Pilotless Passenger Flight Tested

Seen something interesting in the news or on the intertubes? Discuss it here.

Moderators: Zamfir, Hawknc, Moderators General, Prelates

User avatar
Ormurinn
Posts: 1033
Joined: Tue Apr 03, 2012 3:42 pm UTC
Location: Suth Eoferwicscire

Pilotless Passenger Flight Tested

Postby Ormurinn » Mon May 13, 2013 8:43 pm UTC

http://www.bbc.co.uk/news/technology-22511395

Ah, the elimination of potential human error continues apace.

Hope the security in any future aircraft's control systems is absolutely airtight though.
"Progress" - Technological advances masking societal decay.

User avatar
Izawwlgood
WINNING
Posts: 18686
Joined: Mon Nov 19, 2007 3:55 pm UTC
Location: There may be lovelier lovelies...

Re: Pilotless Passenger Flight Tested

Postby Izawwlgood » Mon May 13, 2013 8:46 pm UTC

Assuming ground based pilots can commandeer such a craft and that it responds nearly as well as having a pilot in the craft itself, I think this is fucking awesome and will probably be immensely cost saving to airlines. But yeah, hacking a passenger plane would be bad news indeed.
... with gigantic melancholies and gigantic mirth, to tread the jeweled thrones of the Earth under his sandalled feet.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Pilotless Passenger Flight Tested

Postby DSenette » Mon May 13, 2013 8:48 pm UTC

most commercial flights are basically flown by computers anyway. they even land/take off by themselves sometimes. I can't see the advantage of not having a pilot in the front of the plane when there are people in the back of it. one would imagine you'd still have to have a one to one pilot ratio for the things in the air so I don't know where you're saving that much money.

completely pilotless planes make sense when there are also no passengers.
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

User avatar
Izawwlgood
WINNING
Posts: 18686
Joined: Mon Nov 19, 2007 3:55 pm UTC
Location: There may be lovelier lovelies...

Re: Pilotless Passenger Flight Tested

Postby Izawwlgood » Mon May 13, 2013 8:57 pm UTC

Uh... Not sure I follow your logic DSenette. If planes are basically flown by computers, what's the advantage of having a pilot in the plane?
... with gigantic melancholies and gigantic mirth, to tread the jeweled thrones of the Earth under his sandalled feet.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Pilotless Passenger Flight Tested

Postby DSenette » Mon May 13, 2013 9:08 pm UTC

Izawwlgood wrote:Uh... Not sure I follow your logic DSenette. If planes are basically flown by computers, what's the advantage of having a pilot in the plane?

for when the computer stops working. if the computer failing is the reason you need the human, then having to rely on a remote connection and undoubtedly more computers isn't a really grand option.

I mean, the reason drones are a thing is because they're completely unmanned...if something fucks up you just lose a plane, you don't lose 400 passengers too.

I think the only way this would save money is if you've got a room where one guy is responsible for reality checking multiple planes at once. which, i'm sure would eventually end up with that one guy being responsible for more and more planes as time goes on. and i'm imagining the airlines would try to get by with as few of these guys as possible, and then that would result in them working longer hours, etc... etc... etc...
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

User avatar
Izawwlgood
WINNING
Posts: 18686
Joined: Mon Nov 19, 2007 3:55 pm UTC
Location: There may be lovelier lovelies...

Re: Pilotless Passenger Flight Tested

Postby Izawwlgood » Mon May 13, 2013 9:13 pm UTC

Well sure, so, if the point of failure that requires a pilot is 'computer', and you're replacing a pilot with 'computer', then sure, potentially problem. If the real reason a pilot is around though is to babysit the thing and/or deal with abnormal flights, then this is a pretty good way of doing things, and we might not need pilots in planes anymore than we need train conductors on subways or toll collectors on high ways.

Obviously the airline industry is going to fuck something up, so, yeah, not overworking the remote pilot the same way, say, air traffic controllers are overworked seems like a good idea.
... with gigantic melancholies and gigantic mirth, to tread the jeweled thrones of the Earth under his sandalled feet.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Pilotless Passenger Flight Tested

Postby DSenette » Mon May 13, 2013 9:19 pm UTC

Izawwlgood wrote:Well sure, so, if the point of failure that requires a pilot is 'computer', and you're replacing a pilot with 'computer', then sure, potentially problem. If the real reason a pilot is around though is to babysit the thing and/or deal with abnormal flights, then this is a pretty good way of doing things, and we might not need pilots in planes anymore than we need train conductors on subways or toll collectors on high ways.

Obviously the airline industry is going to fuck something up, so, yeah, not overworking the remote pilot the same way, say, air traffic controllers are overworked seems like a good idea.

trains, subways, and planes are completely different (and....toll collectors? surely you jest)

trains and subways (effectively the same thing) run on rails that are very hard deviate from. a plane is traveling in 3 dimensional space....filled with people.

it's much more reliable to deal with abnormal flights when you're in direct control of the thing. the remote control is introducing another huge point of failure without actually making any real advances in the process.
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

johnny_7713
Posts: 555
Joined: Tue Jun 15, 2010 1:31 pm UTC

Re: Pilotless Passenger Flight Tested

Postby johnny_7713 » Mon May 13, 2013 9:23 pm UTC

DSenette wrote:most commercial flights are basically flown by computers anyway. they even land/take off by themselves sometimes. I can't see the advantage of not having a pilot in the front of the plane when there are people in the back of it. one would imagine you'd still have to have a one to one pilot ratio for the things in the air so I don't know where you're saving that much money.

completely pilotless planes make sense when there are also no passengers.


The only planes that take-off by themselves (for a limited definition of 'take-off') are carrier based fighters that are launched by catapults (there is an automated system on the plane that controls the trajectory even briefly after the catapult launch). Planes can perform automatic landings, but those require expensive RF equipment to be installed at the specific runway you want to land on (so two antenna sets per runway if you want to be able to do it from both ends), as well as on the aircraft. In addition the pilots have to be specially trained and the aircraft has to be maintained to a higher standard. Special procedures have to be followed in the air (to intercept the guidance signal) as well as on the ground (to avoid interference with the radio signal, e.g. by a plane passing in front of the antenna). All of this conspires to make fully automatic landings rare, basically only being done in heavy fog, or once a month or so if necessary to keep the crew in practise.

Depending on airline policy, pilot preference, and considerations like how crowded the airspace is, the autopilot is generally activated somewhere between 400' and 18,000' and deactivated somewhere between top of descent (when the plane leaves it's cruise level) and 1000' above the ground. Even then the autopilot needs pilot input to know what to do. It's far more accurate to say that the pilot flies the plane via the autopilot, rather than saying the autopilot flies the plane.

This test is not so much about having pilot-less passenger flight and more about having drones that can operate within normal civilian airspace, and mixing with normal traffic, unlike current drones which are restricted to military airspace and/or war zones.

Ormurinn wrote:http://www.bbc.co.uk/news/technology-22511395

Ah, the elimination of potential human error continues apace.

Hope the security in any future aircraft's control systems is absolutely airtight though.


I'd suggest the alternative phrasing: 'the concentration of the potential for human error to the control algorithm design phase continues apace' ;)

Izawwlgood wrote:Assuming ground based pilots can commandeer such a craft and that it responds nearly as well as having a pilot in the craft itself, I think this is fucking awesome and will probably be immensely cost saving to airlines. But yeah, hacking a passenger plane would be bad news indeed.


There are some cost-savings, you save fuel of course and you don't need to book your pilot an hotel room, but I don't think that qualifies as 'immense'. If you could do away with pilots altogether that would save quite some money, but if you need to hire them anyway the cost of having them actually on the plane is not that much extra.

User avatar
Izawwlgood
WINNING
Posts: 18686
Joined: Mon Nov 19, 2007 3:55 pm UTC
Location: There may be lovelier lovelies...

Re: Pilotless Passenger Flight Tested

Postby Izawwlgood » Mon May 13, 2013 11:09 pm UTC

DSenette wrote:trains, subways, and planes are completely different (and....toll collectors? surely you jest)

trains and subways (effectively the same thing) run on rails that are very hard deviate from. a plane is traveling in 3 dimensional space....filled with people.

it's much more reliable to deal with abnormal flights when you're in direct control of the thing. the remote control is introducing another huge point of failure without actually making any real advances in the process.
Derple? No; there are toll collectors in my state, whose jobs could very well be replaced by... well... a machine that collects tolls.

Trains and subways certainly don't require as skills a pilot as a plane does, but the human intervention is still in place to prevent injury to passengers. Automating a subway is probably strikingly easier than automating a plane, but that's not really the point I was getting at.

And yeah, I was agreeing with that sentiment, which is why I agreed with that sentiment. The place where it's potentially good news is you don't need 2 pilots per plane, you can instead, potentially, assuming decent fail safes, have 1 pilot flying multiple planes, only checking in for take off/landings. We have autopilots now, and when they were first introduced, surely this same sort of worry was levied. "What if the autopilot fails?!" was certainly something people were concerned with initially.
... with gigantic melancholies and gigantic mirth, to tread the jeweled thrones of the Earth under his sandalled feet.

User avatar
Iulus Cofield
WINNING
Posts: 2917
Joined: Wed Apr 07, 2010 9:31 am UTC

Re: Pilotless Passenger Flight Tested

Postby Iulus Cofield » Mon May 13, 2013 11:48 pm UTC

If you're going to have a human pilot available to take over for a fully automated plane in the event of a hardware problem, then you have to have a human pilot physically on board each plane. If there's a power failure or a radio failure or a bunch of other potential problems, then a remote human pilot can't do anything to prevent a crash. Even in the case of total power and engine failure, a human pilot can still make an emergency landing, but a computer can't. You could install a separate emergency power system for the computer pilot, but that can fail too. Of course, a human pilot can still die unexpectedly, too, but flight is extremely regulated to prevent nearly every conceivable disaster, so I find it very unlikely that we'll ever see a commerical passenger plane with less than two human pilots. Cargo flights, maybe, drones definitely, but never passenger flights.

User avatar
sardia
Posts: 6797
Joined: Sat Apr 03, 2010 3:39 am UTC

Re: Pilotless Passenger Flight Tested

Postby sardia » Tue May 14, 2013 12:16 am UTC

Iulus Cofield wrote:If you're going to have a human pilot available to take over for a fully automated plane in the event of a hardware problem, then you have to have a human pilot physically on board each plane. If there's a power failure or a radio failure or a bunch of other potential problems, then a remote human pilot can't do anything to prevent a crash. Even in the case of total power and engine failure, a human pilot can still make an emergency landing, but a computer can't. You could install a separate emergency power system for the computer pilot, but that can fail too. Of course, a human pilot can still die unexpectedly, too, but flight is extremely regulated to prevent nearly every conceivable disaster, so I find it very unlikely that we'll ever see a commerical passenger plane with less than two human pilots. Cargo flights, maybe, drones definitely, but never passenger flights.

Ah, so we start flying cargo planes with robots, and then shuttle all those pilots over to the commercial flights.

User avatar
Adacore
Posts: 2755
Joined: Fri Feb 20, 2009 12:35 pm UTC
Location: 한국 창원

Re: Pilotless Passenger Flight Tested

Postby Adacore » Tue May 14, 2013 12:20 am UTC

As I understand it, most modern airliners are almost entirely fly-by-wire. If the flight computer dies completely, then a human pilot in the plane will be able to do precisely as much nothing as a computer pilot in the plane could.

User avatar
Ormurinn
Posts: 1033
Joined: Tue Apr 03, 2012 3:42 pm UTC
Location: Suth Eoferwicscire

Re: Pilotless Passenger Flight Tested

Postby Ormurinn » Tue May 14, 2013 12:25 am UTC

Izawwlgood wrote:Trains and subways certainly don't require as skills a pilot as a plane does, but the human intervention is still in place to prevent injury to passengers. Automating a subway is probably strikingly easier than automating a plane, but that's not really the point I was getting at.


There are calls to automate the tube roughly once a year in the U.K, whenever the RMT starts shit stirring.

I was under the impression that many other countries had a lot more automations than we do. I guess they don't have to deal with the RMT...
"Progress" - Technological advances masking societal decay.

User avatar
Angua
Don't call her Delphine.
Posts: 5932
Joined: Tue Sep 16, 2008 12:42 pm UTC
Location: UK/[St. Kitts and] Nevis Occasionally, I migrate to the US for a bit

Re: Pilotless Passenger Flight Tested

Postby Angua » Tue May 14, 2013 6:34 am UTC

The subway isn't automated though - the fee system is, but the drivers are still there. They have to do emergency stuff, fix the train if there's a problem, etc.

Automating things on planes will be good, but given that technical malfunction is still a common cause of air-flight problems (20% versus 80% humans), you need to have someone on board to be able to override the system.
Crabtree's bludgeon: “no set of mutually inconsistent observations can exist for which some human intellect cannot conceive a coherent explanation, however complicated”
GNU Terry Pratchett

User avatar
Adacore
Posts: 2755
Joined: Fri Feb 20, 2009 12:35 pm UTC
Location: 한국 창원

Re: Pilotless Passenger Flight Tested

Postby Adacore » Tue May 14, 2013 6:42 am UTC

Angua wrote:The subway isn't automated though - the fee system is, but the drivers are still there. They have to do emergency stuff, fix the train if there's a problem, etc.

This depends on the subway. See Wikipedia's handy list of driverless trains.

johnny_7713
Posts: 555
Joined: Tue Jun 15, 2010 1:31 pm UTC

Re: Pilotless Passenger Flight Tested

Postby johnny_7713 » Tue May 14, 2013 7:38 am UTC

Adacore wrote:As I understand it, most modern airliners are almost entirely fly-by-wire. If the flight computer dies completely, then a human pilot in the plane will be able to do precisely as much nothing as a computer pilot in the plane could.


Generally there are still mechanical links to at least some of the controls (e.g. the rudder and the elevator trim on an A320 IIRC). In principle those controls are mainly suitable for keeping the plane stable while you try to reboot the flight computer, but a human can at least do a bit more nothing than a computer pilot. (The flight computers are doubly redundant by the way, i.e. there are three of them).

HungryHobo
Posts: 1708
Joined: Wed Oct 20, 2010 9:01 am UTC

Re: Pilotless Passenger Flight Tested

Postby HungryHobo » Tue May 14, 2013 8:41 am UTC

I always see these "what if the computer fails" complaints whenever something like this comes up.

yes a computer can fail but as long as the control system is well made it shouldn't fail often.

to which of course the first reply is "but it can still fail!!!"

well yes. of course it can, just like humans.

A component in one of the computers can burn out or there can be a software bug or there can be quite a few other sources of failure.

on the other hand a blood vessel in the pilots brain can burst, the pilot can have an allergic reaction to his in flight meal. The pilot can get depressed and decide to end it all. the pilot can just make a really bad judgement call or ignore proper procedure because it turned out fine the last 10 times he did it and saved time or the pilot might have had a terrible nights sleep the night before. Something he wasn't trained for can come up or he can just respond the wrong way in a tight spot.

The moment that an autopilots can make less errors than a human pilot, even if they're not perfect it's rational to switch to using them. If on average you'll have 100 crashes with all human pilots but 90 with all auto pilots it's time to use the autopilots.
Last edited by HungryHobo on Tue May 14, 2013 8:45 am UTC, edited 1 time in total.
Give a man a fish, he owes you one fish. Teach a man to fish, you give up your monopoly on fisheries.

User avatar
Iulus Cofield
WINNING
Posts: 2917
Joined: Wed Apr 07, 2010 9:31 am UTC

Re: Pilotless Passenger Flight Tested

Postby Iulus Cofield » Tue May 14, 2013 8:44 am UTC

The issue really isn't whether humans or machines fail more or less often, it's about the attempt to prevent failures altogether through redundancy. Co-pilots aren't just there so the pilot can sneak away to the bathroom, after all.

User avatar
Ormurinn
Posts: 1033
Joined: Tue Apr 03, 2012 3:42 pm UTC
Location: Suth Eoferwicscire

Re: Pilotless Passenger Flight Tested

Postby Ormurinn » Tue May 14, 2013 9:05 am UTC

Iulus Cofield wrote:The issue really isn't whether humans or machines fail more or less often, it's about the attempt to prevent failures altogether through redundancy. Co-pilots aren't just there so the pilot can sneak away to the bathroom, after all.


The question becomes, then, I suppose, how much extra risk will the industry tolerate?

I'd actually foresee the limiting factor being the industry itself - pilotless planes would be bad publicity at first.
"Progress" - Technological advances masking societal decay.

User avatar
Thesh
Made to Fuck Dinosaurs
Posts: 6577
Joined: Tue Jan 12, 2010 1:55 am UTC
Location: Colorado

Re: Pilotless Passenger Flight Tested

Postby Thesh » Tue May 14, 2013 9:15 am UTC

Once driverless cars become common enough, people will open up to pilotless planes.
Summum ius, summa iniuria.

User avatar
Ormurinn
Posts: 1033
Joined: Tue Apr 03, 2012 3:42 pm UTC
Location: Suth Eoferwicscire

Re: Pilotless Passenger Flight Tested

Postby Ormurinn » Tue May 14, 2013 9:31 am UTC

Thesh wrote:Once driverless cars become common enough, people will open up to pilotless planes.


Definately. The government, and industry regulators, on the other hand, will probably be less forward thinking.

I predict we'll start seeing them about 50 years after the technology becomes affordable.
"Progress" - Technological advances masking societal decay.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7590
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Pilotless Passenger Flight Tested

Postby Zamfir » Tue May 14, 2013 9:47 am UTC

As ready noted above: the article in the OP is not about pilotless passenger aircraft. It's a test about flying unmanned drones in the same airspace as civilian aircraft. The test vehicle happened to be a small passenger plane, but that's just because that was the available test platform. For the test they needed an aircraft that could fly both piloted and by remote operation.

Pilotless airliners might happen one day, but it's still distant and this test didn't really add much information about that. The issue there is guaranteeing safety, not one-off feasibility. The drones for civilian airspace on he other hand are right here waiting, and air traffic control has to be aware how those might behave different than they are used to. That's what the test was primarily about.

Chen
Posts: 5576
Joined: Fri Jul 25, 2008 6:53 pm UTC
Location: Montreal

Re: Pilotless Passenger Flight Tested

Postby Chen » Tue May 14, 2013 12:13 pm UTC

In working with electronic controls systems and watching that Mayday show on Discovery, I would be FAR more comfortable with planes that could be flown without humans at the controls. I'm pretty sure Catastrophic hazards (meaning the result is multiple fatalities and/or loss of the aircraft) in aircraft design (be it software or mechanical) needs to meet a probability of something like less than 10^-9 per flight hour. Even the plane crashes that are classified as mechanical failure often come down errors or violations in maintenance procedures which is again human error.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Pilotless Passenger Flight Tested

Postby DSenette » Tue May 14, 2013 12:38 pm UTC

Izawwlgood wrote:
DSenette wrote:trains, subways, and planes are completely different (and....toll collectors? surely you jest)

trains and subways (effectively the same thing) run on rails that are very hard deviate from. a plane is traveling in 3 dimensional space....filled with people.

it's much more reliable to deal with abnormal flights when you're in direct control of the thing. the remote control is introducing another huge point of failure without actually making any real advances in the process.
Derple? No; there are toll collectors in my state, whose jobs could very well be replaced by... well... a machine that collects tolls.

Trains and subways certainly don't require as skills a pilot as a plane does, but the human intervention is still in place to prevent injury to passengers. Automating a subway is probably strikingly easier than automating a plane, but that's not really the point I was getting at.
replacing a toll booth with an automated system is the same thing as putting an ATM out front of your bank....there's no change in service and there's no increase in the danger of catastrophic failure of the system. there's no comparison between automated tolling (which can be done with something as simple as a barcode reader) and a giant tin can full of people rocketing through the air at 500+ MPH. because it's so common place....people forget how freaking difficult flight actually is.

as to the trains (subway and cargo)....they're straight line transports. subways are even easier than trains because they are a multi-track/single direction system. there are never two trains on the same track going in the opposite direction unless there has been a catastrophic failure in planning. automating the subway system is about scheduling and controlling breaking/acceleration. things that are easily automated. it's all about timing and scheduling. there is basically no expectation of external variables changing (the track suddenly going to the left instead of the right. there being a car you don't know about on the track. cows. geese. etc...). cargo/passenger trains (heavy rail) are just a LITTLE more complicated because there are external variables (cars on tracks, cows, etc...) but those are pretty limited too.

planes have 3d space to worry about, air that suddenly isn't there anymore, air that's denser than you thought it was, wind out of nowhere, geese.....too many variables.

Izawwlgood wrote:And yeah, I was agreeing with that sentiment, which is why I agreed with that sentiment. The place where it's potentially good news is you don't need 2 pilots per plane, you can instead, potentially, assuming decent fail safes, have 1 pilot flying multiple planes, only checking in for take off/landings. We have autopilots now, and when they were first introduced, surely this same sort of worry was levied. "What if the autopilot fails?!" was certainly something people were concerned with initially.

until 3 planes have an issue at the same time. the "what if the autopilot fails" issue was handled quite instantly...if autopilot fails, there's a guy in the seat.

Zamfir wrote:As ready noted above: the article in the OP is not about pilotless passenger aircraft. It's a test about flying unmanned drones in the same airspace as civilian aircraft. The test vehicle happened to be a small passenger plane, but that's just because that was the available test platform. For the test they needed an aircraft that could fly both piloted and by remote operation.

Pilotless airliners might happen one day, but it's still distant and this test didn't really add much information about that. The issue there is guaranteeing safety, not one-off feasibility. The drones for civilian airspace on he other hand are right here waiting, and air traffic control has to be aware how those might behave different than they are used to. That's what the test was primarily about.
(and johnny_7713) I didn't catch the cargo plane/this is just a drone test bit. that makes sense. making sure that the robots NOT carrying people around don't hit the ones that are carrying people around. or end up in buildings, etc.

Adacore wrote:As I understand it, most modern airliners are almost entirely fly-by-wire. If the flight computer dies completely, then a human pilot in the plane will be able to do precisely as much nothing as a computer pilot in the plane could.
the fly-by-wire systems aren't directly fly-by-computer......there are some direct pathways to critical controls just in case all the computers on the plane take a shit (based on my understanding of the failsafes in place on modern aircraft...which totes might be wrong)
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

User avatar
jules.LT
Posts: 1539
Joined: Sun Jul 19, 2009 8:20 pm UTC
Location: Paris, France, Europe

Re: Pilotless Passenger Flight Tested

Postby jules.LT » Tue May 14, 2013 12:52 pm UTC

Ormurinn wrote:I was under the impression that many other countries had a lot more automations than we do. I guess they don't have to deal with the RMT...
I think all railway unions pose the same problems, because of how easily they can mess so many people's day up.
That's part of the reason we develop driverless trains.

Aircraft pilots are less strike-prone, and the cost savings won't make up for the huge additional point of failure in a long time.
Bertrand Russell wrote:Not to be absolutely certain is, I think, one of the essential things in rationality.
Richard Feynman & many others wrote:Keep an open mind – but not so open that your brain falls out

User avatar
ucim
Posts: 6858
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Pilotless Passenger Flight Tested

Postby ucim » Tue May 14, 2013 2:34 pm UTC

The new single point of failure is the communications link. Of course we have a lot of experience with that - The internet never goes down, nobody has ever intercepted encrypted communications, and bad people won't try to overload take over the airplane by remote control.

Nah, never happen.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Heartfelt thanks from addams and from me - you really made a difference.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7590
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Pilotless Passenger Flight Tested

Postby Zamfir » Tue May 14, 2013 3:04 pm UTC

Keep in mind that an automated plane wouldn't be unmanned. It could start by going to a single pilot, who works together with a ground based operator. But even without a trained pilot aboard, you would still have a cabin crew. Head of cabin crew would become a more responsible function in the absense if a pilot-captain.

Such a person could still do high-level piloting, if the plane itself is capable of handling the mechanics. Like 'fly to the nearest dedicated safe-zone and start a holding pattern', or even 'land at runway 4 of this airbase' . Also, 'someyhing is wrong, ignore incoming radio commands until further notice', or 'start the emergency broadcast that warns traffic control of problems'

Such actions would provide a buffer against comm problems, buying time to switch to alternative options. It would not be very different from comms problems with a piloted plane.

Of course, current technology is not yet up to this, but it is moving there. And we won't see pilotless airliners until such high-level waypoint piloting is the mature reliable norm for other aircraft, like freighters or combat craft. It would be the end of the process, not a step for the near future.

User avatar
ucim
Posts: 6858
Joined: Fri Sep 28, 2012 3:23 pm UTC
Location: The One True Thread

Re: Pilotless Passenger Flight Tested

Postby ucim » Tue May 14, 2013 7:03 pm UTC

Zamfir wrote:Keep in mind that an automated plane wouldn't be unmanned. It could start by going to a single pilot, who works together with a ground based operator. But even without a trained pilot aboard, you would still have a cabin crew. Head of cabin crew would become a more responsible function in the absense if a pilot-captain...
A monkey can fly an airplane, so long as nothing unexpected happens. The reason you have a real pilot is so that, when the unexpected happens, the airplane can be reused and the passengers can complain in person rather than through their estate. You are proposing that, at that critical time when a real pilot is needed, the head of the cabin crew will be able to safely extricate a modern jetliner from a spiral dive in the middle of a thunderstorm while the controls are being jammed by a terrorist with an iPod and a big antenna?

Zamfir wrote:It would not be very different from comms problems with a piloted plane.
It would be night and day different from comms problem with a piloted plane.

I am hesitant to predict the far future, but this is like predicting automated clothespin applicators for putting clothes out on the clothesline.

Jose
Order of the Sillies, Honoris Causam - bestowed by charlie_grumbles on NP 859 * OTTscar winner: Wordsmith - bestowed by yappobiscuts and the OTT on NP 1832 * Ecclesiastical Calendar of the Order of the Holy Contradiction * Heartfelt thanks from addams and from me - you really made a difference.

User avatar
sardia
Posts: 6797
Joined: Sat Apr 03, 2010 3:39 am UTC

Re: Pilotless Passenger Flight Tested

Postby sardia » Tue May 14, 2013 7:49 pm UTC

ucim wrote:
Zamfir wrote:Keep in mind that an automated plane wouldn't be unmanned. It could start by going to a single pilot, who works together with a ground based operator. But even without a trained pilot aboard, you would still have a cabin crew. Head of cabin crew would become a more responsible function in the absense if a pilot-captain...
A monkey can fly an airplane, so long as nothing unexpected happens. The reason you have a real pilot is so that, when the unexpected happens, the airplane can be reused and the passengers can complain in person rather than through their estate. You are proposing that, at that critical time when a real pilot is needed, the head of the cabin crew will be able to safely extricate a modern jetliner from a spiral dive in the middle of a thunderstorm while the controls are being jammed by a terrorist with an iPod and a big antenna?

Zamfir wrote:It would not be very different from comms problems with a piloted plane.
It would be night and day different from comms problem with a piloted plane.

I am hesitant to predict the far future, but this is like predicting automated clothespin applicators for putting clothes out on the clothesline.

Jose

Let's take a example that is coming up now. Driverless cars. They aren't any better than an optimal driver, but they don't get tired and allow closer spacing between cars. Now if they are vulnerable to programming bugs or jamming, and they crash, does that mean we can't use driverless cars? No, because the chance of a car crashing due to an unexpected thing is small and worth the efficiency gains and fewer accidents that driverless cars give us.

The same can be applied to planes. But, you're falling into the fallacy of fearing a large event that is rare. Everyone is afraid of dying in a crashing plane. Nobody is afraid of dying in a crashing car. And yet the car is infinitely more dangerous than the plane.

DSenette
Posts: 2418
Joined: Fri Mar 06, 2009 8:08 pm UTC

Re: Pilotless Passenger Flight Tested

Postby DSenette » Tue May 14, 2013 8:36 pm UTC

sardia wrote:
ucim wrote:
Zamfir wrote:Keep in mind that an automated plane wouldn't be unmanned. It could start by going to a single pilot, who works together with a ground based operator. But even without a trained pilot aboard, you would still have a cabin crew. Head of cabin crew would become a more responsible function in the absense if a pilot-captain...
A monkey can fly an airplane, so long as nothing unexpected happens. The reason you have a real pilot is so that, when the unexpected happens, the airplane can be reused and the passengers can complain in person rather than through their estate. You are proposing that, at that critical time when a real pilot is needed, the head of the cabin crew will be able to safely extricate a modern jetliner from a spiral dive in the middle of a thunderstorm while the controls are being jammed by a terrorist with an iPod and a big antenna?

Zamfir wrote:It would not be very different from comms problems with a piloted plane.
It would be night and day different from comms problem with a piloted plane.

I am hesitant to predict the far future, but this is like predicting automated clothespin applicators for putting clothes out on the clothesline.

Jose

Let's take a example that is coming up now. Driverless cars. They aren't any better than an optimal driver, but they don't get tired and allow closer spacing between cars. Now if they are vulnerable to programming bugs or jamming, and they crash, does that mean we can't use driverless cars? No, because the chance of a car crashing due to an unexpected thing is small and worth the efficiency gains and fewer accidents that driverless cars give us.

The same can be applied to planes. But, you're falling into the fallacy of fearing a large event that is rare. Everyone is afraid of dying in a crashing plane. Nobody is afraid of dying in a crashing car. And yet the car is infinitely more dangerous than the plane.

the differences between a self driving car and a self flying plane are massive. a car is still only moving around in a 2d space...you're doing speed control, route finding, and object avoidance. it's academically extremely easy (functionally, it's hard as fuck...but on paper it's easy). the efficiency of packing multiple cars on the roads only comes when all of them are robotic and communicating between each other and (ideally) the road.

you don't have to worry about freak wind or things like that. if the car breaks down, it either coasts over to the side of the road, or everyone else just routes around the broken down car in the middle of the lane.
The Righteous Hand Of Retribution
"The evaporation of 4 million who believe this crap would leave the world an instantly better place." ~Andre Codresu (re: "the Rapture")

User avatar
Ormurinn
Posts: 1033
Joined: Tue Apr 03, 2012 3:42 pm UTC
Location: Suth Eoferwicscire

Re: Pilotless Passenger Flight Tested

Postby Ormurinn » Tue May 14, 2013 9:27 pm UTC

DSenette wrote:
sardia wrote:
ucim wrote:
Zamfir wrote:Keep in mind that an automated plane wouldn't be unmanned. It could start by going to a single pilot, who works together with a ground based operator. But even without a trained pilot aboard, you would still have a cabin crew. Head of cabin crew would become a more responsible function in the absense if a pilot-captain...
A monkey can fly an airplane, so long as nothing unexpected happens. The reason you have a real pilot is so that, when the unexpected happens, the airplane can be reused and the passengers can complain in person rather than through their estate. You are proposing that, at that critical time when a real pilot is needed, the head of the cabin crew will be able to safely extricate a modern jetliner from a spiral dive in the middle of a thunderstorm while the controls are being jammed by a terrorist with an iPod and a big antenna?

Zamfir wrote:It would not be very different from comms problems with a piloted plane.
It would be night and day different from comms problem with a piloted plane.

I am hesitant to predict the far future, but this is like predicting automated clothespin applicators for putting clothes out on the clothesline.

Jose

Let's take a example that is coming up now. Driverless cars. They aren't any better than an optimal driver, but they don't get tired and allow closer spacing between cars. Now if they are vulnerable to programming bugs or jamming, and they crash, does that mean we can't use driverless cars? No, because the chance of a car crashing due to an unexpected thing is small and worth the efficiency gains and fewer accidents that driverless cars give us.

The same can be applied to planes. But, you're falling into the fallacy of fearing a large event that is rare. Everyone is afraid of dying in a crashing plane. Nobody is afraid of dying in a crashing car. And yet the car is infinitely more dangerous than the plane.

the differences between a self driving car and a self flying plane are massive. a car is still only moving around in a 2d space...you're doing speed control, route finding, and object avoidance. it's academically extremely easy (functionally, it's hard as fuck...but on paper it's easy). the efficiency of packing multiple cars on the roads only comes when all of them are robotic and communicating between each other and (ideally) the road.

you don't have to worry about freak wind or things like that. if the car breaks down, it either coasts over to the side of the road, or everyone else just routes around the broken down car in the middle of the lane.



For a car you have to worry about road condition, the weather, and the gradient it's driving along.

Theres one massive factor - there are a lot more unpredictable actors in the environment of a car than for a plane. The atmosphere is mostly devoid of physical objects, while the ground is a smorgasbord of things that need to not be hit.

If your aircraft makes a turn or begins to reduce it's speed 0.5 seconds late, it can course correct later. The same thing in a self-driving car is an automatic traffic accident.
"Progress" - Technological advances masking societal decay.

Derek
Posts: 2181
Joined: Wed Aug 18, 2010 4:15 am UTC

Re: Pilotless Passenger Flight Tested

Postby Derek » Tue May 14, 2013 9:47 pm UTC

Yeah, navigating a 3D environment with basically no obstacles is a thousand times easier than navigating a path full of obstacles with small gaps between you and the obstacles compared to your speed. Shit, the military spends millions of dollars developing technologies to intentionally hit aircraft in 3D space. Avoiding them is the easy part.

User avatar
Adacore
Posts: 2755
Joined: Fri Feb 20, 2009 12:35 pm UTC
Location: 한국 창원

Re: Pilotless Passenger Flight Tested

Postby Adacore » Wed May 15, 2013 12:10 am UTC

Humans aren't particularly good at operating with three-dimensional movement, either. We're evolved to live on the two-dimensional surface of the planet, not in the sky. And while there are probably times when a human pilot would have a better response than a computer to some kind of emergency situation, there are probably also plenty of times when the computer makes the correct decision, in addition to the many times a computer will simply be able to take more rapid, precise action than a human simply due to the practicalities of how they operate.

I was about to start looking at aircraft accident data to determine if a computer pilot would've done better than a human one, but of course that would have a huge selection bias, as all the accidents to date have been on planes piloted by humans. Any accident types that would've occured with a computer pilot but not with a human at the helm will, of course, not be present, because to date there's always been a human pilot.

User avatar
Thesh
Made to Fuck Dinosaurs
Posts: 6577
Joined: Tue Jan 12, 2010 1:55 am UTC
Location: Colorado

Re: Pilotless Passenger Flight Tested

Postby Thesh » Wed May 15, 2013 12:30 am UTC

So what you are saying is that, to date, not a single computer piloted commercial aircraft has had a crash, whereas there have been hundreds, if not thousands, of human piloted aircraft crashes?

I'm sold.
Summum ius, summa iniuria.

User avatar
Woopate
Scrapple
Posts: 503
Joined: Fri Mar 12, 2010 10:34 am UTC

Re: Pilotless Passenger Flight Tested

Postby Woopate » Wed May 15, 2013 4:52 am UTC

(testing drone flight in human airspace is a really good thing, Discussion appears to be centered around automated flight on a commercial level so I'm gonna talk about that.)

There's an old pilot joke: one day all aircraft will be flown by a dog and a human. The human will feed the dog, and the dog will bite the human if xhe touches anything. 

That said, I do not think we're anywhere near that yet. 

Many people have been equating automated cars with automated airplanes. These are not remotely equatable. In automated cars, you are constantly dodging and navigating obstacles. But, at the end of the day, you can hit the breaks and stop the car and be reasonably assured of safety in the event of some kind of systems failure. 

In an airplane, it's almost the exact opposite. There are few obstacles, and you can't stop, and the question is not really about dealing with the routine(a 12 year old can fly pretty well under normal circumstances with guidance{except takeoffs and landings}), the question is how well can you deal with a systems failure? What does a computer do if it's fed conflicting information? If the computer loses it's sensors, can it land? Can you account for every problem a human can adapt to by using redundancies and programming? 

Obviously humans are inconsistant in this regard, sometimes managing to hit another airplane in the whole blue sky while their computers are saying "Dude your gonna hit a plane." or fly straight into the dirt, but sometimes they can pull a Gimli Glider or a Hudson River Landing. 

Can computers handle all the unexpected things flight throws your way better than a human can consistantly? This is the first question. The second question is is the difference enough to justify the expense? 

I find it improbable that these computers, the added sensors, the specialists to maintain them, the added infrastructure to transmit to the plane and the specialists for THAT, plus whatever political roadblocks are thrown up around such an "unnatural" thing will be cheaper than just sticking with pilots, especially when it's probable that pilots will still be required decades after implementation. Airlines fight a very delicate balance to stay profitable. As ugly as it is, there's always someone who has to ask if the added cost is worth the added safety.

User avatar
sardia
Posts: 6797
Joined: Sat Apr 03, 2010 3:39 am UTC

Re: Pilotless Passenger Flight Tested

Postby sardia » Wed May 15, 2013 5:48 am UTC

Thesh wrote:So what you are saying is that, to date, not a single computer piloted commercial aircraft has had a crash, whereas there have been hundreds, if not thousands, of human piloted aircraft crashes?

I'm sold.

The actual data is that the vast majority of crashes are caused by human error. Remove the human, and you're left with mechanical error and the addition of programming errors. If you can reduce the programming error to far below the human error rate, then you should rationally use automated planes/cars whatever.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7590
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Pilotless Passenger Flight Tested

Postby Zamfir » Wed May 15, 2013 6:37 am UTC

Woopate, that sounds right. It is far from obvious that drone technology will improve to be as safe as commercial airline pilots. Right now it's quickly improving in dealing with 'routine' situations, but at the end of the road there might enough special situation left that humans handle better than the machine. Or more exactly, where an in-plane pilot aided by flight computers does better than an advanced autopilot with remote human commands. It's not a man vs machine comparison.

But people will build the technology and infrastructure anyway, for use in unmanned flights. There are going to be FedEx drones, no matter what happens to passenger flights. So there will be comparison material both with regards to the problem rate, and the cost. And part of the technology will seep into manned aircraft anyway, if for example autopilot become better than pilots ate revory from spirals, or landings, etc. I expect that in-aircraft piloting will look for and more like drone operation, with more high level commands and less mechanics of flying the plane.

Carnildo
Posts: 2023
Joined: Fri Jul 18, 2008 8:43 am UTC

Re: Pilotless Passenger Flight Tested

Postby Carnildo » Wed May 15, 2013 7:29 am UTC

Derek wrote:Yeah, navigating a 3D environment with basically no obstacles is a thousand times easier than navigating a path full of obstacles with small gaps between you and the obstacles compared to your speed. Shit, the military spends millions of dollars developing technologies to intentionally hit aircraft in 3D space. Avoiding them is the easy part.

Pathfinding is the trivial part of designing the system. The difficulty lies in fault handling.

For a train, fault-handling is trivial: "If fault then stop". Theoretically, there could be a fault (slowly-collapsing bridge?) that would better be handled by speeding up, but those are vanishingly rare.

For a car, it gets harder. In addition to the "if fault then stop" situations, you add a set of situations that should be handled by steering to one side or the other, and possibly ones that require a trickier maneuver such as a deliberate skid. Further, since you've got more options, you need to add logic to decide which one is right for any situation.

For an airplane, it gets even more complicated. In virtually all situations, you need to diagnose the fault (volcanic ash clogging all four engines? fuel exhaustion with no airport in range? rotor disk failure leaving you with two working engines and no working hydraulics?) to a sufficient degree to select the appropriate response. "If fault then stop" is only an option for a very limited range of situations.

User avatar
Woopate
Scrapple
Posts: 503
Joined: Fri Mar 12, 2010 10:34 am UTC

Re: Pilotless Passenger Flight Tested

Postby Woopate » Wed May 15, 2013 7:55 am UTC

Zamfir wrote:It's not a man vs machine comparison.


Of course you are correct. Consider "human" a shorthand for "pilot in cockpit" and "computer" shorthand for "pilot on ground who takes over in emergencies", as that's what I was thinking of when I wrote my post.

A slight tangent on drone cargo aircraft:

Spoiler:
I wonder what the implications are with regards to safety for unmanned cargo aircraft. I hear constantly about bosses in aviation demanding their pilots toe the line with regards to legality in order to optimize profits, such as taking less fuel to squeeze a few extra pounds of cargo on, scheduling pilots to fly with exactly the minimum amount of rest between each flight, encouraging them to fly in dubious weather. Even if these aircraft are unmanned, they are still a risk to people at the airports and anyone who might be beneath them. I don't know what sort of exploits might come about as a result of having no pilot to object. Even if a computer is designed to not operate if the legal specifications are not met, it could still be lied to.

User avatar
Zamfir
I built a novelty castle, the irony was lost on some.
Posts: 7590
Joined: Wed Aug 27, 2008 2:43 pm UTC
Location: Nederland

Re: Pilotless Passenger Flight Tested

Postby Zamfir » Wed May 15, 2013 8:05 am UTC

Carnildo, that would be true if we were considering entirely autonomous systems. But that's not the goal in aircraft automation, the goal is to have system where the aircraft can execute low level and time-critical tasks autonomously, while a remote operator makes higher level decisions.

In that respect, aircraft really have an advantage over cars, which is why unmanned aircraft are more common than unmanned cars. On the ground, unexpected events are common and require direct decisions. In flight, there is typically more grace time. Just look at your examples.

That's not an accident.It's because unexpected events that require direct action tend be highly dangerous in aircraft, even with skilled pilots, so their whole operation method is set up to avoid them.

@woopate, I would expect that that large drones stay (at least initially) restricted to dedicated airfields and lowly populated areas. And it's not as if it will start with unmanned A380Fs, more with small remote sensing drones etc. Then we can see which risks are real, which are easy to solve, which are just not relevant in practice. It doesn't have to be right all at once.


Return to “News & Articles”

Who is online

Users browsing this forum: No registered users and 17 guests