Radio_navigation - Wikiparency

# Radio navigation > [[PORTAL_Aviation|Aviation]] · [[PORTAL_Avionics|Avionics]] spine.  ## Microsims — three.js ### Radio navigation (three.js) <div class="microsim-player"> <iframe src="https://wikitube-3d-microsims.netlify.app/Radio_navigation.html" width="100%" height="620" frameborder="0" loading="lazy" sandbox="allow-scripts allow-same-origin" title="Radio navigation — three.js microsim"></iframe> </div> **Open it full-screen:** [Radio_navigation.html](https://wikitube-3d-microsims.netlify.app/Radio_navigation.html) · library `threejs` · route `microsim/threejs/` ### Related microsims Live sims on neighbouring articles: - [[Satellite_navigation]] - [[Inertial_navigation_system]] - [[Autopilot]] - [[Air_navigation]] - [[Terrain_awareness_warning_system]] - [[Avionics]] *Sim hosted off-article; the article owns the reference, not the runtime (WIKI_RULES §10.4).*  ## Overview Radio navigation is the practice of finding out where you are by measuring something about a radio signal whose transmitter is at a known place. The measurement is never a position. It is an angle, or a distance, or a difference of two distances, and each one of those confines you not to a point but to a **surface**. Your position is the intersection of the surfaces. That is the whole subject, and everything else — the phase comparators, the pulse pairs, the coding delays — is engineering in the service of measuring one number accurately enough that the surfaces cross in a small region rather than a large one. The surfaces are genuinely three-dimensional, and the habit of drawing them on a flat chart hides some of the most consequential facts about them. A VOR bearing does not confine you to a line; it confines you to a vertical half-plane standing on the station, and the line you draw on the chart is that plane cut by your altitude. A DME range does not confine you to a circle; it confines you to a **sphere**, and the circle on the chart is that sphere cut by your altitude — a circle whose radius is smaller than the number in the DME window, sometimes very much smaller. Fly directly over a DME at 30,000 ft and the box will not read zero. It will read about 4.9 nautical miles, because 30,000 ft *is* 4.94 NM and the aeroplane is that far above the antenna. Nothing has malfunctioned. The instrument is measuring exactly what it claims to measure: the straight-line, air-to-ground distance from you to the station, which the trade calls **slant range**. The history of the subject is a hundred-year argument between the two ways of using a radio wave: measure its *direction*, or measure its *time of flight*. The first practical airway system, the American four-course radio range of 1928, measured direction in the crudest possible way — two interlocking Morse patterns, "A" (dot-dash) in one pair of quadrants and "N" (dash-dot) in the other, which merged into a steady tone along four narrow beams. It worked, more than four hundred stations were built, and generations of pilots flew the beam by ear in cloud. It also drifted with the ionosphere at night, bent around terrain, and told you nothing at all about distance. The Second World War produced the alternative. Gee (British, in service from March 1942), LORAN-A (American, 1943) and Decca (British, from the Normandy landings in 1944) all measured time differences between synchronised transmitters, and all produced **hyperbolic** lines of position. These systems were long-legged and worked over water, where there is nothing to put a beacon on; their descendants — LORAN-C from 1957, Omega, Chayka — were the first genuinely global radio navigation aids. Meanwhile the direction-measuring branch grew up: the VHF omnidirectional range (VOR), put into service by the US Civil Aeronautics Administration in 1949 and standardised internationally by ICAO in Annex 10, replaced the four-course range with a beacon that could deliver any one of 360 radials rather than four beams. Distance measuring equipment (DME), a civil descendant of wartime radar beacon interrogation, was standardised alongside it, and the military TACAN system merged the two functions into one antenna. A VOR and a TACAN on the same site is a VORTAC, and much of the world's civil DME service is actually the distance half of a TACAN. Satellite navigation then did to VOR/DME roughly what VOR/DME had done to the radio range — except that it has not finished the job, and the reason it has not is worth stating plainly. A GNSS receiver measures time of flight from transmitters that are 20,000 km away and radiating a few tens of watts; the signal arriving at the antenna is below the thermal noise floor. That makes it superbly accurate and trivially easy to jam. Ground navaids radiate kilowatts from a few tens of miles away. Since 2022 the European Union Aviation Safety Agency has had a standing Safety Information Bulletin on GNSS outages and spoofing, and interference around several conflict zones has become routine rather than exceptional; in late April 2024 Finnair suspended its Tartu service for a month because the only instrument approach at that airport was satellite-based. Both the FAA and the European air navigation service providers have consequently converged on the same policy: thin the conventional network, but keep a deliberately designed **minimum operational network** as the fallback. The FAA's plan retains roughly 594 VORs out of a network from which about 302 were selected for discontinuance, sized so that an aircraft anywhere in the continental United States at or above 5,000 ft AGL is within 100 NM of an airport with a non-GNSS instrument approach. At the same time the FAA's NextGen DME programme is *adding* ground stations, because a DME/DME/IRU area-navigation fix is the designated backup for RNAV routes. Europe has made the same bet with a heavier weighting towards DME. Radio navigation is not a museum subject; it is the layer underneath. ## The physics ### The VOR: a bearing from a phase difference A conventional VOR (CVOR) radiates two 30 Hz signals. The first is a **reference** phase, frequency-modulated onto a 9,960 Hz subcarrier and radiated omnidirectionally, so that its phase is the same everywhere around the station. The second is a **variable** phase, produced by rotating a directional (limaçon-shaped) radiation pattern thirty times a second, which reaches the receiver as 30 Hz amplitude modulation whose phase depends on where you are standing. The two are arranged to be in phase at magnetic north. A receiver therefore only has to compare the phase of the reference with the phase of the variable, and the difference in degrees *is* the magnetic bearing from the station: theta = phase(variable) - phase(reference) A Doppler VOR (DVOR) swaps the roles. The reference becomes amplitude modulation from a central antenna, and the variable becomes apparent frequency modulation generated by electronically commutating a ring of typically 48 antennas so that the radiating point appears to travel in a circle at 30 revolutions per second — a real Doppler shift from a synthetic moving source. The information content is identical, but DVOR is markedly less sensitive to reflections from terrain and buildings, which is why almost every VOR built in the last few decades is one. Annex 10's requirement on radiated bearing accuracy is conventionally quoted as about ±2°, with Doppler installations typically doing better — nearer ±1° — because they are so much less affected by reflections. Airborne receiver error and flight technical error take the practical *total system use* accuracy out to roughly ±4 to ±5° at 95 %, which is the number behind the ±4 NM half-width of a Victor airway out to 51 NM from the station. The geometric consequence is the important part. The locus of all points that would report bearing θ from a station at **s** is a vertical half-plane containing **s**. A bearing pins you *across* the radial and not at all *along* it, and the cross-track uncertainty grows linearly with distance: sigma_across = r * sigma_theta At 1.7° of one-sigma bearing error, that is 0.30 NM at 10 NM and 3.0 NM at 100 NM. This single fact explains most of the operational lore about VOR — why airways splay outward, why you get a better fix from a near station than a far one, and why two VORs on opposite sides of you give a much tighter cut than two in front. ### The DME: a range from a round trip DME is not a passive receiver; it is a two-way transaction. The airborne interrogator transmits a **pair** of pulses (spaced 12 µs on an X channel, 36 µs on a Y channel) in the 960–1215 MHz band. The ground transponder receives the pair, waits a **fixed turnaround delay** of nominally 50 µs, and replies on a frequency offset 63 MHz from the interrogation. The interrogator measures the total elapsed time, subtracts the known 50 µs, and converts what is left at 12.3552 microseconds per nautical mile, round trip because that is 2 × 1852 m divided by the speed of light. The fixed delay is what makes DME an absolute measurement rather than a relative one: it is a bias that is known exactly and removed exactly. It is also why a DME can never read a true zero. The interrogator identifies its own replies among all the others by jittering its interrogation rate and correlating; a single ground transponder can serve on the order of a hundred aircraft, and radiates filler pulses ("squitter") when demand is low so that its automatic gain control stays in range. What DME measures is the straight-line distance from aircraft to antenna. If the ground range is *r* and the aircraft is Δh above the antenna, then D = sqrt(r^2 + dh^2) and D - r = sqrt(r^2 + dh^2) - r Two limits matter, and the microsim draws the whole curve between them at true 1:1 vertical scale: * **Far away**, expand the square root: `D - r ≈ dh^2 / (2r)`. At 30,000 ft and 60 NM the error is 0.2 NM — comfortably inside the noise of everything else. * **Overhead**, `r → 0` and `D - r → dh`. The error becomes the whole of your height. At 30,000 ft that is 4.94 NM; at 41,000 ft it is 6.75 NM. Plot `D - r` vertically above the ground track at true scale and the curve hugs the surface far out and rises to touch the aircraft's own altitude exactly over the station. That is the picture the sim is built around. The same geometry does something subtler and more damaging than a mere reading error. The sensitivity of the *measurement* to *horizontal* position is dD/d(horizontal) = r / D which goes to zero overhead. A range measurement taken from directly above a station carries no horizontal information at all: the sphere is tangent to your altitude plane, and sliding sideways does not change the distance to first order. Equivalently, one sigma of range error becomes `sigma_D * D / r` of position error, which diverges. This is the **cone of confusion**, arrived at from pure geometry with no mention of antenna radiation patterns — though those contribute too, since the vertical pattern of a VOR or DME nulls overhead and both aids typically drop out in a cone above the station. A modern flight management system corrects for slant range by subtracting the known height difference, and the microsim's *slant-range correction* toggle does exactly that. Turn it off and the fix is pushed radially outward from every station by that station's own slant error — which, close in, is a position error of miles. ### Hyperbolic systems: a range difference A hyperbolic chain has one master transmitter and two or more secondaries, all synchronised. A receiver measures the **time difference** (TD) between the arrival of the master's pulse group and a secondary's. The locus of points with a constant difference of distances to two fixed points is, by the classical focal definition, a hyperbola with those two points as foci. In the frame whose first axis runs from master **M** to secondary **S**, with half-baseline `c` and semi-transverse axis `a = (r_M - r_S)/2`, the branch is exactly u = a * cosh(t), rho = b * sinh(t), b^2 = c^2 - a^2 which is how the microsim generates its curves — analytically, not by contouring a grid. The TD a receiver actually reads is not the raw difference. It is TD = (baseline travel time) + (coding delay) + (r_S - r_M)/c The coding delay is a deliberate constant added by each secondary so that the master's pulse group always arrives first everywhere in the coverage area, which removes the ambiguity about which station you are hearing. Chains are identified by their **group repetition interval**: the US north-east LORAN-C chain used GRI 9960, meaning 99,600 µs between pulse groups. LORAN-C ran at 100 kHz as a **groundwave** system, and that has a geometric consequence worth dwelling on. A groundwave path length is a distance measured over the surface; the aircraft's altitude barely enters it. So a LORAN line of position is a vertical **hyperbolic cylinder**, not the hyperboloid of revolution that a genuinely three-dimensional time-difference system (GNSS time-difference processing, say) would produce. Cut it at any flight level and you get the same hyperbola — which is exactly why a printed LORAN lattice chart was valid at every altitude, and why a DME chart is not. The gradient of a range-difference observation is the *difference of two unit vectors*: grad(TD) proportional to u_S - u_M whose magnitude is 2 on the baseline between the transmitters and falls to zero along the baseline **extension** beyond either one. That is why chain coverage diagrams carry baseline- extension warning regions: out there the hyperbolas open into nearly parallel straight lines and the geometry is worthless. LORAN-C's published absolute accuracy was about 0.25 NM (460 m) 2 drms, but its *repeatable* accuracy was tens of metres. The gap between those two figures is not receiver noise; it is the Additional Secondary Factor, the extra delay a groundwave picks up crossing land of imperfectly known conductivity. The microsim uses an effective 0.8 µs one-sigma TD error chosen to reproduce the absolute figure, and says so in the code, because using the receiver's true timing jitter would make LORAN look ten times better than it ever was. ### The fix, and dilution of precision With more than two measurements the surfaces do not meet at a point. Three VOR radials drawn from noisy bearings form a small triangle — the navigator's **cocked hat** — and the job is to find the point that best explains all of them. Linearise each observation about an estimate **p**, giving for measurement *j* a gradient direction **u**ⱼ in the horizontal plane and an equivalent position-domain one-sigma error *e*ⱼ along it: | aid | direction **u** | position sigma *e* | |---|---|---| | VOR bearing | perpendicular to the radial | `r * sigma_theta` | | DME range | along the radial | `sigma_D * D / r` | | LORAN TD | along `u_S - u_M` | `sigma_TD * c / abs(u_S - u_M)` | The information matrix and the covariance are then I = sum_j ( u_j u_j^T / e_j^2 ) C = I^-1 **C** is a 2×2 matrix whose eigenvectors give the axes of the one-sigma error ellipse and whose eigenvalues give the squares of its semi-axes. Its trace gives the distance-root-mean-square error, `drms = sqrt(trace C)`. Dividing that by the one-sigma error of a single reference measurement — here the nominal DME accuracy — strips out the units and leaves a pure statement about **where the stations are**: the horizontal dilution of precision, HDOP. HDOP = sqrt(trace C) / sigma_reference HDOP near 1 means the geometry has cost you nothing. HDOP of 6 means the same instruments, the same accuracy, the same everything, produce a fix six times worse because the lines of position cross at a shallow angle. Two range circles that cross at right angles bound you in both directions equally; two that are nearly tangent leave you free to slide a long way along the tangent. This is precisely why the FAA requires a DME/DME area-navigation position to be computed from stations subtending between 30° and 150° at the aircraft, and why the microsim's *near-collinear* layout — three stations in a line, with the track running along that line — produces an ellipse ten times longer than it is wide even though every individual measurement is as good as it ever was. Finally, none of it happens at all if the signal does not arrive. VHF and UHF navaids are line-of-sight. Standard atmospheric refraction bends the ray downward by about the amount you would get from an earth 4/3 its true size, which puts the radio horizon at d (NM) = 1.23 * ( sqrt(h_aircraft in ft) + sqrt(h_station in ft) ) — 213 NM from 30,000 ft, but only 62 NM from 2,500 ft. The microsim folds the same 4/3-earth bulge into a terrain ray-march, so stations disappear both because a ridge is in the way and because the earth itself is. Descend in the sim and watch the fix collapse as the count of usable lines of position falls from six to two. ## Controls -> what each maps to | Control | Symbol | Range / units | What it changes | |---|---|---|---| | Aircraft altitude | `h` | 1,000 – 41,000 ft MSL (step 500) | Sets the height of the aircraft's altitude plane, and therefore Δh above each station. Drives slant range `D = sqrt(r^2 + dh^2)`, the radius `sqrt(D^2 - dh^2)` of every DME line-of-position circle, and the radio horizon `1.23*sqrt(h)`. This is the single most important control in the sim. | | Aircraft position | mode | *fly the racetrack* / *manual placement* | The racetrack is a 54 NM-straight, 9 NM-radius circuit at 420 kt ground speed, compressed 20×, whose inbound leg passes 0.7 NM abeam the primary station so the overhead pass happens once a lap. In manual mode the aircraft is frozen and placed by the two sliders, the arrow keys (1 NM, or 5 NM with shift) or a click on the terrain. | | East / north of centre | `e`, `n` | ±55 NM (step 0.25) | Horizontal position relative to the centre of the charted square. Live in manual mode; read-only while the track is flying. | | Station layout | — | good spread / poor spread / near-collinear | Moves the three VOR/DME sites. *Good* puts them about 120° apart around the aircraft (HDOP ≈ 1.2, near-circular ellipse). *Poor* crams all three into one quadrant subtending about 25° (HDOP rising past 4.5). *Near-collinear* places them in a straight line with the track along it, so ranges pin you along the line and say nothing across it. | | Measurement noise | `k` | 0 – 3 × nominal (step 0.05) | Multiplies all three one-sigma figures together: VOR 1.7°, DME 0.10 NM, LORAN 0.80 µs. Drives a first-order Gauss-Markov (Ornstein–Uhlenbeck) process with a 5 s correlation time on each of the eight measurement channels, so the lines of position wander the way real ones do instead of flickering. Note that HDOP does **not** move when you drag this: dilution is geometry, not accuracy. | | VOR radial planes | — | on/off | Draws the vertical half-plane of each bearing measurement, its straight line of position at the aircraft's altitude, and the ±1σ wedge at true scale. Also includes or removes those bearings from the least-squares fix. | | DME spheres | — | on/off | Draws the sphere of the nearest station's measured slant range, every station's LOP circle where its sphere cuts the aircraft's altitude, the ±1σ annulus, and the `D - r` error profile. | | LORAN hyperbolic sheets | — | on/off | Draws the two vertical hyperbolic cylinders for the master–X and master–Y time differences, their hyperbolas draped on the terrain, and their ±1σ strips. Also reveals the three chain transmitters, ~300 NM outside the charted square. | | Slant-range correction | — | on/off | On: the receiver converts each measured slant range to ground range using known altitude, `r = sqrt(D^2 - dh^2)`. Off: the raw slant range is fed to a two-dimensional solver as if it were a ground range, which is what a bare DME needle and an old-fashioned VOR/DME plot do. Watch "fix error now" in the HUD. | | Uncertainty ellipse + scatter | — | on/off | The 1σ error ellipse from the covariance matrix, plus 300 Monte-Carlo draws from that same covariance through its Cholesky factor. Both are drawn 20× actual size — the only exaggerated quantity anywhere in the scene, and labelled as such in the HUD, the legend and the caption. | | Pause / Reset | — | — | Reset restores the default state and re-seeds the noise generator, so any run is reproducible. Reduced-motion users start paused with no autoplay motion. | **Live HUD:** aircraft east/north and altitude in feet, nautical miles and flight level; the nearest station's ident and radial; ground range `r`; slant range `D`; the slant error `D - r` in NM and per cent; the 1σ fix uncertainty as drms and as ellipse semi-axes; HDOP; and the instantaneous fix error. The live equation is `D = sqrt(r^2 + h^2)` with the current numbers substituted, plus the round-trip interrogation time `D × 12.3552 + 50 µs`. A collapsed panel gives per-station radial, slant range, ground range and signal status, the radio horizon at the current altitude, and both LORAN time differences in microseconds. ## Learning objective After using this sim you should be able to state, without looking anything up, what surface each kind of radio navigation measurement corresponds to; explain why a DME over-reads by an amount equal to your height above the station when you fly over it, and why that error is negligible at range; predict whether a given arrangement of ground stations will give a tight or a smeared fix, and say which direction the smear will run; and distinguish clearly between the *accuracy* of a measurement and the *dilution* that geometry applies to it. You should also be able to explain why descending makes a VHF fix worse for reasons that have nothing to do with the receiver. ## Limits and connections The model is deliberately honest about scale and deliberately simplified about physics. Worth knowing before you quote any number out of it: * **Flat earth, true bearings.** The 110 NM square is Euclidean. Real bearings and ranges are geodesic, and — more importantly — real VOR radials are **magnetic**, aligned to the local variation at the time the station was last flight-checked. Magnetic variation is not modelled here at all, and the drift of the magnetic field is a live operational problem: stations must periodically be re-aligned, and a few have been decommissioned rather than re-aligned. * **True 1:1 vertical scale everywhere except two clearly labelled things.** The uncertainty ellipse and its scatter cloud are drawn 20× actual size, because a good three-station fix is a few hundred metres across in a 110 NM picture. The aircraft symbol is roughly fifty times life size. Terrain *shading* is computed from an exaggerated surface normal, the way every printed relief map does it, but the terrain *geometry* is genuinely 1:1 — it has to be, or the DME sphere would stop being a sphere. * **The LORAN chain is compressed.** Real LORAN-C baselines were 600–1000 NM; the sim's are about 300 NM, so its hyperbolas are more strongly curved across the charted area than a real chain's would be. The transmitters are still placed well outside the mapped square, which is the honest half of the picture. * **No propagation modelling.** There is no Additional Secondary Factor computation, no sky-wave contamination, no VOR *scalloping* or *bending* from terrain reflections — the effects that motivated Doppler VOR in the first place. Terrain enters only as a hard line-of-sight obstruction plus the 4/3-earth bulge. * **No receiver dynamics.** Real DME goes into "memory" and coasts when it loses lock; real VOR raises an OFF flag inside the cone of confusion; real receivers have tracking loops with lag, and real avionics apply a Kalman filter with an inertial reference rather than solving each fix from scratch. Here every frame is an independent snapshot least-squares solution, seeded from the previous fix. * **Noise is a stand-in.** A first-order Gauss-Markov process with a 5 s correlation time is a reasonable caricature of a wandering bias, but real navaid error budgets combine a slowly varying siting error, a receiver error and a flight technical error with quite different spectra. * **DME channel capacity and pulse structure are described, not simulated.** No interrogation jitter, no reply efficiency, no X/Y channel pairing. The connections run outward in every direction. [[Satellite_navigation]] is the same intersection-of-surfaces problem with the transmitters in orbit, four unknowns instead of two (the receiver clock is the fourth), and pseudoranges instead of ranges — and the DOP algebra in this sim is literally the same algebra, which is why the acronym HDOP is borrowed from GNSS. [[Inertial_navigation_system]] is the complement: no external signal at all, unjammable, but drifting with time, which is why a modern position solution blends inertial with DME/DME or GNSS rather than choosing. [[Autopilot]] and the flight management system are what consume the fix. [[Terrain_awareness_warning_system]] shows what happens when a position solution and a terrain database are combined, and why an error of a mile in the wrong place is not an academic matter. [[Air_navigation]] frames the whole problem, and [[Avionics]] is the box the receivers live in. ## References * International Civil Aviation Organization. *Annex 10 to the Convention on International Civil Aviation: Aeronautical Telecommunications, Volume I — Radio Navigation Aids.* ICAO, Montreal. (The primary standard for VOR, DME and ILS: signal formats, accuracies and coverage requirements.) * Kayton, Myron, and Walter R. Fried, eds. *Avionics Navigation Systems.* 2nd ed. New York: John Wiley & Sons, 1997. ISBN 978-0-471-54795-8. (The standard engineering reference; chapters on terrestrial radio navigation and on position-fixing geometry.) * Bowditch, Nathaniel. *The American Practical Navigator* (Pub. No. 9). National Geospatial-Intelligence Agency. Chapter on Loran navigation. (Hyperbolic lattices, time differences, coding delays, and the distinction between absolute and repeatable accuracy.) * Misra, Pratap, and Per Enge. *Global Positioning System: Signals, Measurements, and Performance.* 2nd ed. Lincoln, MA: Ganga-Jamuna Press, 2006. ISBN 978-0-9709544-1-1. (Weighted least squares, the covariance matrix and dilution of precision, developed for GNSS but identical in form to the terrestrial case.) * Parkinson, Bradford W., and James J. Spilker Jr., eds. *Global Positioning System: Theory and Applications, Volume I.* Progress in Astronautics and Aeronautics, vol. 163. Reston, VA: AIAA, 1996. ISBN 978-1-56347-106-3. * Langley, Richard B. "Dilution of Precision." *GPS World*, vol. 10, no. 5, May 1999, pp. 52–59. (The clearest short treatment of what DOP is and is not.) * RTCA. *DO-189: Minimum Operational Performance Standards for Airborne Distance Measuring Equipment (DME) Operating within the Radio Frequency Range of 960–1215 MHz.* RTCA, Washington DC. * Federal Aviation Administration. *Aeronautical Information Manual*, Chapter 1, Section 1: Navigation Aids. (Service volumes, VOR accuracy checks, the cone of confusion, DME slant range.) * Federal Aviation Administration. *Instrument Flying Handbook*, FAA-H-8083-15B, 2012. * Federal Aviation Administration. "Provision of Navigation Services for the Next Generation Air Transportation System (NextGen) Transition to Performance-Based Navigation (PBN): Plan for Establishing a VOR Minimum Operational Network." *Federal Register*, 26 July 2016. (The VOR MON criteria: coverage at and above 5,000 ft AGL, and a MON airport within 100 NM of any point in the continental United States.) * United States Coast Guard, Department of Homeland Security. "Terminate Long Range Aids to Navigation (Loran-C) Signal." *Federal Register*, 7 January 2010. (US LORAN-C transmissions ceased on 8 February 2010; the Northwest European Loran System chains followed at the end of 2015.) * European Union Aviation Safety Agency. *Safety Information Bulletin 2022-02: Global Navigation Satellite System Outage and Alterations Leading to Communication / Navigation / Surveillance Degradation*, and subsequent revisions. * Komons, Nick A. *Bonfires to Beacons: Federal Civil Aviation Policy Under the Air Commerce Act, 1926–1938.* Washington DC: US Department of Transportation, Federal Aviation Administration, 1978. (The lighted airways, the four-course radio range and the origins of the federal airway system.) **On the spine:** [[Avionics]] · [[Fly-by-wire]] · [[Autopilot]] · [[Head-up_display]] · [[Traffic_alert_and_collision_avoidance_system]] · [[Terrain_awareness_warning_system]] · [[Radio_navigation]] · [[Radiation_hardening]] · [[Radar]] · [[Air_traffic_control]] · [[Aviation]].  **Part of the [[Avionics]] hub** — main article for section X7, *Navigation*. Related sections: Pitot–static system · Global Positioning System · Instrument landing system.  ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Radio_navigation) : [Wikitube](https://en.wikitube.io/wiki/Radio_navigation) --- *PORTAL_Avionics three.js batch · 2026-08-05 · sim staged in `_3d_deploy_stage/`.*