google' cars have huge issues surrounding them so we really shouldn't get ahead of ourselves.

After...

The internet sleuths are amazing..
http://www.asias.faa.gov/pls/apex/f?...0PIERCE,N7502S

Some Dutch guy managed to do it:



(I hope that clip won't offend the more sensitive readers)

Dont know what happens to license but I bet this guys military career took a hit

Wat? Non parlez Francaise.

Mirage = these aircraft.

Ventre = belly

That's probably all you need to know. It's the same clip twice.

Yeah. Also, as he's sliding down the runway he says he didnt lower the landing gear, so its not a malfunction but his mistake. Someone says in the comments that this was due to fatigue from combat excercises and that this didnt affect his career too much but the controller lost her job for some reason. Impossible to know for sure, I couldnt find any info on the incident. I dont know why this would have affected the controller.

Red eye last night LAX-JFK. The captain announced that due to storms over Chicago, he was going to shift course and, as a result, the flight would (and did) arrive 45 minutes early. This means the regular route is intentionally slow. Query whether that is (i) for fuel efficiency, (ii) due to allocation of landing slots & gates at JFK, (iii) because the flight is more popular if people can get more sleep or (iv) other.