Should ATC be automised using ML/AI technologies in real life?

Real World Aviation
#1

Hey Guys,

I was just thinking, when we have ML/AI technologies, why can’t we in real life have Automated ATCs?
An ATC gives the same command everyday nearly 99 percent of the time. When there is a new request /Emergency then a human ATC can take over.

Once this is achieved, we can have flights communicating directly with ATC instead of pilots. Sort of two way virtual communication. The flight could then get all clearances and land itself in the destination without any or very less human input.

It might sound dystopian but I believe we have the technology to do that.

What do you guys think?

#2

Is this for Infinite flight or real life? I’m confused

1 Like
#3

It’s actually for real life

1 Like
#4

This would best fit #real-world-aviation then. However, you have to be TL2 to post in that category. Keep being active, posting and liking and you’ll get there!

1 Like
#5

Ummmm, I think dis should be in real world aviation

#6

AI ATC in real life would be a disaster. It can’t spot out things a human can with our senses. Unicom is there to replace a field without any ATC services so you just announce your position without any clearance. It is vital to have another human who can actually talk to you instead of coding basically talking to you.

2 Likes
#7

Allow me to point out one huge disadvantage. We may never see another “Kennedy Steve” guy again with AI around. Just a monotone robot with no sense of humor. 😢

2 Likes
#8

I actually wrote something on AI in the control tower. It is something that could definitely work and, as a matter of fact, does work. There are many airports that are using AI to preform functions that would otherwise be mindlessly completed by humans.

Artificial Intelligence is not forecasted to takeover controller’s jobs any time soon. Many instances call for the human brain; emergencies, emotions, and creativeness simply can’t be computed.

Link to full Artificial Intelligence section

3 Likes
#10

Nice post! This is an interesting thing to discuss, so I’ve moved it over to #real-world-aviation.

I’m wary of AI ATC, but I don’t think it’s far off. We have self-driving cars, and even self-landing airplanes that don’t need a ground-based ILS signal (Garmin).

There are a lot of facets to this problem — defining the location of the aircraft in space, determining where it’s heading, what priority and separation needs to be applied, how to handle emergencies and route around weather…

I’m not sure what the best way to do it would be, but enough other things involve AI that I’m sure it could be done.

3 Likes
#11

You see, computers are designed to be predictable and follow set algorithms, and that introduces several problems:

  • While it might work well in a normal operation, in case of an emergency the computer might do something wrong, because for the AI to do something it should be trained to do so.
  • Moral questions with choosing between 2 evils is a problem for AIs in all spheres, but one thing is a tesla, and the other one are planes with tens or hubdreds of people onboard.

And also, don’t forget about the techinacl side of things:

  • Getting a dataset to train the AI would be very tricky.
  • Current communication standards are based on speech, which would either mean new voice recognition technologies that can deal with immense noise on the frequency or new text based communication, which would be good for AI, but bad for pilots.

Overall, I think we have a long way to go until AI ATC

1 Like
#12

While I don’t see it being used at very traffic heavy airports such as DEN, NRT or CDG, it could be somewhat reasonable at small airports such as GJT, DLH or GOH.

Also, while I don’t fully agree with the idea of AI ATC or AI pilots, I could see the possibility of fully autonomous ATC before autonomous planes.

1 Like
#13

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.