Air Traffic Control | Hacker News Reader

Air Traffic Control

computer.rip

264 points1317115 comments

Kind of curious how air traffic control evolved in other countries, and how the international flight system works with handoffs between countries, particularly in Europe and the Caribbean where national borders tend to be relatively small.

This was indeed normalized very early on in post WWII european construction:

https://en.m.wikipedia.org/wiki/Eurocontrol

In a nutshell it goes like this: you file a flight plan to go from A in country A to B in country B. Every country that you cross will know about this.

Depending if you fly IFR or VFR and at what altitude, you will talk to either a flight information service (FIS), Radar or Area Control Center (ACC).

Hand-off is usually done a few miles from the border of the next country you're going to pass. You tell them who you are and where you're going to. They can let you fly as planned or give you another route or altitude.

From a pilot's point of view is pretty simple and straightforward.

How the flights plans get routed is also of interest, they typically go through AFTN: https://en.wikipedia.org/wiki/Aeronautical_Fixed_Telecommuni...

How were the flights filed in the pre-internet era?

There is standard form that you would fill out at the Briefing office and they would fax it to all relevant parties.

Before that, I have no clue if something existed or you would just ring up the destination aerodrome to let them know you're coming.

I’ve filed a VFR plan by phone (more than a decade ago), don’t know for sure if it works the same for IFR but I’d imagine so. You’d call up and basically read out the form fields in one long monologue.

Internationally, the concept of regulated airspace began taking shape after World War I. Under the Treaty of Versailles, the International Commission for Air Navigation (ICAN) was created in 1919, developing the first air traffic regulatory framework initially signed by 19 states.

There is cooperation between states, with the best example I know of being the Maastricht Upper Area Control Centre (MUAC) which was established in 1972 by Eurocontrol and manages the upper airspace (above 24,500 feet) over Belgium, Luxembourg, the Netherlands, and northwest Germany.

Eurocontrol is the first and, to the best of my knowledge, the only successful attempt to pool controllers between countries in Europe.

ICAO exists for exactly this.

https://en.wikipedia.org/wiki/International_Civil_Aviation_O...

Re WWII use of radio:

My grandfather flew Typhoons, and they operated 'cab rank' as part of the 2nd Tactical Air Force. The Army would radio in coordinates of German tanks or fortified positions, and the Typhoons would come in with their rockets / bombs / cannons. I wish he was still around so I could ask him how that was done. A central dispatcher? Or did they talk to the Army directly? Not sure.

Forward air control would get a map grid or other description of the target from Infantry and Aircraft would be dispatched to the target . https://en.wikipedia.org/wiki/Forward_air_control_operations...

And before GPS they'd reckon their position off of landmarks and terrain?

Yup, and there were apparently fake towns set up with lights that the German night bombers would hit instead of the real populated areas:

https://en.wikipedia.org/wiki/Starfish_site

There was a lot of development around radio navigation aids in the interwar period. I believe it was even possible to navigate using commercial AM radio broadcasts, as long as the location of the broadcasting station was known. So while the basic method was computing location based on speed and compass bearing from a landmark (possibly accounting for wind drift), I believe a certain level of radio navigation was still going on in cases of adverse weather.

Longer-range bomber flights also did it the same way ocean-faring ships have done in the past few hundred years: they had a roof window for taking celestial fixes with a sextant. Come to think of it, that's how the Apollo capsule also confirmed its location.

There's a lot to read about this. This is probably a good start since it branches to many other related articles: https://en.wikipedia.org/wiki/Oboe_(navigation)

> Come to think of it, that's how the Apollo capsule also confirmed its location.

Isn't that still how it's done? - When leaving earth orbit there isn't GPS or such, the guidance to the moon or further still uses star trackers.

The Voyager space probes for sure use visual fixes on guide stars, such as Alpha Centauri (and our own Sun?) to ensure correct orientation. Of course, they don't need to navigate much, since given the relative velocities, their thrusters wouldn't make much difference.

And there is a quite-famous scene in Apollo 13 where they literally line-up the Moon itself in the reticule so that their engine burn puts them on exactly the correct trajectory.

I believe star maps are also baked into ICBMs

Some SLBMs (Submarine Launched Ballistic Missiles) use star sighting to correct their trajectory mid-flight. ICBMs launched from silos know exactly where they're starting from, whereas a submarine is both moving, at different depths and has some error to knowing exactly where it is. Missiles use inertial guidance, so knowing your starting your point is crucial.

How accurate a missile needs to be is a whole other dimension though. If the value of a missile is as retaliation to destroy a city (countervalue) then it can be a large warhead and "miss" by quite a margin but needs some form of credible survivability of an enemy first strike. If the missile is to be used to destroy enemy military installations (counterforce) then it needs to be a lot more accurate but usually the implication is as a first strike so less survivability is required.

When you have nuclear weapons that you can drop from manned aircraft, ICBMs from silos and SLBMs controlled by different military branches there is going to be a lot of politics over what the missile is for, which will determine its required accuracy, which will be a factor on if it needs star sighting.

ICBMs early on applied star tracking to increase precision, starting with pretty simple analog systems (I recall something about shifting a tape with simulated tracker signal to a position matching the launch time) to modern digital map systems.

Rad labs volume 2 has a great description of the radio navigation/positioning systems in use at the time: https://archive.org/details/mitradiationlabo0002john/mode/1u...

I believe several of these were also used by civilian aviation before and after the war

TL;DR: both sides used systems with multiple ground antennas that allowed pilots to essentially triangulate their position or at least know they're in a given lane.

I understand that the germans used radio navigation to guide their nighttime bombers over Britain. Wikipedia suggests that the British quicly countered this.

One thing i find interesting is the fact that by sending beacons ahead of the bombing party, they literally broadcasted their intentions to the enemy but that didn't seem to worry them too much.

There were a great number of inventive aerial navigation systems that predate GPS by many decades. One of my favourite books is Most Secret War[0], a biography by R.V. Jones that details much of the cat and mouse that went into detecting and countering the various systems used by the Axis and the Allies to bomb each other.

[0] https://www.goodreads.com/book/show/1127842.Most_Secret_War

Armed forces have been using variations of grid maps to plan bombardment since time immemorial. Do you really think that before GPS people just walked around lost all of the time?

https://en.wikipedia.org/wiki/Ordnance_Survey_National_Grid

https://en.wikipedia.org/wiki/Military_Grid_Reference_System

Dead Reckoning - Pilots have been using this forever (100+ years). It's just maths: speed + heading + time + wind = position. Still widely used in VFR flying! It was made more accessible with the E6B "computer" from the 1930s, but isn't a computer at all - just a circular slide rule with aviation formulas built in and still used today. Every student pilot still learns this stuff.

Non-Directional Beacon (NDB) / Automatic Direction Finder (ADF) - Introduced 1930s-40s, standard in WWII. Ground station sends signal, aircraft needle points to it. Pretty simple but not super accurate. Still technically operational but barely used anymore. PPLs still learn it in training but honestly who uses this regularly? It's like knowing how to use a fax machine - technically still around but why would you?

Very High Frequency Omnidirectional Range (VOR) & Distance Measuring Equipment (DME) - These work together - VOR gives you direction from a station, DME tells you distance. Developed in the 1940s, and surprisingly still pretty common. Lots of GA pilots use these as primary or backup nav. Created all those "highways in the sky" that formed the airways system. VOR is definitely the most useful of the old-school nav aids, and you'll find these receivers in most cockpits.

Inertial Navigation (INS) - Commercial airliners got these in the 1960s. No external signals needed - just measures acceleration to track position changes from your starting point. Completely self-contained, works anywhere on Earth. Modern versions use laser gyros instead of mechanical parts. The drawback is drift - errors add up over time (1-2 miles per hour). Still used alongside GPS on long-haul flights.

US Postal Arrows: Also, i've not seen them myself - but the US in the 1920s before radio navigation, the US built a system of MASSIVE concrete arrows (up to 70 feet long!) across the country to guide airmail pilots. This was officially called the "Transcontinental Airway System." They were painted bright yellow and paired with beacons/lighthouses for night flying. The system eventually stretched from NY to SF.

Most got dismantled during WWII, but dozens of these concrete arrows still exist today scattered across the western US. It was literally a "follow the yellow brick road" situation for pilots. They'd fly from arrow to arrow during daylight, and at night there were gas powered lights showing them. More info here, https://www.core77.com/posts/25236/what-are-these-giant-conc...

Until you mentioned the E6B I totally had forgotten that half a lifetime ago, as a student pilot I programmed my Radio Shack Pocket PC to do some of the E6B calculations.

It worked, but the usability of the tiny keyboard was crap in a bouncing, vibrating Cessna 152, forcing me to look down at my lap while flying. My instructor once said "you're going to kill yourself using that thing" so I eventually just gave up and bought the electronic E6B.

That's awesome. I found 2 x E6B's in my garage last week after a clean out.

Do you still have the code? I'd love to see how you wrote it.

Try using SkyDemon or ForeFlight on an unmounted mobile phone, checking your way point and dropping it under your seat, whilst flying in congested airspace. It's a mistake you only make once :o

No, that was a looong time ago. I actually gave the Pocket Computer away a while back for someone's "computer museum."

In addition to what others have said, letting artillery mark targets with colored smoke and simply flying to a grid location and keeping an eye out for these signals also works well.

As an aside, wasn't there some HN post a few months back about how carrier warfare in the Pacific led to more and more complex schemes of ATC, aka "Fighter Directors", including early Naval computers devoted to figuring out where to put airplanes in the right place at the right time right time?

I read in A Bridge Too Far, the Hawker Typhoons that attacked German positions at the start of the Operation (called in by the Forward Air Comtroller), flew under the opening barrage and one Typhoon disintegrated as it was hit by a shell. I've often wondered if there was any system in place to prevent that sort of thing.

I also read there was some difficulty in target designation as the pilots had different maps, so the grid coordinates were off.

Hitting an aircraft in flight was hard enough if you were aiming at it, I doubt they felt any need to take countermeasures against such a rare occurrence.

As others have mentioned, in WWII, many of the things modern militaries take for granted were in their infancy. In modern times, there are lots of staff meetings and procedures involving nothing other than deconflicting fires. Fast jets hit this, artillery hits that, fighters cover here, surface-to-air assets cover there, etc.

But those lessons are written in blood due to exactly incidents like the one you mention. The "big sky, little airplane" theory is not something to bet your life on.

The general lack of front line radios makes me think he wouldn't generally have been talking directly to the troops on the ground but the entire field of air combat much less close air support was in it's infancy so a lot of the doctrine was being made up on the fly. By the end of the war they had made a lot of improvements including bringing control of CAS closer to the frontline units that would be calling it in.

More about the US side of the conflict:

https://www.reddit.com/r/WarCollege/comments/jfrpiu/how_did_...

Not Europe (assuming from your grandfather flying Typhoons) but here's a paper about how the Marines in the Pacific were developing the idea of close air support in the Pacific: https://apps.dtic.mil/sti/tr/pdf/ADA369287.pdf

A random book I found that might have more specifics about the RAF and British use of CAS: https://crecy.co.uk/product/close-call-raf-close-air-support...

Close air support is the phrase you're looking for for your own searches though if he was bombing tanks in the field.

An interesting parallel to the Air Force's SAGE was the Navy's NTDS, Naval Tactical Data System. It had to do much the same as SAGE, with target correlation and the like, and had to do it from moving ships (and later with the E-2 Hawkeye Airborne Tactical Data System from planes as well). Purely military, and I can't tell that many of its developments ever ended up in civilian society like elements of SAGE did, but it's remarkable just how much they could do with the technology of the time.

The best source I've seen about NTDS itself is https://ethw.org/First-Hand:No_Damned_Computer_is_Going_to_T... while Norman Friedman's book _Fighters Over the Fleet_ talks about the fighter control context around NTDS, like what the system was before NTDS, the factors that caused the previous system to break down, and parallel British efforts at the same problem

Hi author, I love your blog, but could you add a toggle option to disable the background image? I end up going to Chrome Devtools and disabling the CSS rule.

I suppose the PDF is a good workaround for now.

The RSS feed is another good option!

That’s true, but I need to set aside time to clean up my RSS reader (too many feeds!).

This piece hit hard. It’s strange how something so invisible to most of us is simultaneously one of the most complex, tightly-coupled systems running in the background of daily life.

But is it that complex? Airplanes arrive in your sector, you write the number and destination on a slip of paper, and put it on your board. You tell the plane, if needed, where to fly, and then you tell it to talk to the next sector controller when it leaves your area.

It seems rather non-complex, and I think it has to be, so it can be robust and offer room for errors etc.

A lot of the complexity is the sheer volume of air traffic. Not so much at cruise, but as planes transit through populated areas. ATC over NorCal is a mess of overlapping zones, ~dozen airports, and hundreds of airplanes at any given time of day. NYC metro is similar.

Everything is simple at the highest levels of abstraction.

It is the details of actually making it work that raise the complexity levels and/or kill you if you don't get them right.

Remember: a decision that in the abstract with infinite thinking time is easy can be extremely stressful when it has to be made in seconds in real time and getting it slightly wrong will potentially cause a catastrophe. And ATC has to keep doing this throughout their entire work shift.

Standing joke in aviation: what's the same between a pilot and a controller?

If they screw up, the pilot dies.

The historical perspective is kind of interesting. What interests me more is how this stuff can be modernized. Because the current byzantine system of task saturating people by making them yell numbers and letters at each other over state of the art 1950s technology for doing so is hardly efficient, scalable, or ultimately secure when it inevitably breaks down under the stresses it induces on pilots and staff involved.

Communication failures and pilot errors are usually major contributing factors to accidents. Pilot errors become more likely when individuals are preoccupied with processing vast amounts of information and demands thrown at them via radio in critical parts of flight (e.g. when finding their way to a runway).

Most of the radio calls today (>95%) are bog standard verbal exchanges about things that should not require any humans in the loop. Anything to with weather is just numbers that can and should be be exchanged digitally. Much faster, less room for error. There should be zero confusion about what the altimeter settings are at a particular airport, what the active runway is, where the wind is coming from, etc.

Anything to do with "who are you and why are you here", which is a disturbingly large proportion of verbal exchanges, sounds like it could be established both more securely, robustly, and efficiently. We have computers now with things like secure hashes, uuids, certificates, etc. Any time you have enough bandwidth to talk to somebody you definitely have enough bandwidth to throw quite a bit of data around securely and reliably.

If you call your granny via whatsapp, facetime, or whatever, there's no need to tell her who you are. Because the app tells her before she even answers. It's completely redundant information. She already knows. There's no valid technical reason why ATC cannot have the same comfort of knowing who they are talking to and what they are flying. It's 2025! Not 1965. ATC should have full context when they talk to people. They shouldn't have to ask for that context verbally. Routine course changes, altitude changes, etc. could be communicated and confirmed via computers. Voice channels should be used for emergencies and non standard situations only.

There are a few things happening in aviation that are making this more urgent after many decades of stagnation in technological changes. Battery electric is going to make flying vastly cheaper and safer than it is today. That's going to increase the amount of relatively inexperienced pilots and plane movements. And secondly, there are an increasing amount of autonomously flying planes, drones, etc. Those are actually going to dominate traffic in the years ahead. Pilots are expensive and are becoming kind of redundant. The amount of flights that ATC needs to juggle is going to increase by at lease one or two orders of magnitude.

The current system won't scale with that, it will have to change. Now would be a good time to start figuring that out.

I take your point but simply throwing silicone at the problem doesn't make things all of a sudden better. Wind data can already be uplinked to aircraft through ACARS anyway. The problem is aviation is very dynamic. You can't fly the plane and get a bing bong message cleared to land and not be distracted by it. Voice comms are flawed but one of the benefits is you can hear them and monitor the aircraft state. ATC have systems which add another layer of protection for a misidentified runway for example.

This is a very good summary of the situation.

We have digital datalink weather, pilots love it and use it frequently. It would be nice to have D-ATIS at smaller airports, but otherwise weather is a solved problem.

For everything else, voice communications are a feature. Pilots are trained to spend most of their time heads-up looking out the window, not heads-down looking at computer screens. ATC via text message breaks that flow. (We do actually have ATC via text message, using CPDLC. It's most useful on oceanic flights as an alternative to HF radio, where things happen slowly. Once you're getting close to the destination and events speed up, it's back to voice.)

In theory the biggest benefit to CPDLC is uploading a clearance into an FMS automatically, which is already a thing... but even then you have to closely supervise and occasionally step in when it does it incorrectly.

> If you call your granny via whatsapp

Remember flight MH370? The industry doesn't want obviously good solutions if it costs more than $1 per passenger. There is room to automate part of this, and have ATC and pilot human oversight with manual contingencies. But no one is going to pay for it, not through normal or usual and customary processes. Most of the major US stakeholders (airlines) are poor.

I remember lots of accidents where there was miscommunication, task saturated pilots making mistakes, etc. That's not an excuse to sit on our hands and decide that aviation safety peaked decades ago and that there's no need to modernize.

Also, this is bigger than the US. Aviation is a world wide business. I don't think the Chinese are looking at the FAA for learning how to make drones that they produce interface with Chinese ATC. This might be a case where the US will follow instead of lead.

> don't think the Chinese are looking at the FAA for learning how to make drones that they produce interface with Chinese ATC. This might be a case where the US will follow instead of lead.

The Chinese use systems bought from European companies, made to the same standards as used in EU. I'm working on one of such systems.

If they cared about cost at all, they'd be serving pre-brewed coffee and tea. The aviation standard issue coffee machines are incredibly old, wasteful, and stupid. It works so it stays.

The reason why aviation isn't accepting new tech, I think, is because tech industry, especially Web, hasn't earned enough trust with real-time high trust systems. The computer used in the Apollo 11 lander had BSoD multiple times and still landed men on the Moon and back.

You can buy a car on virtual credit cards from an iPhone, or view charts and maps for pilots on an iPad, but you can't call airport tower on Discord and ask for landing clearance - I think that's the threshold line up to which aviators trust Web-related techs.

By the way, Armed Forces of Ukraine do use Discord for coordinating military operations and requesting artillery barrages. Risks of using untrustworthy technology far outweigh costs of not using it, to them. Maybe there are some clues there?

> Risks of using untrustworthy technology far outweigh costs of not using it, to them. Maybe there are some clues there?

I hope your clue is not to start a global war to change the baseline for safety and make FAA stop being so anal about certification and allow new coffee machines

Related: United serves instant decaf on most flights and it’s glorious. I too wish they did all coffee as instant, especially with how good some instant is now.

> Most of the radio calls today (>95%) are bog standard verbal exchanges about things that should not require any humans in the loop. Anything to with weather is just numbers that can and should be be exchanged digitally. Much faster, less room for error. There should be zero confusion about what the altimeter settings are at a particular airport, what the active runway is, where the wind is coming from, etc.

> Anything to do with "who are you and why are you here", which is a disturbingly large proportion of verbal exchanges, sounds like it could be established both more securely, robustly, and efficiently. We have computers now with things like secure hashes, uuids, certificates, etc. Any time you have enough bandwidth to talk to somebody you definitely have enough bandwidth to throw quite a bit of data around securely and reliably.

The problem is, for that to work, you'd need to push a worldwide coordinated effort to upgrade everyone. Hundreds of thousands of GA aircraft, historical aircraft, about 30k commercial aircraft and most likely hundreds if not thousands of different models. For each of these you'd need to develop and most importantly certify a digital avionics package. And you'd need to train about half a million pilots for commercial flight, on top of that all the GA pilots.

And no, "give pilots an iPad" doesn't count. That's what many airlines are already doing so pilots have an alternative to paper charts of approaches or on-ground navigation... but the old paper and radio? That all still has gotta be present because that stuff is actually certified to be redundant and (largely) fail-proof.

There's a reason innovation speed is very slow in aviation: it took decades worth of work and many thousands of deaths to get aviation to be what it is: the most safe way of transportation by far. You probably have a higher chance of getting hit by a drunk driver than to even get a small injury on a commercial airplane, because all the failure modes of airplanes and individual parts have been worked out. Every kind of new part, new systems and new material introduces new failure modes that have to be discovered, mitigated and the mitigation be distributed around the world.

You are describing why something that hasn't changed for decades won't change for decades. Because nothing ever changes. All of that stops being true when change is forced. My closing argument provides that change vector: a couple of orders of magnitude increase in flight movements and autonomous flight.

From a technical and practical point of view this stuff isn't actually hard: you start with the big airlines. Maintaining and refurbishing airplanes is something they spend lots of money on. Hundreds of thousands per plane per year. So, there's plenty of budget and a little bit of new equipment won't really move the needle.

Just look at how quickly ipads got adopted by pilots. Many pilots won't fly without one now. Reason: they are good and they provide redundancy when the multi million thing in front of them has an electrical failure. It provides GPS, can talk to ADSB via bluetooth. You can have an artificial horizon on it, etc. Pilots won't use them as a primary instrument (that would be illegal) but they are well capable as an emergency replacement with the right apps. This industry can move fast when it needs to. Ipads are cheap. They provide convenience and safety. So pretty much any pilot will want one in their plane.

The issue isn't technical but bureaucratic. If ipads had to have gone through some FAA controlled design committee, they would never have gotten this popular. But they didn't. The history of this stuff is that airlines got organized first before the FAA took over. They'll have to force these changes.

> From a technical and practical point of view this stuff isn't actually hard: you start with the big airlines.

You still need to get any new integration certified, and not just on the plane side, but also on the ground. And it's ... problematic because you would then have two wildly different systems in parallel, and pilots would need to switch between radio and digital all the time. Too much potential for chaos.

  The current system won't scale with that, it will have to change.
  Now would be a good time to start figuring that out.

Modernization has been in the works for years now. Chronic underfunding means that nothing much has come from it.

  Anything to with weather is just numbers that can and should be be
  exchanged digitally.

It often is, via ACARS.

  It's completely redundant information.

Redundancy is the point.

From what you said I take it you're not a pilot.

> Routine course changes, altitude changes, etc. could be communicated and confirmed via computers. Voice channels should be used for emergencies and non standard situations only.

Do you think when planes get closer to the airport, pilots want to keep their eyes on the group chat while trying to land? :)

A few comments seem to make the assumption that "via computers" implies screens, but that assumption is not valid; computers can talk.

I agree with the thrust of the pushback though. I think it seems like there would need to at least be a human voice fallback, and then you'd want to exercise that pretty often, because having an emergency fallback that is never exercised is a recipe for disaster.

> A few comments seem to make the assumption that "via computers" implies screens, but that assumption is not valid; computers can talk.

I don't think any pilot likes having a computer blab all the time when they need to focus on a landing for example.

What makes a computer blabbing all the time different from an ATC blabbing all the time? Just because a computer can speak some things doesn't mean it has to speak all the things.

It doesn't know what's important and what not at a moment's notice.

ACARS and CPDLC handle a lot of this. A change of altitude/flight level can be sent from ATC, accepted by a pilot, and updated in the autopilot without anyone speaking over VHF.

Reply on news.ycombinator.com