Air_traffic_control - Wikiparency

# Air Traffic Control > [[PORTAL_Aviation|Aviation]] · [[PORTAL_Avionics|Avionics]] spine.  ## Microsims — three.js ### Air Traffic Control (three.js) <div class="microsim-player"> <iframe src="https://wikitube-3d-microsims.netlify.app/Air_traffic_control.html" width="100%" height="620" frameborder="0" loading="lazy" sandbox="allow-scripts allow-same-origin" title="Air Traffic Control — three.js microsim"></iframe> </div> **Open it full-screen:** [Air_traffic_control.html](https://wikitube-3d-microsims.netlify.app/Air_traffic_control.html) · library `threejs` · route `microsim/threejs/` ### Related microsims Live sims on neighbouring articles: - [[Radar]] - [[Instrument_flight_rules]] - [[Aircraft_flight_dynamics]] - [[Avionics]] - [[Global_Positioning_System]] - [[Helicopter]] *Sim hosted off-article; the article owns the reference, not the runtime (WIKI_RULES §10.4).*  ## Overview **Air traffic control** (ATC) is a ground-based service that keeps aircraft apart. ICAO Annex 11 states its objectives in a specific order: prevent collisions between aircraft; prevent collisions between aircraft and obstructions on the manoeuvring area; expedite and maintain an orderly flow of traffic; provide advice and information useful for the safe and efficient conduct of flights; and notify search and rescue when an aircraft needs help. The first two are about geometry. The third is about throughput. The whole discipline lives in the tension between them. It is worth being precise about what the service is *not*. A controller does not fly the aeroplane and does not command it in the military sense. A controller issues **clearances** — permissions to occupy a piece of airspace on stated conditions — and **instructions** that the crew is expected to comply with unless compliance is unsafe. The **pilot in command retains final authority**. ICAO Annex 2, paragraph 2.4, says that the pilot-in-command "shall have final authority as to the disposition of the aircraft while in command"; 14 CFR 91.3(a) in the United States says the pilot in command "is directly responsible for, and is the final authority as to, the operation of that aircraft", and 91.3(b) permits deviation from any rule to the extent required to meet an in-flight emergency. A clearance is not a guarantee that the manoeuvre is safe, and it never overrides the commander's judgement. Nor is separation provided to everybody everywhere. Which flights are separated from which depends on the class of airspace. In Class A only IFR flights are permitted and all are separated from each other. Lower down, in Class E, IFR flights are separated from other IFR flights and merely receive traffic information about VFR flights, who are separating themselves by looking out of the window. In Class G there is no control service at all. Terrain and weather are also, in general, the crew's problem: the controller has minimum vectoring altitudes and ground-proximity alerting to work with, but "cleared direct" is not a promise that the direct track is clear of a mountain. What the controller actually manipulates is a small set of variables — heading, level and speed, plus the routing itself — applied to a picture that is not the world but a processed surveillance track. That picture lags. A rotating en-route radar delivers a fresh position every ten to twelve seconds; a terminal radar every four or five; ADS-B and multilateration update roughly every second. Altitude arrives quantised: classic Mode C reports in 100 ft steps, Mode S in 25 ft steps. The separation minima in force encode the quality of that picture. Five miles is not a physical constant. It is a number chosen so that the true positions remain safely apart given how imperfectly the displayed positions are known. ## The physics ### The protected volume Give every aircraft an invisible bubble and the rules become geometry. The bubble is a **cylinder**, not a sphere, because the horizontal and vertical minima are different numbers and are tested independently. Separation between a pair is lost only when *both* conditions fail at once: $ d_h < S_h \quad \text{AND} \quad d_v < S_v $ where $d_h$ is horizontal distance, $d_v$ is the difference in altitude, and $S_h$, $S_v$ are the minima in force. Equivalently, and more usefully for drawing: give each aircraft a cylinder of radius $S_h/2$ and half-height $S_v/2$, and separation is lost exactly when the two cylinders intersect. Two aircraft at precisely the minima have bubbles that touch. That is the convention the simulation uses. The shape of that cylinder is the whole lesson. At the en-route radar minimum of 5 NM and 1000 ft, the protected volume is 30,381 ft across and 1000 ft deep — an aspect ratio of about **30 to 1**. It is a dinner plate. At the terminal minimum of 3 NM it is still 18 to 1. Set the simulation's vertical exaggeration slider to 1 and the picture stops flattering the vertical axis: the shelves of flight levels collapse into a slab and the protected volumes become discs. Two consequences follow. First, **the vertical dimension does most of the work.** A thousand feet is 0.165 NM. Buying 1000 ft of vertical separation costs a fraction of what buying 5 NM of horizontal separation costs, and it can be bought while both aircraft keep their tracks and their planned routes. Two airliners can pass with their radar returns nearly touching — one or two miles apart laterally — and be perfectly legal, because they are on adjacent flight levels. This is not a loophole; it is the design. Second, **the airspace is emptier than it looks.** A protected volume of radius 2.5 NM and half-height 500 ft has a volume of $\pi r^2 h = 3.2$ cubic nautical miles. A sector 100 NM square spanning FL290 to FL410 contains about 19,750 cubic nautical miles. Pure packing would allow thousands of aircraft. Real en-route sectors handle ten to twenty. The binding constraint is not instantaneous geometry at all: it is the requirement that separation be maintained *into the future*, and the finite capacity of one human being to hold that future in mind. Sector capacity is a workload number wearing a geometry costume. ### Why altitude is the precise axis Vertical separation is cheap because it is *well measured in the relative sense*. Above the transition altitude every aircraft sets its altimeter to the standard pressure datum of 1013.25 hPa and flies a **flight level**, not a height. Two aircraft on FL350 and FL360 may both be some hundreds of feet away from their true geometric altitude, but they are wrong in nearly the same direction and by nearly the same amount, because they are sampling the same atmosphere with the same rule. The difference between them is far better known than either absolute value. Reduced Vertical Separation Minimum (RVSM) is the engineering programme that made this precision auditable. To operate between FL290 and FL410 an aircraft must have two independent altitude measurement systems, an automatic altitude control system able to hold the selected level within about ±65 ft in straight and level flight, and an altitude alerting system that warns at about ±300 ft. Height-monitoring units measure the total vertical error of the fleet in service against a geometric reference, and the whole scheme is justified against an agreed target level of safety — an ICAO figure of 5×10⁻⁹ fatal accidents per flight hour attributable to the vertical dimension. The payoff is the cleanest capacity story in aviation. Between FL290 and FL410 the old minimum of 2000 ft gave seven usable cruising levels. Under RVSM, at 1000 ft, it gives thirteen — **six additional flight levels**, produced without a metre of concrete or a square mile of new airspace. The simulation's vertical-minimum control does exactly this: switch it and count the shelves. Aircraft parked on a level that has just ceased to exist are commanded to the nearest legal one, which is a compressed version of what happened, in reverse, on each implementation night. RVSM entered service in the North Atlantic in March 1997 (initially between FL330 and FL370), in European airspace on 24 January 2002, and in United States domestic airspace on 20 January 2005. Above FL410 the minimum reverts to 2000 ft. Which levels a given aircraft may use is set by the **semicircular rule** of ICAO Annex 2, Appendix 3, which allocates levels by magnetic track. Tracks of 000° to 179° take one column of the table and tracks of 180° to 359° the other. Under a 2000 ft minimum the direction alternates every 2000 ft — FL290 eastbound, FL310 westbound, FL330 eastbound, and so on. Under RVSM it alternates every 1000 ft. The microsim colours its shelves accordingly: cool for the eastbound column, warm for the westbound one. (It uses true track; magnetic variation is not modelled.) ### Detecting the conflict Take the relative geometry of a pair. Let $\mathbf{r}_0$ be the horizontal position of B relative to A, $\mathbf{v}$ their relative ground velocity, $z_0$ the difference in altitude and $w$ the difference in vertical speed. Assuming both keep their present track, level and speed, $ \mathbf{r}(t) = \mathbf{r}_0 + \mathbf{v}t, \qquad z(t) = z_0 + wt $ The horizontal condition $|\mathbf{r}(t)| < S_h$ is a quadratic in $t$: $ |\mathbf{v}|^2 t^2 + 2(\mathbf{r}_0\!\cdot\!\mathbf{v})\,t + \left(|\mathbf{r}_0|^2 - S_h^2\right) = 0 $ whose two roots, when real, bracket the interval during which the pair are horizontally inside the minimum. The vertical condition $|z(t)| < S_v$ is linear and brackets its own interval. A **conflict** exists if those two intervals overlap, and if the overlap falls inside the look-ahead horizon. The first instant of the overlap is the predicted time of loss of separation. This is exact, cheap, and it is what the simulation draws: at that instant it renders the two protected volumes as ghosts, interpenetrating, joined by a line. Setting the derivative of $|\mathbf{r}(t)|^2$ to zero gives the horizontal **closest point of approach**: $ t_{\text{CPA}} = -\,\frac{\mathbf{r}_0 \cdot \mathbf{v}}{|\mathbf{v}|^2} $ a quantity every controller estimates by eye a hundred times a shift and every ground conflict-alert system computes explicitly. Its sign carries the news: negative means the pair are already diverging. Real systems run this in two bands. **Short-term conflict alert** (STCA) looks perhaps two minutes ahead and is deliberately tactical: a safety net, tuned so its alerts are rare enough to be believed. **Medium-term conflict detection** (MTCD) looks eight to twenty minutes ahead against the planned trajectory, and is a planning aid rather than an alarm. The look-ahead slider sweeps across both regimes, and the trade is immediately visible: a short horizon misses problems you could have solved cheaply, and a long horizon fills the screen with conflicts the traffic would have resolved by itself. The linear assumption is the weak point in both, and honestly so — it ignores turns, and it ignores wind. ### Three tools, three currencies Once a conflict is found, the controller has three tactical instruments. They are not interchangeable, and the simulation bills each of them in its own units. **A vector** — a heading change — pays in **track miles**. It is the most powerful tool and the most flexible, but it is not instantaneous. An airliner at cruise is bank-angle limited, typically to about 25°. Rate of turn is $\omega = g\tan\phi / V$, which in the usual mixed units is $1091\tan\phi / V_{\text{kt}}$: at 470 kt and 25° of bank, about 1.08° per second, so a 30° turn takes nearly half a minute before the aeroplane is even pointing the new way. Worse, the turn radius $r = V^2/(g\tan\phi)$ is about **6.9 NM** — larger than the entire 5 NM en-route separation minimum. A late vector cannot deliver. Every mile flown off the direct track is a mile of fuel and a minute of somebody's day, and the aircraft must eventually be turned back. **A level change** pays in **feet and kilograms**, and in a scarce resource. The energy argument is simple: lifting mass $m$ through height $\Delta h$ costs $mg\Delta h$ of potential energy, and a modern turbofan at cruise converts fuel into useful work at an overall efficiency around 0.35 with a fuel of roughly 43 MJ/kg. For a 70 tonne twinjet and a 1000 ft step that is about 14 kg of fuel — a floor, since it ignores the extra drag during the climb. A descent costs nothing in that accounting, which is misleading: the aircraft then cruises below its optimum level, where specific range is worse, so the bill arrives later. And a level, unlike a heading, is a shared resource. Putting one aircraft on FL360 takes FL360 away from everyone else in that piece of sky. **A speed instruction** pays in **delay**, and it is the weakest of the three. Above FL250 speed control is issued as a Mach number, and the usable authority is small — perhaps 0.02 to 0.04 Mach either side of the planned cruise Mach before the flight falls off its optimum. At high level the speed of sound is fixed at 573.6 kt by the isothermal stratosphere, so 0.03 Mach is about 17 kt: it opens in-trail spacing at roughly 17 NM per hour of flight, and costs the slowed aircraft around two minutes over a 400 NM leg. Applied to two aircraft in trail on the same route, this is elegant and almost free of side effects. Applied to a crossing conflict it is close to useless, because slowing an aircraft moves the crossing time without moving the crossing point. The simulation makes that failure visible: choose the speed tool on a crossing pair and it will tell you that the best available speed reduction does not clear the conflict. Underneath all three, the aircraft model is real rather than kinematic decoration. True airspeed is derived from Mach number through the International Standard Atmosphere — temperature falling at 6.5 K/km to the tropopause at 11,000 m and constant at 216.65 K above it, with $a = \sqrt{\gamma R T}$ — so M0.80 at FL350 gives 461 kt and the same Mach at FL390 gives 459 kt. Turns are bank-limited. Level changes run at a fixed rate with a capture. Speed changes are limited by a cruise thrust margin of 0.15 m/s², which puts a 0.03 Mach reduction at about a minute. ## Controls -> what each maps to | Control | Symbol | Range and units | What it does | | --- | --- | --- | --- | | Aircraft | *N* | 2 – 24 (count) | Number of aircraft simultaneously inside the block. Together with density it fixes the sector size. | | Traffic density | *ρ* | 2 – 40 aircraft / 10,000 NM² | Sets the sector side length as *L* = √(10⁴·*N*/*ρ*). Low values are oceanic-sparse; high values compress a busy terminal-area workload into a small block. | | Horizontal minimum | *S<sub>h</sub>* | 3 / 5 / 10 NM | 3 NM is the terminal radar minimum (ICAO Doc 4444 permits it where the surveillance supports it; FAA JO 7110.65 uses it inside 40 NM of the antenna). 5 NM is the en-route radar standard. 10 NM is the wide-buffer case: a real 10 NM longitudinal minimum exists in procedural airspace using on-board distance measuring, with the leading aircraft at least 20 kt faster in true airspeed. | | Vertical minimum | *S<sub>v</sub>* | 1000 / 2000 ft | The RVSM switch, applied between FL290 and FL410. 1000 ft yields 13 usable levels, 2000 ft yields 7. Aircraft on a level that ceases to exist are re-cleared to the nearest legal one. | | Look-ahead | *T* | 1 – 20 min | Horizon of the conflict probe. Around 2 min is the STCA regime; 8 – 20 min is medium-term conflict detection. | | Resolution tool | — | vector / level / speed | Which instrument the **Issue** button uses. Each searches its own option space, scores candidates by forward-integrating the real flight model, and bills its own cost account. | | Issue instruction | — | button, key **I** | Applies the cheapest version of the selected tool that clears the earliest conflict without creating a new one, and reports the result in standard phraseology. If nothing clears it says so. | | Vertical exaggeration | *k* | 1 – 40 × | Drawing only; no readout is affected. At *k* = 1 the block is at true scale and the protected volumes are visibly flat discs. | | Time compression | — | 0 – 60 × real time | Simulated seconds per wall-clock second. 0 is paused. Conflicts develop over minutes, so the default is 25×. | | Protected volumes | — | toggle | Draws each cylinder as a translucent fill plus an explicit wire cage, and shows the ghost volumes at the predicted point of intersection. | | Flight-level shelves | — | toggle | The ladder of legal cruising levels, coloured by the semicircular rule: cool eastbound, warm westbound. | | Radar data blocks | — | toggle | Callsign, current flight level with a climb/descend trend, and ground speed in tens of knots. | | *(readout)* Closest pair | — | NM and ft | The pair with the smallest normalised margin max(*d<sub>h</sub>*/*S<sub>h</sub>*, *d<sub>v</sub>*/*S<sub>v</sub>*), reported in real units. A value of 1 means exactly at the minima. | | *(readout)* Time to CPA | *t*<sub>CPA</sub> | mm:ss, horizontal | −(**r**₀·**v**)/\|**v**\|², clamped at zero for a diverging pair. | | *(readout)* Conflicts | — | count | Pairs whose protected volumes are predicted to intersect inside the look-ahead, with current losses of separation counted separately. | | *(readout)* Cost accounts | — | NM, ft and kg, s | Track miles added by vectors, feet commanded and climb fuel spent by level changes, and delay accrued by speed control. | ## Learning objective Come away able to state the separation test as a geometric one — two protected cylinders, an intersection that requires *both* the horizontal and the vertical margin to fail — and able to explain, from the 30:1 aspect ratio of that cylinder, why a thousand feet is worth so much more than a mile. Be able to derive the time of closest approach from the relative velocity vector, and to say why a linear conflict probe is both the right tool at two minutes and the wrong tool at twenty. Be able to reproduce the RVSM arithmetic: seven levels to thirteen between FL290 and FL410, six additional levels, achieved by improving altimetry rather than by building anything. And be able to argue about the three tactical tools in terms of what each one costs and who pays: miles, fuel and a shared level, or minutes. ## Limits and connections **What the model leaves out.** There is no wind. Ground speed equals true airspeed throughout, which is a serious simplification: winds above 100 kt are routine at cruise level, they differ between adjacent levels, and it is precisely because of them that longitudinal spacing in procedural airspace is worked in *time* and held with the Mach number technique rather than measured in miles. There is no altimetry error and no surveillance lag: every aircraft's state is known exactly and instantaneously, which is the opposite of a controller's situation. Nothing deviates for weather, holds, or fails, and there is no coordination with adjacent sectors, no flow management and no arrival sequencing — all of which consume more of a real controller's attention than tactical conflict resolution does. **Wake turbulence separation is a different thing** and is not modelled. It is a distance or time minimum based on the mass categories of the two aircraft, applied mainly on approach and departure, and it is often larger than the radar minimum: a light aircraft behind a heavy one needs considerably more than 3 NM. Radar separation and wake separation are applied together, and whichever is greater wins. **The conflict probe is deliberately naive** in the same way real ones are. It extrapolates linearly from the current state. An aircraft in a turn, or one about to level off, will be mispredicted. The simulation's *resolution* logic is honest about this in a way the *detection* logic is not: candidate manoeuvres are scored by forward-integrating the full flight model, turn rate and climb rate and thrust limit included, so the advice it offers is advice the aeroplane can actually fly. **Where the field is genuinely unsettled.** How much of the separation task should be automated is not a solved question. Ground-based conflict detection and resolution has been demonstrated in research systems since the 1990s, and trajectory-based operations promise to move the work from tactical intervention to strategic de-confliction; whether that improves safety, or merely moves the controller into a monitoring role that humans perform badly, is argued about seriously. The capacity benefit of RVSM is likewise easier to state than to measure: the six extra levels are real, but how much of the observed throughput gain is attributable to them rather than to concurrent changes in route structure, surveillance and automation is contested. ### TCAS, and why the rule is now unambiguous Everything described above is a **ground** service. Airborne there is a second, entirely independent net: the **traffic alert and collision avoidance system** (TCAS II; ACAS II in ICAO usage). It is not a separation service and does not know about flight levels, sectors or clearances. It interrogates the transponders of nearby aircraft, estimates time to closest approach, and issues a **traffic advisory** roughly 35 to 48 seconds out and a **resolution advisory** roughly 20 to 35 seconds out, depending on altitude band. Resolution advisories in TCAS II are **vertical only** — climb, descend, or maintain a vertical rate — and are coordinated between the two aircraft over the Mode S data link so that they always select opposite senses. The vertical miss distance it aims for is a few hundred feet, rising to around 700 ft at high level. Compare those numbers with the ATC minima. TCAS acts tens of seconds before impact and settles for hundreds of feet; the controller works in minutes and thousands of feet. **By the time a resolution advisory fires, separation has already been lost or is seconds from being lost.** TCAS is a last resort, and a last resort that can and does contradict a controller's instruction. That contradiction was, for a time, not clearly resolved in the regulations. On **1 July 2002**, at 21:35 UTC over Überlingen in southern Germany, a Bashkirian Airlines Tupolev Tu-154M (flight 2937, Moscow to Barcelona) and a DHL Boeing 757-200F (flight 611, Bergamo to Brussels) collided. All 71 people on board both aircraft were killed: 69 on the Tupolev, among them 45 children, and the two pilots of the 757. The German Federal Bureau of Aircraft Accident Investigation (BFU) documented the sequence in report AX001-1-2/02, published in May 2004. Both aircraft were at FL360 on converging tracks in airspace controlled by Skyguide from Zurich. A single controller was working two positions during a quiet night shift, an unofficial but tolerated practice; one radar system had been placed in a degraded fallback mode for scheduled maintenance, which disabled the optical short-term conflict alert; and the direct telephone lines were also out because of that maintenance, so the neighbouring centre at Karlsruhe, whose own alert had triggered, could not get through to warn Zurich. Less than a minute before the collision the controller instructed the Tupolev to descend. Seconds later both aircraft received resolution advisories: the Tupolev's told it to climb, the 757's told it to descend. The Tupolev's crew followed the controller. The 757 followed its RA. Both aircraft descended, into each other. The investigation identified as causal, alongside the organisational and technical failures at the ANSP, that the regulations governing ACAS were **incompletely and partly contradictorily worded**. A near-collision over Suruga Bay, Japan, in January 2001 had already exposed the same ambiguity without producing an unambiguous rule. It is unambiguous now. ICAO's ACAS operating procedures in PANS-OPS (Doc 8168) require pilots to respond immediately to a resolution advisory by following it as indicated, never to manoeuvre in the opposite sense to an RA, and — explicitly — to **follow the RA even if it conflicts with an air traffic control instruction to manoeuvre**. The single stated exception is where following it would itself jeopardise the safety of the aircraft, which is the pilot-in-command's authority reasserting itself at the last possible point. The crew notify ATC as soon as workload permits and return to the clearance when the conflict is resolved. The complementary provision in PANS-ATM (Doc 4444) closes the loop from the other side: once an aircraft departs from its clearance in compliance with an RA, the controller **ceases to be responsible** for separation between that aircraft and any other affected as a direct consequence of the manoeuvre. In the United States the same priority is written into 14 CFR 91.123(a), which permits deviation from a clearance in response to a TCAS resolution advisory. Two rules, then, that sound contradictory and are not. The controller separates traffic and the crew complies with the controller — until the airborne system says otherwise, at which point the airborne system wins, because it is the only one still operating on the timescale that matters. And behind both, unchanged, sits the commander's final authority over the aeroplane. ## References - International Civil Aviation Organization. *Annex 2 to the Convention on International Civil Aviation: Rules of the Air*, 10th edition. Montreal: ICAO, 2005 (with subsequent amendments). See §2.4 (authority of the pilot-in-command) and Appendix 3 (tables of cruising levels). - International Civil Aviation Organization. *Annex 11 to the Convention on International Civil Aviation: Air Traffic Services*, 15th edition. Montreal: ICAO, 2018. See §2.2 (objectives of the air traffic services) and Appendix 4 (ATS airspace classes). - International Civil Aviation Organization. *Doc 4444, Procedures for Air Navigation Services — Air Traffic Management (PANS-ATM)*, 16th edition. Montreal: ICAO, 2016. Chapters 5 (separation methods and minima), 8 (ATS surveillance services) and 15 (ACAS procedures for controllers). - International Civil Aviation Organization. *Doc 8168, Procedures for Air Navigation Services — Aircraft Operations (PANS-OPS), Volume I: Flight Procedures*. Montreal: ICAO. ACAS II operating procedures. - International Civil Aviation Organization. *Doc 9574, Manual on a 300 m (1 000 ft) Vertical Separation Minimum Between FL 290 and FL 410 Inclusive*, 3rd edition. Montreal: ICAO, 2012. - International Civil Aviation Organization. *Doc 7488, Manual of the ICAO Standard Atmosphere (extended to 80 kilometres)*, 3rd edition. Montreal: ICAO, 1993. - Federal Aviation Administration. *Order JO 7110.65, Air Traffic Control*. Washington, DC: FAA (current edition). Chapter 5, radar separation minima. - Federal Aviation Administration. *Advisory Circular 91-85, Authorization of Aircraft and Operators for Flight in Reduced Vertical Separation Minimum (RVSM) Airspace*. Washington, DC: FAA. - United States. *Code of Federal Regulations, Title 14, Part 91*: §91.3 (responsibility and authority of the pilot in command) and §91.123 (compliance with ATC clearances and instructions, including the TCAS resolution advisory exception). - Bundesstelle für Flugunfalluntersuchung (German Federal Bureau of Aircraft Accidents Investigation). *Investigation Report AX001-1-2/02*: mid-air collision near Überlingen, Lake Constance, 1 July 2002. Braunschweig: BFU, May 2004. - Eurocontrol. *ACAS II Guide: Airborne Collision Avoidance System II* (incorporating TCAS II Version 7.1). Brussels: Eurocontrol. - Reich, P. G. "Analysis of Long-Range Air Traffic Systems: Separation Standards." *Journal of the Institute of Navigation* (Royal Institute of Navigation), vol. 19, 1966, published in three parts. The origin of the collision-risk modelling framework still used to justify oceanic and vertical separation minima. - Kuchar, James K., and Lee C. Yang. "A Review of Conflict Detection and Resolution Modeling Methods." *IEEE Transactions on Intelligent Transportation Systems*, vol. 1, no. 4, December 2000, pp. 179–189. - Erzberger, Heinz. "Automated Conflict Resolution for Air Traffic Control." *25th International Congress of the Aeronautical Sciences (ICAS)*, Hamburg, 2006. - Hoekstra, Jacco M., and Joost Ellerbroek. "BlueSky ATC Simulator Project: An Open Data and Open Source Approach." *7th International Conference on Research in Air Transportation (ICRAT)*, 2016. An open-source air traffic simulator whose conflict-detection formulation matches the one used here. - Nolan, Michael S. *Fundamentals of Air Traffic Control*, 5th edition. Delmar Cengage Learning, 2011. Standard undergraduate treatment of separation standards and control procedures. - Eurocontrol Experimental Centre. *User Manual for the Base of Aircraft Data (BADA)*. Brétigny-sur-Orge: Eurocontrol. Source of representative cruise masses, Mach numbers and climb performance of the aircraft types named in the simulation. **On the spine:** [[Aircraft]] · [[Aircraft_flight_dynamics]] · [[Fixed-wing_aircraft]] · [[Helicopter]] · [[Turbojet]] · [[Jet_engine]] · [[Sonic_boom]] · [[Contrail]] · [[Air_traffic_control]] · [[Avionics]] · [[Aviation]].  **Part of the [[Aviation]] hub** — main article for section A23, *Air traffic control*. Related sections: Automatic Dependent Surveillance–Broadcast · Airband · [[Traffic_collision_avoidance_system]].  ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Air_traffic_control) : [Wikitube](https://en.wikitube.io/wiki/Air_traffic_control) --- *PORTAL_Aviation three.js batch · 2026-08-05 · sim staged in `_3d_deploy_stage/`.*