Would you fly on an unmanned commercial air transport?

[jhae2.718]

As an aero, I'm rather interested on what PFers would feel about flying on an UAV airliner...

I'm kind of mixed; on one hand I would have no problems, but on the other hand I know some of the people who would be designing the control laws!

I don't think the general public would want to fly on an aircraft without a pilot on board.

Neat IEEE Spectrum article on the subject, and the catalyst for the thread: http://spectrum.ieee.org/aerospace/aviation/when-will-we-have-unmanned-commercial-airliners/0

 

[Drakkith]

I would provided all applicable safety requirements are met. It's not like aircraft don't already have flight control systems failures. I see little difference really.
This of course assumes that the technology can support full scale, safe flying, even in adverse conditions.

 

[Andre]

Pilotless is an absolute no-no because you cannot program Situational Awareness and Airmanship.

But I see no problems for single pilot operations.

 

[Drakkith]

Pilotless is an absolute no-no because you cannot program Situational Awareness and Airmanship.

But I see no problems for single pilot operations.

Wouldn't you still have pilots controlling the aircraft through remote control, similar to military UAV's?

 

[jhae2.718]

Along similar lines, there are issues even with pilots dealing with automated systems.

 

[jhae2.718]

Wouldn't you still have pilots controlling the aircraft through remote control, similar to military UAV's?

There's still the SA issue, at least in the near term. You're not going to get the same situational awareness over sensors as you would with a physical presence in the cockpit.

 

[Greg Bernhardt]

Right now I would not ride a plane that doesn't have someone capable of managing the plane from within the plane.

 

[Drakkith]

There's still the SA issue, at least in the near term. You're not going to get the same situational awareness over sensors as you would with a physical presence in the cockpit.

What SA issue?

 

[jhae2.718]

Operator disconnect from physical cues, difficulty in scanning for visual cues, plus possible communications lag in the imagery projected in the ground control station. For example, Predator, if I recall correctly, only has a forward FOV of 30 degrees for the operator.

Edit: of course, you could always add additional cameras and such, but there's still going to be that disconnect. There's going to need to be more human factors research in the area.

 

[Drakkith]

Operator disconnect from physical cues, difficulty in scanning for visual cues, plus possible communications lag in the imagery projected in the ground control station. For example, Predator, if I recall correctly, only has a forward FOV of 30 degrees for the operator.

Edit: of course, you could always add additional cameras and such, but there's still going to be that disconnect. There's going to need to be more human factors research in the area.

I am unfamiliar with the details of everything, so all I can say is that if these factors are taken care of safely then I am all for it.

 

[jhae2.718]

I am unfamiliar with the details of everything, so all I can say is that if these factors are taken care of safely then I am all for it.

If you've ever used a PC flight simulator, you can observe the lack of SA from a forward-fixed instantaneous field-of-view.

 

[Drakkith]

If you've ever used a PC flight simulator, you can observe the lack of SA from a forward-fixed instantaneous field-of-view.

Yes but that is easily fixable as you said. Given that we already have UAV's I believe it is very likely that it is only a matter of time until the technology matures enough for commercial use. Like anything it will take time to get mainstream.

 

[jhae2.718]

I'm mostly talking in the short term; long term, I agree completely. A lot of the technology that would be needed for unmanned CATs would also be required for UCAS capable of air combat missions, so there's definitely going to be work in the area.

 

[Drakkith]

I'm mostly talking in the short term; long term, I agree completely. A lot of the technology that would be needed for unmanned CATs would also be required for UCAS capable of air combat missions, so there's definitely going to be work in the area.

Oh. Yeah, in the short term this isn't going to happen.

 

[AlephZero]

As another "aero", I wouldn't have any problem with it so long as the system met the same standard as current regulations for piloted flight. And I can't see that happening any time soon.

I would take it as "self-evident" that any unmanned system would have to be capable of acting on the equivalent of air traffic control instructions, so in that sense it would never be completely automonous.

Don't forget that most of the time civil aircraft are not being flown (hands-on) by the crew already.

@Andre, you could envisage a "convoy" system where N aircraft flying in formation have one flight crew. That is already being tested for road vehicles as a way to improve safety and increase traffic density (i.e. have 1 good driver compared with N average-to-bad ones).

It's hard to judge the real risks involved here. You can easily counter every "heroic" feat of piloting like successfully ditching a plane into the Hudson river after a birdstrike, with an incident where the only reason a plane flew into the side of a mountain was the crew didn't realize the mountain was there until they hit it.

 

[Andre]

Interesting to see that so many people are more than happy to put their lives in the mechands* of a mindless system. Oh sure no need for a pilot to control an aircraft manually, form take off to landing, the autopilot can do everything, but is that the whole story?

So what are all the problems? I already mentioned SA and airmanship, that is best reflected in the definition of a superior pilot, who is a pilot who uses his superior judgement to avoid getting into situations which require his superior skills. How would a robot do?

Directly related is the ability to recognise an accident chain. Accidents are usually the result of complex situations, like "if not this and if not that and if not more this", etc, etc, then nothing would have happened. Pilots lear to catch one or more of those "thisses" and "thats" in time and for any acident that did happen, there may have been a dozen near-accidents that were caught in time by a thinking human being, before it getting a problem. How would the pilotless aircraft do?.

Okay as suggested we put a remote man in the control loop, who takes the decisions. Apart from the loss of SA as mentioned, how is this guy getting his information anyway? We are talking about gigabytes of visual clues per seconds. Surely no problem for super fast computers. But you'd have to send that information using the Electromagnetic Spectrum and obviously physical laws limit the flow of information there considerably. And then we are talking about only one remotely piloted vehicle. But how many starts and landings are there on your local airport? And mind that there is a big competition going on continuously in the division of the usuable bandwidth for radio transmissions. It's full.

There are a few more elements but the post should not get too long. So I'll return for that.




*I put "hands" in there first, but there are no hands obviously

 

[zoobyshoe]

I would be willing to do it if, after 30 years of letting other people be the guinea pigs, it turned out to be no more dangerous than the present system.

 

[Jobrag]

Ladies and Gentlemen welcome to the first pilotless commercial flight, this system has been tested extensively, please be assured that absolutely nothing can go wrong, go wrong, go wrong, go wrong, go wrong...

 

[MarcoD]

I used to be in IT and I would never fly an automated airplane. I agree with automation because computers can solve problems flying whereas tired worn-out humans would only make mistakes in certain situations. But the reverse is also true, computers can never solve unexpected scenarios.

I doubt fully automating a plane is worth the increased risk. I mean, seriously, the backup system, a pilot, is just worth his money because he typically supervises a few hundred lives.

 

[Filip Larsen]

I think it is interesting to note, that current systems of automated flight (autopilot and autoland systems) are often designed so that they only can be engaged and operate within a limited parameter set of parameters and only within these limits is the system "guaranteed" to work. Outside of these limits the pilot must be ready to take over (in part or in full depending on the situation), meaning that such systems can be designed around a much simpler safety model than a fully autonomous system can be. Auto-flight systems of today should probably be viewed more as systems designed to reduce the workload on the pilots while maintaining safety and expedient flight than systems designed to be early versions of a fully autonomous autopilot.

 

[Andre]

...Auto-flight systems of today should probably be viewed more as systems designed to reduce the workload on the pilots while maintaining safety and expedient flight than systems designed to be early versions of a fully autonomous autopilot.

Right, maybe an acceptable parallel is photography. In the old days the photographer would have to measure light and determine the best shutter speed and aperture at a given film sensitivity and he'd have to focus too. Today you just push a button. Never mind all that, but there still a human being in the loop, taking the decisions what to shoot.

Further complications re my previous post...

...So what are all the problems? ... a superior pilot, who is a pilot who uses his superior judgement to avoid getting into situations which require his superior skills. How would a robot do?

Directly related is the ability to recognise an accident chain. ...

... But you'd have to send that information using the Electromagnetic Spectrum and obviously physical laws limit the flow of information there considerably. ...

Also think air traffic control, that's a whisker more complicated than traffic lights. Flights have to be deconflicted and sequenced. Ideally flights follow predetermined patterns which are easy for unmanned vehicles, but reality requires frequent deviations with seconds of decision- and reaction times. How to coordinate that with multiple ground based UAV controllers, hundreds of miles away?

Then what is a commercial flight anyway? Basically it's a mobile organisation with multiple crews for catering, technical control, load supervision, flight planning and decision making away from homebase on all those things. Bottomline is, that there will still be people on board with functions. Maybe that's most easy if one of them also would happen to be a pilot?

 

[Dale]

I would be willing to do it if, after 30 years of letting other people be the guinea pigs, it turned out to be no more dangerous than the present system.

This is my take too. They should fly them on military and cargo flights for decades before moving them to passenger airliners. After being tested for several decades and proven at least as safe then I would fly on one.

However, I would be highly skeptical that the data would prove favorable to pilotless aircraft.

 

[S_Happens]

I'm not an aero, but after working in chemical process automation I can say that I'm not interested in riding in a plane that doesn't have a competent pilot on board to oversee the automation.

No matter how much you automate a system there will be situations that are simply doomed to failure if not recognized/corrected from an outside source. It can be part situational awareness and part compromise based on optimizing normal operating conditions. There is no question that there WILL be times when full manual override is needed. Obviously you can't predict when or what ALL of those situations will be.

It's amazing sometimes how many redundancies can fail, or how a single stuck input/bad variable can mean total failure.