Pilotless Flight UK Airspace Milestone

[Always Up]

Another milestone passed. The commentator reckons 10 years before we see this in mainstream use. I wonder in what role we will see them first and how long before they carry passengers - I would hate to say never. But I think they still have one hell of a job to convince the first passengers to fly without a crew - pioneering in the true sense. Hey just thought perhaps Ryan Air would offer a pilot on a surcharge basis :ok:

BBC News - Pilotless flight trialled in UK shared airspace (http://www.bbc.co.uk/news/technology-22511395)

[VX275]

No passengers? What was the pilot if he wasn't flying it? Answer. A passenger!

I'll only be impressed when the flight is repeated with no flesh and blood on board.

[VinRouge]

I can imagine single pilot ops with pilot crew resting whilst George does everything. Pilot will remain purely for emergencies. Too much potential for a whole new meaning to "unlawful Interference"

[Lonewolf_50]

Pilotless Flight

Sorry, I thought this thread was going to be about Airbus ...

*ducks and hides* :}:E:8

[Stuffy]

Boeing did it in 1981, with a modified B737-200. They had a cockpit in the cabin for back-up.

I believe it flew automatically across the Atlantic, the pilots taking over for the landing.

It was 30+ years ago. I think the aircraft was painted in NASA colours.

It stayed in the same time box. If it fell behind, the aircraft would speed up for a short time. If it became early the aircraft would slow a little. All so the B737, stayed in that time box.

It was successful, yet it was during 1981 !

[BEagle]

At least an unmanned airliner would protect passengers from the type of 'pilots' who destroy serviceable airliners and kill their passengers as happened with the Air France A330 over the Atlantic and the Turkish Airlines Boeing 737 at Schipol.

I cannot see many fare-paying passengers agreeing to travel 'drone class' though.

[OvertHawk]

I can see Fedex, DHL and the other big freight carriers being very interested in this... Freight will not complain if there is no flesh and blood up front.

OH

[Red Line Entry]

Joe Public clearly has a gut-reaction aversion to flying as a passenger on a pilotless aircraft.

However, surely the most sensible approach would be to compare the predicted accident rate in a pilotless aircraft vs the historic accident rate in a piloted one? In one you gain by avoidance of aircrew error, in the other you gain by the ability to react to unusual circumstances.

[Trim Stab]

But during the 500-mile journey, the specially adapted plane was controlled by a pilot on the ground, instructed by the National Air Traffic Services.

So remotely piloted, not "pilotless". So just as much chance (if not more) of human error. And how does the pilot on the ground make all the other judgment calls that an airline pilot makes during a flight (e.g. avoiding weather, turbulence, birds, other stray aircraft). Does he have to rely on technology to replace Mk1 eyeball?

kill their passengers as happened with the Air France A330 over the Atlantic

That accident was triggered by technical failure (pitot icing) which sent conflicting information to the PFD which was then mis-diagnosed by the crew. But would a computer have diagnosed the problem better? Conflicting information is just the sort of situation that humans are generally better at evaluating than computers. It would be very hard, if not impossible, to write software that could pre-empt every unusual situation which has and could occur on the flight-deck.

[BEagle]

That accident was triggered by technical failure (pitot icing) which sent conflicting information to the PFD which was then mis-diagnosed by the crew. But would a computer have diagnosed the problem better?


Not so much 'mis-diagnosed' as 'grossly mis-handled' by a clearly less than competent crew.

[VX275]

Joe Public clearly has a gut-reaction aversion to flying as a passenger on a pilotless aircraft.



How does Joe Public / SLF know there is a flight crew on their aircraft these days? With the cockpit door closed and bolted unless the cockipt has been spied through a gap in the airbridge the only contact the SLF have with the cockpit is the disembodied "Captain Speaking" voice. Even in the days when one could see a crew up front through an open door I always thought that some of those broadcasts were recordings, they might just as well be ground transmissions.

I think that Joe Public's opinion on crewless flight decks is similar to 'his' dislike of rear facing seats, everyone knows what Joe Public wants but no one has ever asked him.

[Whopity]

It would be very hard, if not impossible, to write software that could pre-empt every unusual situation which has and could occur on the flight-deck.Just programming some basic aerodynamics would have solved this one Sinclair level computing!

[Fareastdriver]

If a pilotless aircraft flies into a skyscraper; who do you blame?

[Trim Stab]

Just programming some basic aerodynamics would have solved this one Sinclair level computing!

No, because there was conflicting and fluctuating ASI information from the pitots. Software which could reliably deconflict every possible combination of sensor failures (not just pitot - but all the others too) on a complex aircraft would be a massive job. Undoubtedly better trained pilots than the AF pilots could have saved that aircraft, but I doubt that a computer would have succeeded.

[dead_pan]

I doubt that a computer would have succeeded

Surely you could throw all manner of problems at a computer and see how it copes, then make changes and improvements as required. This could be done a lot faster than testing a human crew in a simulator, allowing a much larger range of faults and scenarios to be run.

Still, it would be nice to think that someone was on hand, if only to reassure to the PAX in the event of an emergency. I suppose some bloke in a nice comfy chair at the airline's maintenance center could do the job - a career for all those drone 'pilots' when they leave Waddington?

[HaveQuick2]

Would the Hudson River incident have been handled the same by computers?

[t43562]

I went to a lecture at Cranfield by the gentleman in the video last week. I'm interested because of the computing aspects.

They have looked at a lot of issues e.g.:


Integrating with air traffic control seamlessly without the need to change ATC at all. i.e. at least initially the ground pilot's voice is sent to the plane and then transmitted from there by whatever normal radios the aircraft might use.
When there are a lot of unmanned vehicles they are going to have problems with bandwidth. Satellite is too expensive and limited. Solution: ad hoc network (data can hop from one aircraft to the next).
How well to the ground pilots cope with the latency that can be introduced in areas with more or less air traffic. Solution: make sure the best low-latency communications are available in such areas, allow the aircraft to "find the best channel."
How the aircraft will notice weather conditions, other aircraft and inform the ground pilot and then take action if he eventually does nothing.

It's all here:

Presentations from ASTRAEA National Conference 2012 (http://www.astraea.aero/presentations/presentations-from-astraea-2.html)

[AR1]

Flown without pilot intervention? - We could call it 'Autopilot'

[dead_pan]

Would the Hudson River incident have been handled the same by computers?

Depends how intelligent it was and/or how quickly a man could be put in the loop if it did throw in the towel.

Perhaps in such an extreme situation the PAX could be polled on the forced landing options available? That way if they did all perish at least we would know they did so democratically.

[t43562]

This sums it up fairly well. I don't think the intention is for it to be used with passengers. They also make the point that most of these things can be useful with manned aircraft too - in the same way that modern cars can read speed limit signs and warn you if you're too fast or auto-park or can brake when they perceive some very immediate danger.

The bit about UAVs refuelling in flight was interesting.

Probably worth skipping the first 3 minutes.

i9ztB-JXVJw&t=3m320s

[Fox3WheresMyBanana]

Worth bearing in mind the computing arrangements for driverless trains. As a guy designing them put it to me "If a supercomputer is a 10/10 and your mobile phone a 2/10, the computing power on this train is a...0/10"
The more there is, the more that can go wrong.

Also;
1) humans are very bad at monitoring, as opposed to controlling
2) Even if they do take over at the correct point, how will they fly it with no practice?
3) Sorting out a complex computer failure with unreliable info and flying a plane takes a lot of skill, which is....expensive to hire. Anyone see a trend in aviation towards paying pilots what they're worth?

The only model I see working is a central control with a few highly skilled pilots for the rare emergencies.

Your gravestone will be marked 'Uplink failed!', or more likely

'Windows has encountered a problem and needs to kill your passengers.....':uhoh:

[dead_pan]

...or 'Game Over' :}

[CoffmanStarter]

I, for one, won't be getting on any aeroplane without a properly trained human pilot upfront ... End of :cool:

This was scary enough most days :eek:

http://www.tfl.gov.uk/assets/images/square/new-DLR-train.jpg

[Trim Stab]

If this technology became mainstream I also wonder where they would get the ground pilots from? Most of the reason people become pilots are not applicable to ground pilots.

[Fox3WheresMyBanana]

Coffman - the designer guy did the DLR controls. They are amazingly simple. each component is tested and it's failsafe. Same for the signalling. I happily ride at the front knowing this.

[t43562]

For anyone that's interested there's a chance to ask questions about the programme that resulted in this pilotless flight:

The Engineer Q&A: ASTRAEA autonomous aircraft project | News | The Engineer (http://www.theengineer.co.uk/aerospace/news/the-engineer-qa-astraea-autonomous-aircraft-project/1016347.article)

[Wensleydale]

'Windows has encountered a problem


It seems ironic that you need windows for a pilotless aircraft.

[parabellum]

It won't happen until all security risks have been overcome, including a terrorist take-over of a ground station or the use of a jamming station that could take over control. Either scenario will cause a blood bath of biblical proportions.

Interesting, this post was made after pittextra's!:confused:

[Basil]

Pittsextra, Hear, hear!

Ain't going to happen for a long time. In '89, I flew a new, very large, computerised aircraft. On a number of occasions (on the ground) we had to power the whole thing down and re-establish electrics just to reboot the on-board computers when they were acting up a bit. I can assure you that I will NOT be flying as passenger in an unmanned aeroplane. Just look at the computers in your day to day life; would you really trust them?

p.s. Prunetime is 40min slow so this may appear before the posting to which it refers - computers - eh? :hmm:

[Courtney Mil]

You do for the passengers to look out of.

[Pittsextra]

IMO I don't think we will see this in our lifetimes without huge change been accepted by everyone else sharing its airspace.

One of the biggest problems is uncontrolled airspace where currently "see and avoid" ensures X doesn't bump into Y. When one also considers that, especially at this time of year, many GA pilots are resuming their flying activities after a winter break this uncontrolled airspace is also full of the rusty and/or inexperienced....

Secondary to this is the control of such machines in terms of airworthiness and airmanship. There are some great stories surrounding the operation in the US of Israeli made UAV's.

The CIA were given these UAV's to operate as it was thought that the asset had a better fit there than with the USAF.

Following some early losses it was suggested that the cause had been icing, which confused the CIA operaters as they couldn't square icing and the heat of Nevada.

Then of course who is responsible for the default system if the communication with its base fails? Again there is a story where Israeli made UAV's tried to return "home" from the US as whilst everyone believed there was a fail safe system nobody seemed to check where that home was..

So to suggest these machines are "pilotless" will never be entirely accurate.

[Flap62]

BEagle,
The Air France incident would have had exactly the same outcome if flown by autonomous means. If the system (pitot) fails then the computer has false information and cannot do anything but react to it. You are right that the crew should have done a lot better in that situation but that is why we need people in the loop.

[Davef68]

How does Joe Public / SLF know there is a flight crew on their aircraft these days? With the cockpit door closed and bolted unless the cockipt has been spied through a gap in the airbridge the only contact the SLF have with the cockpit is the disembodied "Captain Speaking" voice. .

On the 40+ flights I have made as SLF in the last year, i can say with some confidence I have seen the flight crew on every one. Doors are usually closed/bolted after the pax have been seated, and the crew often appear after the flight to say 'goodbye'

[Basil]

Flap62, The Air France incident would have had exactly the same outcome if flown by autonomous means.
Now you mention that, it got me thinking (and it hurts :p ).
What if there was a fallback mode to fly Mach then AoA then GPS groundspeed/forecast wind then power/pitch ? Hmm could have got the aircraft out of the situation.
Perhaps Airbus will think about incorporating those modes.

[Trim Stab]

What if there was a fallback mode to fly Mach then AoA then GPS groundspeed/forecast wind then power/pitch ? Hmm could have got the aircraft out of the situation.
Perhaps Airbus will think about incorporating those modes.


But then what happens if (say) the AoA vane fails? Or the ECU which is inputting engine power to the flight control system? Or any one of the hundreds if not thousands of other data sensors on a complicated machine such as the Airbus? Writing software that could pre-empt and prioritise all of the millions of possible fail-permutations would be a Herculean task and even then the software would be full of bugs that might not be uncovered until too late.