Control Towers Without Humans

LevelWing

Pre-takeoff checklist
Joined
Jul 2, 2013
Messages
465
Location
GA
Display Name

Display name:
LevelWing
NOW ARRIVING: AIRPORT CONTROL TOWERS WITH NO HUMANS INSIDE

This is an interesting concept. I can see the positives and negatives to this. Despite the positives, I can't see how taking controllers away altogether is a good idea. I think the implementation of the high definition cameras can aid existing towers with better awareness, especially in lower visibility conditions, if they don't already have them.

The article said that the NATCA was involved in the testing but I don't see how they'd support this if it meant the eventual loss of jobs.
 
Between drones and unmanned towers,aviation is going to get more dangerous rather than safer.
 
I thought this would be an automated system, but no, "No jobs have been eliminated but ultimately such systems will allow tiny airports to pool controllers."

What happens when there are conflicts at more than one airport at the same time?
 
What happens when there are conflicts at more than one airport at the same time?

Same thing that happens when there's more than one conflict at the same airport at the same time: the controller multitasks.

But with remote controllers, there may be another option too: load-balance by shifting some of the conflicts to a currently idle controller.
 
Last edited:
I think it could be great if they had onsite radar for each airport. Not sure having a manned tower is as good as having good radar re.Frederick MD Cirrus vs Helicopter incident. So many times you have guys in the towers looking for you with binoculars. I think if done right, they could expand towers to a lot more fields that are uncontrolled which would be great for some but not all.
 
Next thing they will be building airports with NO control towers.:yikes::rofl:
 
I think it could be great if they had onsite radar for each airport. Not sure having a manned tower is as good as having good radar re.Frederick MD Cirrus vs Helicopter incident. So many times you have guys in the towers looking for you with binoculars. I think if done right, they could expand towers to a lot more fields that are uncontrolled which would be great for some but not all.
San Marcos has a radar feed from San Antonio (I think) but it's not certified so they can't legally use it (as in saying "radar contact") but they have it for situational awareness.

Next thing they will be building airports with NO control towers.:yikes::rofl:
The article discussed that a little saying that would be a capability of smaller airports in the future but it would be quite some time before it was ready for larger airports.
 
I just got back from Norway last night. They are experimenting with remote towers there and, according to an airline captain I talked to, it is going well and will probably be expanded. These are fairly low-traffic airports.
 
Plenty of control towers could do this. Heck, mine could easily, as could several others in my state. There are class Charleys that are quitter than the Delta down the street.
 
It's more cost effective, but will likely place a higher load on controllers in the remote location needing to watch the cameras. Maybe it will lead to more uncontrolled airports getting ATC coverage.
 
Are these cameras 3D? I've never seen a camera that is as good as plain 20/20 vision. Perhaps some of these airports just don't need a tower at all.
 
It's more cost effective, but will likely place a higher load on controllers in the remote location needing to watch the cameras. Maybe it will lead to more uncontrolled airports getting ATC coverage.

Or ADS-B IN receiver, post 2020.
 
Soon the machines are not going to need us.
 
I trained at KBDR, a non-Federal control tower/non-radar tower/Class D airspace. On a nice Saturday it was common to have 6 to 8 students in the pattern doing Crash and Go's.

I will be amazed if a remote operator can keep up with that safely while handling other towers, too.
 
Are these cameras 3D? I've never seen a camera that is as good as plain 20/20 vision. Perhaps some of these airports just don't need a tower at all.

Really? Are you comparing them to Polaroids?

The great thing about this system is, like the article stated, being able to use infrared and other technologies to assist the controllers. I particularly like the software they're testing which would detect any motion/pixel changes on movement area and bring it to the controllers attention.
 
It's more cost effective, but will likely place a higher load on controllers in the remote location needing to watch the cameras. Maybe it will lead to more uncontrolled airports getting ATC coverage.

They are talking about trying a remote tower at my airport KFNL. Plan to pick 3 or 4 busy GA airports and see if they can control from either Denver Center or Approach.
 
They are talking about trying a remote tower at my airport KFNL. Plan to pick 3 or 4 busy GA airports and see if they can control from either Denver Center or Approach.

They are?! Do you have any more info on this?
 
If I have real time traffic info in my cockpit AND I still have a human one radio call away if I need help, I'm open to this... but I need both those things or no way.
 
Really? Are you comparing them to Polaroids?

The great thing about this system is, like the article stated, being able to use infrared and other technologies to assist the controllers. I particularly like the software they're testing which would detect any motion/pixel changes on movement area and bring it to the controllers attention.

Why would you think I'm not serious? :dunno: it's more than just a matter of the reported resolution. It's a 3D world and some guy sitting at a government-issued computer with latency and a ****ty LCD monitor isn't going to have as good a view or as quick a scan across angles as the human eye connected to a human head. Can they change camera angle as fast as a turn of the head? React immediately to deviations when a student turns in front of someone on final?

Multispectral is nice, but is it 3D? Most IR images are relatively very low resolution and the ones that aren't are exceptionally expensive. I assume it will be combined with radar or ADS-B, but that's the easy stuff.

This seems like technology that would be fine for aircraft that are squawking and talking and able to be followed easily but humans are better for the one-off situations where aircraft deviate, go NORDO, or just screw up.

They think this will replace humans but I don't think it will be as good or as cheap as some of us are willing to imagine.

If they want to save money I dare say they should take a bigger ax to the number of near-idle towers out there. Trying to have one guy monitoring x number of towers is what they're going for and I just don't see the purpose.
 
Last edited:
They are?! Do you have any more info on this?

Not anything recent. Maybe now that there is commercial service again it will surface. Seems like quite a few of the commercial pilots don't like having to deal with us little guys. I think they need to find the budget first. There's a meeting Thursday, I'll try to ask if there more info. There are folks pushing for a tower and this was one idea for a much cheaper fix.
 
We're well along the path at KJYO. It's truly scary what they're saying. It won't be marked as towered on the charts. They're sure everyone will "know" The marketing folks are in charge.
 
... Trying to have one guy monitoring x number of towers is what they're going for and I just don't see the purpose.
Maybe some of that can happen, but there would be significant management benefits to simply having all of the individual tower controllers in the same physical facility. Scheduling, unplanned absences, training, morale, travel, ... It's a long list. I'd be very surprised if controlling ten airports with all the staff in one place wouldn't cut costs by half -- without having anyone handling more than one airport at at time.

And ... it would be upgrade an existing airport or to build a new one without ever building a brick and mortar tower.
 
Binocular depth perception is only useful out to about 10 meters, so not a consideration for this situation.

There is more to this than binocular vision. You have the capability to "focus" on a small area much better than any camera, and move it around much more rapidly.

Ever try to track a satellite with a telescope? Try it. Especially something low like ISS. You will be able to follow it beautifully with just your eyes. With a telescope or even a pair of binoculars, you might be able to see it better when it's in the field of view, but you'll lose it quickly and then see nothing.

Before you tell me that's contrived, ISS looks a lot like a passing airliner, just after sunset or just before sunrise.
 
Multispectral is nice, but is it 3D? Most IR images are relatively very low resolution and the ones that aren't are exceptionally expensive.

CAP tried this, and has largely (or maybe completely) abandoned it.

You would think a multispectral imager would be a godsend for search and rescue, but reality didn't conform.
 
... Ever try to track a satellite with a telescope?
Or a 172 while looking through a soda straw? So what?

I'd expect that a remoted tower would physically look like the training tower pictures that I've seen; basically large display panels instead of window glass. A mouse and a maginifier icon running a supplemental camera could substitute for the binoculars of a physically local controller. Supplemental IR image overlays could actually improve controllers' ability to see things in low viz and night conditions. Sort of a HUD for them. It would be easy, too, to have supplemental cameras to show places that could not be seen from the tower; "invisible" areas are not all that uncommon.

Going further, AI could detect and cue the controllers to movements on the screens that might otherwise go unnoticed -- a person or vehicle trespassing in an unsafe area for example. Lots of other potential for machine support to the controllers.
 
One would be surprised what a tower controller with years in a tower can perceive. Depth perception doesn't matter when you know what a Baron on a 5 mile final looks like. Three weeks ago I saw a Pilatus on radar skew to the wrong airport 7 miles away. Tower had eyes on while they tried to coral the pilot back.

Easy mistake for the pilot, human being used their mark I eyeball to prevent a spill out on our main airports final. Pretty good system as it is.
 
Do you actually work with AI?

It's not anywhere near as easy as you're assuming. Particularly, quantifying and predicting it's probability of detection.

It's a lot easier to say "AI" than to specify what it actually means. And those details make the difference between success and failure.
 
Or a 172 while looking through a soda straw? So what?

I'd expect that a remoted tower would physically look like the training tower pictures that I've seen; basically large display panels instead of window glass. A mouse and a maginifier icon running a supplemental camera could substitute for the binoculars of a physically local controller. Supplemental IR image overlays could actually improve controllers' ability to see things in low viz and night conditions. Sort of a HUD for them. It would be easy, too, to have supplemental cameras to show places that could not be seen from the tower; "invisible" areas are not all that uncommon.

Going further, AI could detect and cue the controllers to movements on the screens that might otherwise go unnoticed -- a person or vehicle trespassing in an unsafe area for example. Lots of other potential for machine support to the controllers.

I mean no disrespect, but your proposal would require every barely class D to have JFK levels of capabilities including some that don't yet exist. They can't afford to pay us let alone the probably 20 billion in improvements you're thinking about.

Great ideas though!
 
Do you actually work with AI? ...
Well, I designed my first image detection hardware about 30+ years ago, looking for tanks on an IR battlefield image, and spent about 20 years in high-tech. So yes, I think I understand what is involved.

Discriminating decoys from RVs during a midcourse ABM battle can be a bit difficult, but cuing motion with good POD in specific areas of a high resolution, fixed, and relatively featureless background is trivial. Actually, I wouldn't even call it AI.

From there it can get tougher, but there is a lot of potential once the tower windows become screens. You could do a lot without even getting close to the image processing that Google is doing in their driverless cars.
 
I mean no disrespect, but your proposal would require every barely class D to have JFK levels of capabilities including some that don't yet exist. They can't afford to pay us let alone the probably 20 billion in improvements you're thinking about.

Great ideas though!
I wasn't suggesting that all tower windows become monitors, for just the reason you say. Lots of reasons, not the least of which is cost, to prevent that from happening.

But I wonder if building a training-tower type cab with video screen windows would be all that expensive, given that the training cabs have already been developed and a lot of the hardware would be commercial OTS. I don't know much about them; maybe you do? And the sensor package for a basic remote tower could be fairly simple -- a few video cameras on a pole with decent redundancy. Radios, wx instruments, etc. are there anyway -- remoted or not. So, while a Star Wars video screen tower could get expensive (and fun!), it seems to me that a basic capability need not be.
 
Having been inside Future Flight Central several times, it's not at all like looking out a window. Basically, you can see aircraft on short final and on the ground at a large airport like LAX, plus a few landing lights.

You can see A LOT more from a real tower.

Dynamic range is a problem for imaging displays.
 
I say beta test it with manned cabs for a few years till confidence is achieved...

Anything less is suicide for the pilots...

Controllers don't care, they never get killed by mid air collisions...
 
Last edited:
Coming soon to an uncontrolled (but monitered) field near you! Now you won't need an operating control tower to charge User Fees for your County Airport; we can do it for you...:mad2:
 
San Marcos has a radar feed from San Antonio (I think) but it's not certified so they can't legally use it (as in saying "radar contact") but they have it for situational awareness.

A lot of Class D towers, especially those with nearby radar approach controls, may have a radar display in the tower. They are VFR towers, not qualified radar controllers. The display is for SA on inbounds coming from approach and departures leaving their airspace. They can use it to advise, but cannot control based on it.
 
I'm not seeing these things being used to replace controllers to the point of "running the pattern." You're number two, follow him, extend downwind, hold short, cleared for take off etc. Not in the near future anyway. I can see a value at some airports though. Like at currently uncontrolled Airports where Instrument Approaches are being conducted, especially at airports without a Surface Area. Detecting aircraft flying around Clear of Clouds under the bases as an approach is breaking out of those bases is an item of concern now that with RNAV there are so many Approaches to just about everywhere. Even with a Surface Area, where supposedly no one is going to be operating beneath a Ceiling less than 1000' there are still issues. Detecting and preventing "Runway incursions" is where it could really shine. No Radio Guys, sorry, but it would establish the communication requirement.
 
Back
Top