Robo Pilot

[BlankBox]

https://www.newscientist.com/article/2214731-robot-pilot-that-can-grab-the-flight-controls-gets-its-plane-licence/

...how long before they take over ?

[Mostly Harmless]

Soon.
https://vignette.wikia.nocookie.net/lostinspace/images/3/3f/Robot.jpg/revision/latest?cb=20090823210437

Real soon.
https://upload.wikimedia.org/wikipedia/commons/thumb/a/ac/Robbie_the_Robot_San_Diego_Comic_Con_2006.jpg/240px-Robbie_the_Robot_San_Diego_Comic_Con_2006.jpg

[UltraFan]

In the 90s there was a popular joke:
- How would it be if Microsoft was building airplanes?
- It would fly at the speed of light, use a glass of fuel per 1 million km and crash every 15 minutes.

(Windows'95 anyone?)

Back then any discussion about pilotless planes irreversibly ended with "Would you fly a plane without a pilot?", because the unilateral answer to that was a firm "NO".

Two things changed. One, people are A LOT more "seasoned" for the technology. We trust our bank money to our cellphones and pay for things by tapping our watches on a cash register. People trust the machines. Tesla automobiles blow up on every corner and run people over but they still sell by thousands.

And two, the technology has improved. We click our mice (mouses?) on a button in an "online-store" and in a few hours the thing we only saw on a webpage is delivered from another continent to our doorstep... or thrown over the fence if you use UPS. The technology is reliable, simple and friendly. And it actually IS better. The autopilot doesn't sleep, doesn't eat, doesn't care about rest periods or unionization.

Think about it, (roughly) 99% of all accidents occur because the pilot forgot to turn a system on or off, didn't remember some bullet point in the operation manual, got distracted, misjudged the situation or made a decision too late or too wrong. The pilot is the risk factor. Remove it - and most crashes of the latter years will be avoided. The robot will never panic, will never misclaculate an approach or insert an correct number in the MTOW. It would never cock up an approach to SVO or fly from Greece to Prague on one engine. And if something goes wrong, it will at least do nothing instead of doing something stupid.

So these days, as the flying public is ready to tolerate any inconvenience or humiliation to get to their destination, the answer to the above question would be "If the tickets are cheap". As soon as they get a working system certified and as long as a fleet of robots is cheaper than a pilots union, it will be in every cockpit. The day is coming!

[Sobelena]

The only problem is that the robot will need to be programmed by a human. Programmers have never achieved 100% success, hence updates and patches. Even after extensive testing something would crop up at some point. That point may just be with a plane full of pax.

[BizJetJock]

99% of all accidents occur because the pilot forgot to turn a system on or off, didn't remember some bullet point in the operation manual, got distracted, misjudged the situation or made a decision too late or too wrong.
Except that this completely ignores the fact that those accidents represent about 0.01% of the number that would occur if pilots didn't, day in and day out, deal with computers and machinery malfunctioning. Until reliability levels go up by several orders of magnitude then pilotless flying will be the realm of military and cargo only. It will happen, but I am not losing any sleep for my career!

[Lookleft]

The 737 Max with its embedded software and single point of failure has just rewritten the stats on the cause of aircraft crashes.

[ShyTorque]

I thought "Robo pilot" meant it was about a small helicopter.

But it's obviously about a different Robinson family:

https://www.youtube.com/watch?v=RG0ochx16Dg

[calypso]

99% of all accidents occur because the pilot forgot to turn a system on or off, didn't remember some bullet point in the operation manual, got distracted, misjudged the situation or made a decision too late or too wrong.

I guess you can provide a Source for such a fascinating statistic?

[Euclideanplane]

I guess you can provide a Source for such a fascinating statistic?
Probably not with a sizable probability. Some give the figure 99% as a substitute for the value 110% that
they might have preferred, but which tends to attract pedants. Apparently didn't work this time.

[triploss]

The only problem is that the robot will need to be programmed by a human. Programmers have never achieved 100% success, hence updates and patches. Even after extensive testing something would crop up at some point. That point may just be with a plane full of pax.
As long as "that point" is reached less frequently with robo-pilot than when human-pilot is flying, then robo-pilot wins.

Human pilots don't achieve 100% success either. Robo-pilot only has to be better than human-pilot, it doesn't have to be 100% reliable.