Pilotless planes

fr336

Will it ever be wise to rely solely on computers? Surely that would be madness, but I have no doubt the profit obsessed business world and in turn the media would convince the public go to along with it eventually :rolleyes:

flazio

Not a hope, same way we've yet to have driver less buses, trains or trams in this country. No better computer for the unexpected then the well trained human brain.

I sometimes wonder what would,one day, make airport security obsolete. I think the ability to slam a Big Red Button in the cockpit & hand control over to a computer, no more user input, would be it.

Aka Ishur

flazio wrote: »

Not a hope, same way we've yet to have driver less buses, trains or trams in this country. No better computer for the unexpected then the well trained human brain.

London Tube is moving to be completely driverless.

relaxed

No computer would be able to land a plane in the Hudson.

fr336

flazio wrote: »

Not a hope, same way we've yet to have driver less buses, trains or trams in this country. No better computer for the unexpected then the well trained human brain.

Indeed, though they were looking at driverless tubes in London for a while It did get postponed indefinitely due to safety concerns however:p

notharrypotter

Pilotless planes

As in totally automated or Remotely Piloted Aircraft?

Registration Requirements for Remotely Piloted Aircraft (RPA) IAA

Remotely Piloted Aircraft Systems (RPAS) EU

Do not think we are currently anywhere near the former yet.

View Profile

At most, there could be remotely operated aircraft. But that still requires a qualified pilot, even if he/she is sitting on the ground.
I don't see how that will save airlines much money.

tricky D

The Crash at San Francisco thread had a bit of (off-topic) debate on this from this post on: http://www.boards.ie/vbulletin/showpost.php?p=85509732&postcount=121 which is worth a scan.

Tabnabs

The Docklands Light railway in London is and has been fully automated since the 80s. With aircraft, of course it will happen one day, the question is, how far into the future that is?

tricky D

The comparison to rail systems isn't a great one as the amount of variables is significantly fewer mainly being 2D (excepting over/underpasses), with a large degree of linear restriction and with very few weather variables in comparison to aviation. There's a lot more of the unexpected and multiple pattern-based considerations in aviation which computing isn't so good at. Machine Intelligence hasn't moved swiftly in these types of areas, but eventually, though probably not in my lifetime.....

stevenmu

Driverless cars are coming, it's almost certain at this point that they will be commonplace within 20 years, quite possibly sooner. The famous google driverless cars are already statistically safer than human driven cars.

Driving cars is actually a much more difficult task for a computer than flying a plane, airspace and fight plans are already relatively controlled and easy to organise in advance, and you don't get random obstacles or kids jumping out from behind walls etc.

The perception is still there that flying is too dangerous for computers to be completely in control, but I suspect that when driverless cars become commonplace and (if) they are proven to be statistically much safer, then there will be both pressure and support for pilotless planes too.

fr336

What happens in a world of rapidly increasing population and fewer and fewer jobs? A few big world wars to "clear out", I suppose...

Rather depressing post there apologies.

Shamrock231

Cost benefit of all the extra equipment that would be needed on board aircraft, which adds weight and on the ground vs. the cost of paying someone 80k a year to fly the thing starts to look far fetched... It'd simply cost too much to develop the aircraft, and put in place the infastructure, for something that can currently be done relatively cheaply in the scheme of things with a pilot sitting up front.

artanevilla

A lot of people probably wouldn't get on a pilot-less plane, plus there's always need for redundancy in aircraft, a redundancy in a computer flown aircraft would be the pilot....

Strumms

It will no doubt happen. That much is certain but it could be 100 years or more. The technology has to be there, at an attainable cost and obviously all this has to lead into the customer belief that it is safe to get on a pilotless aircraft. The comparison to a light rail system isn't accurate as there is little that can go wrong with such a system in comparison to an aircraft which is a very complex machine and subject to all sorts of variables, laws, weather and gravity etc...

As regards pilots and the cost there of... It's not just the cost alone of paying a pilot you have the cost of training initial and recurrent, per diems, uniforms, travel etc... Companies will always try and introduce automation or any cost reduction if it makes sence. I'll always like the idea of two highly trained professionals up front .. It won't always be that way but nothing will chance in my lifetime or that of my kids either so anyone going for a PPL right now .. Well I wouldn't worry...

wretcheddomain

If something goes wrong the cabin crew can do a defragmentation or use PilotBot Search and Destroy.

stevenmu

Shamrock231 wrote: »

Cost benefit of all the extra equipment that would be needed on board aircraft, which adds weight and on the ground vs. the cost of paying someone 80k a year to fly the thing starts to look far fetched... It'd simply cost too much to develop the aircraft, and put in place the infastructure, for something that can currently be done relatively cheaply in the scheme of things with a pilot sitting up front.

Most commercial flights will have at least 2 pilots, so that's 160k to start, before you also include things like training costs, accommodation and other expenses etc. That's also a recurring expense, so you have to compare that against however long a computer system would last. Say a computer system would last 5 years, that is 800k on wages alone.

The weight of the extra equipment could easily be less than the weight of two pilots, and you can also remove all the cockpit equipment and instruments used by the pilots. It also has the added benefit that it could be placed anywhere on the aircraft. In theory there could actually be minimal equipment on the aircraft itself and the heavy computation could be done remotely, but that's probably going a bit too far. Either way, it would free up the space used by the pilots and cockpit for more paying passengers (and for a premium you could probably reserve the front seats which would have a fantastic view).

But the biggest cost benefit would probably be around safety. If pilotless planes proved to be safer, then the money saved on payouts after accidents would be huge, and in turn the insurance savings would be huge too.

In fact, the entire matter will probably be decided by insurance companies, both those providing cover to the aircraft manufacturers and to the airlines themselves. If they decide that pilotless planes are safer then they will push the matter through pretty quickly.

Tabnabs

stevenmu wrote: »

Driverless cars are coming, it's almost certain at this point that they will be commonplace within 20 years, quite possibly sooner. The famous google driverless cars are already statistically safer than human driven cars.

Actually according to Sebastian Thrun, the guy leading the Google car project, he reckons within 5 years they'll be commercially available. He also says the current situation is one accident every 50,000 hours or about every 4 years. This is unacceptable and will need to be vastly improved upon before going to market.

In the plane, why have two pilots when one could take off and land and then hand over to the computer for the rest? Surely the first evolution towards fully automated passenger aircraft?

View Profile

stevenmu wrote: »

Most commercial flights will have at least 2 pilots, so that's 160k to start, before you also include things like training costs, accommodation and other expenses etc. That's also a recurring expense, so you have to compare that against however long a computer system would last. Say a computer system would last 5 years, that is 800k on wages alone.


But the biggest cost benefit would probably be around safety. If pilotless planes proved to be safer, then the money saved on payouts after accidents would be huge, and in turn the insurance savings would be huge too.

There is no computer that can successfully operate an airliner from pushback to on stand independently from humans.
Therefore licenced pilots are still required and thus they need to be paid. Not really a saving there then.

Also, airlines (thankfully) may never have to payout big compensations because they may never have an accident with loss of life or serious injury.

notharrypotter

He also says the current situation is one accident every 50,000 hours or about every 4 years. This is unacceptable and will need to be vastly improved upon before going to market.

Aviation safety is currently in the order of 1* 10(-6) with efforts in Europe to increase this to 1*10(-7).

1 accident every 50,000 is not even at the races.

Google's car is worried about 2 axis (front/back and left/right) on a relatively flat surface with a maximum speed of 120 kph.

Your average commercial airliner in cruise is doing 450 KTS depending on weather this equates to a ground speed of 400 to just shy of 600 Mph.

Typically in busy terminal airspace its 250 KTS or less below 10,000 ft.

ATC use speed control and vectors to achieve the required spacing for arrivals.

So we need a computer capable of processing in 4 dimensions.
Height
Speed
Time
Location

We will be waiting a little while methinks

billie1b

stevenmu wrote: »

Most commercial flights will have at least 2 pilots, so that's 160k to start

Its not the 1970's dude, you'd be lucky if a captain alone would earn 160k over a 2-3 year period

folbotcar

We're a long way from pilotless passenger carrying airliners. Only after years of reliable pilotless cargo operations will anyone even think of getting rid of the crew. But we're years from removing pilots from cargo aircraft.

There's a general impression among the general public and among people who should know better that flying is almost fully automated and that pilots have little input into the actual flying of modern airliners. The reality is quite different.

You only have to look at recent accidents and incidents to see where automation was either no help at all or may have actually caused the problem. The Airbus into the Hudson is one example. Another is the Heathrow 777 were only the skill of the pilots saved everyone on board. No computer could have coped with either situation.

We have AF447 where the automatics couldn't cope and essentially handed over to the crew who couldn't cope either. But the point being that the automatics lost control first not the crew.

Then there was the Turkish airlines 737 that crashed at Schipol. There was a faulty radar altimeter which caused the autopilot to assume the aircraft was on the ground and retarded the throttles on short finals. The crew failed to see this in time and the aircraft crashed. So a combination of pilot and computer error. That would be fine if it was a one off but as it turned out the altimeter problem had occurred previously on the same aircraft and on other 737s. In every other case the crew saw it in time and recovered. No thanks to the automatics.

But you don't have to look at the disasters. You only have to look at the daily decisions all pilots make when things go wrong from weather to passenger disturbances.

There's not point in making fatuous comparisons with cars and trains. Aeroplanes are different and of an order far more complex than anything that travels on the ground.

We are a long way from pilotless airliners. Our children might see it but I doubt it.

murphaph

relaxed wrote: »

No computer would be able to land a plane in the Hudson.

Yeah and no computer would ever beat a chess grand master....oh, wait.

At some, probably distant time in the future all vehicles including aircraft will be fully automatic and there will be fewer car accidents and plane crashes as a result.

folbotcar

It will be the distant future at a time when computers themselves stop crashing at every opportunity as mine just did ten minutes ago. Computers are far from a mature technology.

When driverless cars, trains trucks and buses are perfected then maybe someone should consider trying to pilotless aircraft. The routine, indeed the banality of air travel these days shrouds the sheer complexity of the process. That's very hard to get across to people.

irishmover

This is already in operation and thus far has been successful.

I also notice people are talking about the need for computers to develop substantially in order for pilot-less aircraft to function. Nobody seems to be talking about how much GNSS (Global Navigation Satellite Systems) or GPS for the lay person will need to advance in order to accurately determine the position of a 1000km/hr aircraft.

Computers aren't the only thing which needs to advance in order for this to work.

The reason why cars aren't far off is due to the advancement in Terrestrial Laser Scanning over the past 15 years. GNSS isn't the main means of the car being driver-less hence why this is possible.

fr336

Cars I could agree with - they definitely need safer drivers a lot of the time;), but it's slightly insulting to compare that to pilots tbh.

BonkeyDonker

Is it not the case now that an aircraft can be flown from just airborne to landing on autopilot?? All that would be needed was an interface to the flight management system to allow any ATC instructions - possibly themselves computer generated (think 4d flightplans) and a reconfiguration of the system to do some of the manual congifuration stuff (flaps, gear etc.)
The interface would also allow a ground based pilot to intervene in the case of an emergency - but you could have one or two pilots "monitoring" several flights at once.

Plus as said above, you could remove the cockpit, and charge extra for the privilege of being the first to the accident when something goes wrong :eek:

Popoutman

folbotcar wrote: »

It will be the distant future at a time when computers themselves stop crashing at every opportunity as mine just did ten minutes ago. Computers are far from a mature technology..

Computing *is* a mature technology. I'm working as a systems admin for a multinational, and I regularly have unix and linux servers with uptimes over one year. I have one cluster with uptimes of 3000 days. The most common reasons for systems going down is for preventative maintenance on the power circuits in the datacentres, or for upgrades to the servers themselves. My linux-based desktop computers at home running my torrentserver and firewall are regularly up for more than 6 months on commodity hardware.

No, as for Microsoft operating systems that's a different story. We are too conditioned to these crashing or rebooting regularly. It's specific to Microsoft, not specific to computing in general.

bigblackmug

This is why there will be pilotless planes
http://en.wikipedia.org/wiki/World%27s_largest_airlines#By_Scheduled_Freight_Tonne-Kilometres_Flown_.28Millions.29

The potential savings are too high to ignore.

I can see organised countries like Germany actually encouraging shift of freight from road/rail to air so that they don't have to invest as much in maintenance of their road infrastructure.

when/if freight lines show their reliability then there will be a push for pilotless passenger planes; a couple of decades away before this happens.

IRLConor

Yeah, it'll happen with freight first and probably only semi-automated too. A remote pilot supervising and/or performing takeoff and landing and letting the computer control the rest of the flight.

I'd say we'll see fully-automated freight before we see even semi-automated passenger service though.

Popoutman

Air freight is a lot more energy intensive than either rail or road, and that will end up being a large factor in the costings and taxing on freight movements. It's significantly cheaper to move a few tons by road than by air, even if the automation was cheap.

artanevilla

irishmover wrote: »

This is already in operation and thus far has been successful.

I also notice people are talking about the need for computers to develop substantially in order for pilot-less aircraft to function. Nobody seems to be talking about how much GNSS (Global Navigation Satellite Systems) or GPS for the lay person will need to advance in order to accurately determine the position of a 1000km/hr aircraft.

Computers aren't the only thing which needs to advance in order for this to work.

The reason why cars aren't far off is due to the advancement in Terrestrial Laser Scanning over the past 15 years. GNSS isn't the main means of the car being driver-less hence why this is possible.

And the fact that that GPS is controlled by the US Military...

bigblackmug

Popoutman wrote: »

Air freight is a lot more energy intensive than either rail or road, and that will end up being a large factor in the costings and taxing on freight movements. It's significantly cheaper to move a few tons by road than by air, even if the automation was cheap.

Air freight without presurised cabins using turboprops would be quite cheap.
Pilotless freight planes will not be former civilian passenger jets. They will be dedicated built for purpose machines.
They can even be single prop as there is nobody in them so you don't care so much if they fall out of the sky(within reason).

gctest50

"It's cool and fun to fly a plane, but that doesn't mean I'm going to be better at flying it than a computer. The computer doesn't get tired," she says.

http://esd.mit.edu/staging/Faculty_Pages/cummings/cummings.htm

http://www.usnews.com/news/articles/2013/05/13/britain-tests-pilotless-passenger-plane

irishmover

artanevilla wrote: »

And the fact that that GPS is controlled by the US Military...

That's coz GPS is American. GNSS on the other hand isn't. That refers to the collective. Galileo system, Glonass and all the rest are not controlled by America.

So I must disagree.

artanevilla

GNSS isn't operational yet.

irishmover

artanevilla wrote: »

GNSS isn't operational yet.

You do realize what GNSS refers to? It's not one system. It's all systems.

Coming from a surveyor who utilizes GPS and Glonass I can most certainly tell you the GNSS is operational. As GPS is GNSS and Glonass is GNSS.

artanevilla

Sorry I meant the European GNSS. Which is the only viable one for pilot-less a/c as it's the only civilian one currently.

L1011

artanevilla wrote: »

GNSS isn't operational yet.

Galileo isn't operational yet, huge difference.

GLONASS alone works perfectly well since coverage was restored in 2011 anyway.

artanevilla wrote: »

Sorry I meant the European GNSS. Which is the only viable one for pilot-less a/c as it's the only civilian one currently.

That's still wrong...

irishmover

artanevilla wrote: »

Sorry I meant the European GNSS. Which is the only viable one for pilot-less a/c as it's the only civilian one currently.

It was explained above me so I won't correct you.

I might have sounded blunt and for that I apologise but I myself am beginning to grow increasingly irritated by the lack of knowledge regarding GNSS. I even had a Surveyor (albeit non degree qualified) try to convince me that a GNSS tagged Survey grade receiver meant Glonass and not the GNSS I was referring to.

The misunderstanding of it is crazy.

However, back on point, GNSS will definitely need to advance further if we're to see accurate positioning of a 1000km/hr tin can.

Tabnabs

The boffins seem to think there's a reasonable chance for pilots to be replaced with computers within the next 20 years.

IngazZagni

Tabnabs wrote: »

The boffins seem to think there's a reasonable chance for pilots to be replaced with computers within the next 20 years.

]

It doesn't say replaced. It says leading to job losses. I think there is no doubt that there will be an increasing trend especially in the military for the use of automated equipment replacing human driven equipment. I'd expect an increase in automated fighter jets over the next few decades. I'd also expect to see an automated cargo jet in my lifetime carrying actual goods people have ordered but I don't expect to see a fully automated passenger jet in the next 50 years at least.

Look at all the commercial aircraft planning on being introduced by 2020. All of them still require a minimum of 2 flight deck crew and have a similar level of automation. I wouldn't hold my breath.

folbotcar

I think they'll be talking about replacing pilots for the next hundred years. It's going to need a very advanced form of artificial intelligence before it happens. The problem with replacing pilots on the flight deck is that you need to replace their entire role.

There does seem to be a fundamental misunderstanding of the modern pilot's job. A pilot will spot something wrong before it becomes an issue and correct for it long before it becomes dangerous. I gave the example of the Turkish airlines accident in my previous post. I made the point that the same problem occurred in other 737s. In every other case the crew spotted it and corrected it before it became a problem. Again the Heathrow 777, without crew input it would have come down short of the runway with everyone dead. Until artificial intelligence comes close to that of humans.

Automating airliners is not the same as automating trains or cars. It's also worth pointing that driverless mainline trains aren't exactly the norm are they? We're getting way ahead of ourselves in trying to get pilots off the flight deck when the relatively simple train still has need for a driver in most cases.

No it will be long time before pilot jobs become ground based.

murphaph

I don't think anyone suggested it was just around the corner.

folbotcar

murphaph wrote: »

I don't think anyone suggested it was just around the corner.

You might not but some people seem to think it is. The recurring phrase is: 'Sure, don't they fly themselves these days.' Not helped by a certain O'Leary character and his tongue in cheek comments.