10

I was just wondering: is it possible to have a commercial plane with no cockpit windshield, just external cameras and video screens?

The question come in my mind after reading the nth occurrence of laser in pilot's eyes. Moreover, a structure without windscreen is surely lighter: when there are not passenger, we avoid to put windows! Which are the limits and the opportunities to this solution?

With "ever" I don't mean "in centuries" but in 30/40 years.

Gianni Alessandro
  • 3,154
  • 7
  • 30
  • 51
  • 4
    I am sure such idea was already mentioned several times here. Technically, it is completely possible. However, it has significant problem. It adds a point of failure and nobody wants to add that. – Jan Hudec Feb 19 '16 at 15:35
  • 1
    It's been done in testing before – fooot Feb 19 '16 at 15:43
  • @fooot oh, it is very interesting! That's what I meant. I will spend sometime online for some researches. Thank you. – Gianni Alessandro Feb 19 '16 at 15:50
  • @fooot I don't think there was any visibility from the aft cockpit. It was mainly to test IFR, so that was not considered a problem – aeroalias Feb 19 '16 at 16:22
  • 16
    Preventing laser attacks using aircraft without windows is like preventing burglary by building homes without windows: It's not a solution. A good solution would be to have kW-lasers onboard which fire automatically on the source in an act of self-defence. – mins Feb 19 '16 at 16:26
  • Slightly off-topic, but the Spirit of St Louis has no windscreen but rather a sideways-oriented periscope, and many drone aircraft use cameras for obvious reasons. – KJP Feb 19 '16 at 16:50
  • @aeroalias some of the tests did include various video and sensor feeds that could be displayed in the aft cockpit. – fooot Feb 19 '16 at 17:00
  • @mins I don't see that as a solution to the current laser problem. If we were talking about some kind of weapons-grade laser attacks maybe. But I think most of the laser incidents are not so much attacks on the plane but just stupid people who don't know or don't care how reckless it is. I can see those same people being even more motivated to do it if they can trigger a response from the aircraft. – TomMcW Feb 19 '16 at 17:19
  • @TomMcW: Just a joke. By kW-laser I meant a power in the order of magnitude of the kW. That should be enough :-) – mins Feb 19 '16 at 17:27
  • @mins Smoke 'em where they stand! – TomMcW Feb 19 '16 at 17:38
  • 3
    The biggest problem is depth perception, without the ability to adjust the view based on the position of the observer, its going to be difficult to judge distances. This will be especially confusing since the camera is not in the same location as the pilots head, making it hard to translate it to a 1:1 eye:camera system and having the perspective in a different position than the pilot(s). – Ron Beyer Feb 19 '16 at 17:59
  • 2
    @TomMcW: Re "stupid people", think of it as improving the gene pool :-) – jamesqf Feb 19 '16 at 18:29
  • @mins Like this? – Bassinator Feb 20 '16 at 00:47
  • In practice they would use VR goggles. The technology is mature enough that it is expected to hit the consumer market within a year or two. And it has the issues with depth perception and lag already solved (more or less). And it would probably be used only for remote controlled aircraft. Maybe for supersonic optimized military aircraft even with a pilot, but hard to think of an use case that would require carrying a pilot for such. – Ville Niemi Feb 20 '16 at 03:40
  • @HCBPshenanigans While that would get the job done, I'm pretty sure that is a megawatt class laser. :) One of the tracking lasers from that could work, though. – reirab Feb 20 '16 at 05:54
  • 1
    Some cons I consider obvious are: the much lower dynamic range of camera/screen systems, low resolution, delay of about 30milliseconds, and cameras are susceptible to laser attacks as well (in fact might already suffer from a flight heading towards the sun for a few minutes) – Hagen von Eitzen Feb 20 '16 at 15:37
  • 2
    @reirab Go big or go home. – Bassinator Feb 20 '16 at 18:57
  • @VilleNiemi Oculus Rift has already had dev kits out for a year or so and consumer version releases this summer, IIRC. Also, Google Cardboard has been around for a year and a half. – reirab Feb 20 '16 at 19:22
  • @reirab Yes, I know. That is why I am optimistically assuming something useable to be on the market within a year or two. – Ville Niemi Feb 21 '16 at 23:45

5 Answers5

15

It's entirely possible to do this now, it's simply not a good idea. The drawbacks to safety outweigh the benefits of doing so. If you have no cockpit windows an electrical or systems failure would leave pilots totally blind, without any references whatsoever. The "mark I eyeball" works in a wide variety of conditions and does not require electrical power of any kind. Plus, there's usually 4 of them in the cockpit, and chances are at least one of them will work.

GdD
  • 53,842
  • 6
  • 147
  • 212
  • 2
    With modern fly by wire systems, doesn't a total electrical system failure result in loss of control anyways? If a windowless cockpit was implemented, it would likely have the same number of redundancies as other essential flight systems. – End Anti-Semitic Hate Feb 19 '16 at 23:53
  • 1
    Misleading visual cues is major cause of incidents. Relying on vision didn't prevent the Grand Canyon Crash 1956. OK argument for slow moving civil aviation but gains as much you lose for passenger airlines. – user2617804 Feb 20 '16 at 05:00
  • 1
    @RockPaperLizard The power supplies for it could be equally redundant, but it would be hard to make the screens themselves equally redundant while preserving their position. Also, placing redundant cameras close enough that they are capable of producing the correct field of view and also far enough apart that they wouldn't be likely to get taken out by a single event (e.g. bird strike) could be tricky. Doubly so if that single event is something like icing or a hail storm. – reirab Feb 20 '16 at 06:02
  • 1
    @reirab Those are good points. I'm not sure they are insurmountable, but good points nonetheless. – End Anti-Semitic Hate Feb 20 '16 at 06:12
  • @user2617804 Your comments are true. It makes me wonder how we can envision the cockpit of the future assuming pilots could not directly view their surroundings. – End Anti-Semitic Hate Feb 20 '16 at 06:14
  • @reirab - I can't say I agree about the camera proximity thing. Post-processing need not introduce significant latency because the transformations and deformations are a function of relative position and angle; for predetermined camera positions and angles the transformation matrix can be precomputed and applied in hardware by an FPGA. This depends on having fixed cameras, so you'd have to have lots. Good cameras these days are cheap, small, light and low power. You wouldn't put a fixed camera in a dome, you'd set it behind a pinhole in the aircraft skin. Low drag. – Peter Wone Jun 02 '16 at 23:43
  • @PeterWone That's normally used for 3-D rendering. Seeing the world with cameras, making high-resolution 3-D models from the flat images, creating the correct textures for the faces, and then rendering the scene and projecting back into the appropriate 2-D views is non-trivial. It's not impossible, but it's not easy. You'd probably want something more like a programmable GPU (or several programmable GPUs) than an FPGA. Doing that fast enough to maintain a usable latency for a vehicle whose most important out-the-window views involve racing through the scene at 200 mph isn't easy. – reirab Jun 03 '16 at 03:34
  • @PeterWone At any rate, windows are much cheaper and have far fewer likely failure modes. It's also worth noting that using screens to emulate the windows assumes a fixed viewing position... which is a bad assumption when you have 2 (or more) pilots who need to be able to look at the same part of the glass and see two different images based on their position relative to the glass. An Oculus-style wearable headset for each crewmember would probably be a better solution. – reirab Jun 03 '16 at 03:39
  • Why correct the camera image to the "proper view"? I would think in many cases the eyeball view from the cockpit would really not be the ideal, but people tolerate it because they'd rather sit in the cockpit than on the outside of the aircraft. Certainly adding additional cameras would be helpful in many cases (e.g. avoiding mid-air collisions in cases where at least one plane lacks ATACS and each plane happens by chance to stay in the other plane's blind spot). – supercat Oct 31 '16 at 16:06
14

Is it possible? certainly yes. Airbus has applied for a patent for a windowless cockpit in which the external view is displayed in the cockpit using cameras and screens. The technology itself is available and some significant progress has been made in recent years in the related equipment.

Will it become a reality? Maybe. At present, there is nothing wrong with windshields in aircraft (pointing laser at aircraft is illegal; anyway, lasers can damage camera sensors too). Why fix something that ain't broken? In the future this may well be used, maybe in supersonic aircraft (For example, spike aerospace has proposed a design without cabin windows; interestingly, the design still has cockpit windows). Addition of cameras and displays adds one more layer of complexity, which is at present unnecessary.

Also, there are other issues to consider. In order to give the pilot with a view at least as good as the exiting aircraft, images from multiple cameras have to be 'stitched' and displayed. The system should be able to do this without any lag (F-35 had a similar problem while joining feeds from its DAS). Unless there is a compelling reason for the addition of cameras and displays, the windscreen is here to stay.

FreeMan
  • 16,245
  • 16
  • 87
  • 166
aeroalias
  • 100,255
  • 5
  • 278
  • 429
  • 2
    "The system should be able to do this without any lag" There are studies about how much "lag" a human can notice and that's not even "can't compensate for", IIRC it's around 60ms which in computer time is eternity these days. I highly doubt that joining static cameras onto a static display was the problem, it was probably the fact that the displays on the F-35 were not static. Also the F-35 uses very old computer technology. – Sam Feb 19 '16 at 19:43
  • 5
    One reason to do this would be that windshield are expensive (>$20K) and have a fairly high failure rate, in my experience #1 windshields last about 6 years at a commercial airline. They are also very heavy. They create a weak point in the fuselage requiring heavier structure and usually force a less than idea aerodynamic shape. So there are reasons other than laser pointers to do it. – OSUZorba Feb 19 '16 at 23:21
  • 2
    @Sam, I am not sure what the lag is, but I know that lag in video causes issues on KC-10 refueling operators. Basically there is enough lag to cause the operator to start pilot induced osculation. I know there was a lot of work on the KC-767 and KC-46 remote aerial refueling stations to remove as much lag as possible, which of course is easier with today's technology. The other problem with remote aerial refueling operations is the loss of depth perception. – OSUZorba Feb 19 '16 at 23:26
  • 2
    Well I hope they don't get that patent then! Clearly the NASA plane has already done this before so it's not a new idea. – curious_cat Feb 20 '16 at 03:51
  • @Sam Around 60 ms? Are you suggesting that a human wouldn't notice a 17 FPS frame rate? If so, I know a lot of gamers who will adamantly disagree with you. – reirab Feb 20 '16 at 06:06
  • 2
    @reirab - 60ms latency, not a 60ms refresh cycle. You're seeing things 60ms after they happen. – Compro01 Feb 20 '16 at 10:05
  • 1
    @reirab, that is not at all what I am saying. Latency and refresh rate have nothing to do with each other. – Sam Feb 20 '16 at 14:34
  • 1
    60ms may seem like an eternity in computer time, but a lot of digital video systems work by having a camera capturing 60 frames/second (start with 16ms of latency from one frame to the next), wait for an entire frame to be loaded into RAM before starting processing on it (which may be another 15ms), then do some processing, and then start feeding it to a display (which could be another 15ms). If the input and output frame timings aren't synchronized, that could add another 15ms to the mix. Pretty sad considering that old analog video had a 17ms worst-case latency for fast motion. – supercat Oct 31 '16 at 16:02
2

That should be ... exciting ... during an electrical failure.

Good luck getting it approved for use anywhere in the world, you can't even build a pure glass cockpit in many jurisdictions. Likewise you can have all the cameras you like provided you can still look out the window.

It's certainly technically feasible but it's not a good idea.

Peter Wone
  • 343
  • 3
  • 7
1

Somewhat, but the lack of depth perception may be an issue on the taxiways.

But if lasers is the concern, I think they should install something similar to the auto darkening lenses used by welders to the front windows.

phiz
  • 11
  • 1
  • You could at least reduce, if not completely solve, this problem by using VR headsets (e.g. Oculus Rift.) You could even attach cameras to the headset itself and have the system programmed to display the cockpit interior except when the pilot is looking where the windows would otherwise be. – reirab Feb 20 '16 at 06:17
-2

Technically possible but not very comfortable. As a first step, windowless passenger aircraft could be attempted. In crowded Indian buses with windows being swamped by opaque advertising material it is very disconcerting not knowing where you have reached. Passenger experiences can be taken as a starting point. Also the experience of astronauts could be considered....

sn tekur
  • 1