Sensing and Navigation: A Practical Reference

An active electronically scanned array replaces the single transmitter and mechanical gimbal of a conventional radar with a few hundred to a few thousand transmit-receive modules, each with its own amplifier and phase shifter. Steering the beam becomes a matter of adjusting phase across the array, which takes microseconds instead of the seconds a mechanical scan needs. That speed changes what a radar can do: it can interleave search, track, weather, and datalink functions in the same second, form several beams at once, and hop frequency pulse to pulse across a wide band.

Agility is the main advantage. An AESA can hold tracks on dozens of targets while continuing to search, and pulse-to-pulse frequency agility makes it much harder to jam than a fixed-frequency set. Reliability is also better, because losing 10% of the modules degrades the array gracefully rather than failing it. Low-probability-of-intercept waveforms let it look without being obviously detected. The weaknesses are cost and heat. Modules are the expensive part, gallium nitride devices need serious cooling, and the scan volume is limited to roughly ±60° from boresight before the effective aperture shrinks too far, so wide coverage needs multiple faces.

Pick an AESA for any modern fighter, air-defense, or naval fire-control application where multi-function operation and jam resistance justify the cost. If the requirement is simple wide-area surveillance at long range with a modest budget, a mechanically scanned or PESA radar delivers most of the detection performance for a fraction of the price. If the target set is low-observable, an AESA at X-band helps less than people expect — the shaping and coatings on stealth aircraft are optimized precisely against this band, so pair it with VHF/UHF surveillance or passive sensors rather than expecting the fire-control radar to solve it alone.

X-band 8–12 GHz for fire control · a few hundred to a few thousand transmit-receive modules per array · detection roughly 150–250 km against a 1 m² target on fighter-class sets · beam steering in microseconds · scan volume about ±60° off boresight per face · module cost in the hundreds of dollars at volume, down from thousands · radar cost roughly $2M–10M per fighter set.

Digital radio frequency memory jammers replay coherent false targets that look real to the processor. Stand-in jamming from an expendable drone puts noise inside the main beam. Towed decoys pull the track off the aircraft. Against low-observable targets the radar simply does not detect at useful range. Emissions also give the platform away to any electronic support receiver, which usually detects the radar well before the radar detects the target.

Raytheon APG-79 and APG-82, Northrop APG-77 (F-22) and APG-81 (F-35), Leonardo Captor-E (Typhoon), Thales RBE2-AA (Rafale), the SPY-6 and SPY-3 naval families, and the ground-based Patriot LTAMDS and Israeli EL/M-2084.

Transmit-receive module cost dominates the radar. Gallium nitride replaced gallium arsenide over the past decade and roughly tripled power density while improving efficiency, which is the main reason AESA has moved down-market to trainers and drones. Module prices have fallen from thousands of dollars to hundreds at volume. That trend is what enables AESA in missile seekers and small UAVs, and it is the single most useful number to track when assessing anyone's radar roadmap.

4 minAESA radar technology | 3D Animation | Thales | C4RealC4Real · 100k+ views

18 minWhat Are Phased Arrays?MATLAB · 100k+ views

17 minThe insane engineering of the F-35 AESA radars!Millennium 7 * HistoryTech · 100k+ views

Phased Array Antenna (Radar Tutorial) · Radar Principle (Radar Tutorial)

A passive electronically scanned array uses one central transmitter feeding a large array of phase shifters, so the beam steers electronically but every element radiates the same signal at the same frequency. It was the first practical phased array architecture and dominated from the 1960s through the 1990s. Compared with a mechanically scanned radar it gains beam agility; compared with an AESA it lacks the ability to form multiple independent beams, to hop frequency pulse to pulse, or to degrade gracefully — a transmitter failure takes the whole radar with it.

Cost is the advantage. One high-power transmitter, usually a traveling wave tube, is much cheaper than a thousand solid-state modules, and PESA gives most of the scan-rate benefit of a phased array. Peak power can be very high, which suits long-range search. The weaknesses are single-point failure, limited waveform flexibility, and jamming vulnerability: because the whole array transmits one frequency at a time, a jammer only has one thing to work against. Bandwidth through the central feed is also limited, which constrains resolution and low-probability-of-intercept operation.

Pick PESA when you need electronic beam steering at long range on a budget, particularly for ground-based and naval surveillance where size and power are not constraints. If the mission involves contested electromagnetic conditions, AESA's frequency agility is worth the cost difference. If cost is the overriding constraint and the target set is unsophisticated, a mechanically scanned radar is cheaper still. In practice most new Western designs skip PESA entirely, and its continued relevance is largely in legacy fleets and in export systems from suppliers who have not made the transition.

One central transmitter feeding hundreds to a few thousand phase shifters · peak power in the hundreds of kilowatts from a traveling wave tube · detection 150–400 km against fighter-sized targets on large sets, on vendor figures · beam steering in microseconds · scan volume about ±60° off boresight per face · unit cost roughly a third of a comparable AESA.

Noise jamming is more effective than against an AESA, because the radar's frequency agility is limited to pulse groups rather than individual pulses. Chaff and DRFM false targets work. Anti-radiation missiles home on the single high-power transmitter easily. Sidelobe jamming is a standard approach, since a central-feed array typically has worse sidelobe control than an individually weighted AESA.

The Soviet and Russian Zaslon on the MiG-31 (the first airborne phased array in service), Bars and Irbis-E on Su-30/Su-35, the Patriot MPQ-53 and MPQ-65, the AN/SPY-1 Aegis arrays, and the Chinese Type 346 in its earlier variants.

PESA is a mature and largely sunset architecture in the West, sustained by upgrade programs on fielded systems like Patriot and Aegis, both of which are transitioning to AESA replacements. In export markets it remains commercially significant, because a PESA priced at a third of an AESA is genuinely attractive to buyers facing unsophisticated threats. The economic story is essentially the transmit-receive module cost curve: as modules got cheap, PESA's cost advantage evaporated.

5 minPESA and AESA for radar systems - ISAE SUPAEROArthurSR44 · 10k+ views

5 minPhased Array AntennasMark Hickle · 100k+ views

20 minPhased Arrays - Steering and the Antenna Pattern | An Animated Intro to Phased ArraysMarshall Bruner · 100k+ views

Phased Array Antenna (Radar Tutorial) · Navy Aegis Ballistic Missile Defense (BMD) Program (Congressional Research Service)

A mechanically scanned radar rotates or nods a single antenna to sweep the beam through space. Almost every radar built before 1990 works this way, and a very large number still do: airport surveillance, weather, naval air search, and most ground-based early warning. A rotating antenna at 10–15 rpm gives a full 360° picture every four to six seconds, which is fine for tracking airliners and poor for tracking anything that maneuvers. The mechanical gimbal is simple, cheap, and the source of most of the maintenance.

Cost per unit of aperture is the great advantage. A large rotating dish delivers gain that would be expensive to reproduce with an array, so mechanically scanned radars remain the cheapest way to get long detection range. They are also simple to maintain with ordinary technicians. The weaknesses are revisit rate and flexibility: the beam is where the antenna points, so track updates come once per rotation, and a maneuvering or fast target can be lost between looks. The radar also cannot interleave functions, and the mechanical drive is the component that wears out.

Pick a mechanically scanned radar for wide-area surveillance where targets are cooperative or slow and the budget matters: air traffic control, weather, coastal surveillance, and long-range early warning. If you need to engage maneuvering targets or maintain many simultaneous tracks, an electronically scanned array is necessary. A common and sensible architecture pairs the two — a cheap rotating radar for volume search cueing an expensive electronically scanned set for engagement — which gets most of the coverage benefit at a fraction of the array cost.

Rotation 10–15 rpm, so a full 360° picture every 4–6 seconds · detection roughly 100 km for airport surveillance sets and up to 450 km for long-range early warning · one track update per rotation · antenna 4–12 m across on long-range sets · beamwidth typically 1–2° · emplacement and displacement in tens of minutes for tactical sets.

Slow revisit rate is the exploitable weakness: a target that maneuvers between looks breaks track. Terrain masking and low-altitude flight defeat ground-based sets. Noise and deception jamming are effective because the waveform is predictable. Anti-radiation missiles home easily on a continuously rotating high-power emitter, which is why these radars now come with rapid displacement requirements measured in minutes.

AN/TPS-77 and TPS-75 tactical radars, the AN/SPS-49 naval air search radar, ASR-9 and ARSR airport surveillance radars, the Russian P-18 and Kasta families, and the rotating back-to-back arrays on AWACS aircraft, which combine a mechanical scan in azimuth with electronic scan in elevation.

This is the commodity end of radar. Prices per unit of detection range are far below phased arrays, the supplier base is broad and includes non-traditional vendors, and the technology diffuses easily. Ground-based early warning radars of this type are widely exported and widely copied. Their continued relevance rests on the same economics that keep cheap sensors in every field: adequate performance at a price that lets you buy enough of them to cover the area.

13 minHow Radar Works | Start Learning About EW HereThe Ops Center By Mike Solyom · 100k+ views

4 minPrimary Radar or Primary Surveillance Radar (PSR)Lets Learn Aviation · 10k+ views

15 minVintage Furuno Marine Radar Scanner AutopsyAussie50 · 10k+ views

Radar Principle (Radar Tutorial) · Airport Surveillance Radar (ASR-11) (Federal Aviation Administration)

Over-the-horizon radar operates in the high-frequency band, 3–30 MHz, where signals either refract off the ionosphere (skywave) or follow the conductive sea surface (surface wave). Skywave systems detect targets 1,000–3,000 km away, far beyond the line-of-sight horizon that limits microwave radar. The wavelength is 10–100 meters, so resolution is terrible — a track is a blob tens of kilometers across — but the same long wavelength means aircraft are comparable in size to the wave and stealth shaping does very little. These are strategic wide-area warning sensors, not fire-control radars.

Coverage and counter-stealth performance are the reasons OTH radar exists. One installation watches millions of square kilometers, and low-observable shaping designed against centimetric radar has little effect at HF. Cost per square kilometer covered is unbeatable. The weaknesses are severe: resolution and accuracy are measured in tens of kilometers, the ionosphere changes with the diurnal cycle, season, and solar activity so performance varies hour to hour, there is a skip zone from about 0 to 1,000 km where the radar sees nothing, and the transmit arrays are hundreds of meters to kilometers long and completely fixed.

Pick OTH radar for strategic early warning over ocean and air approaches at continental scale, and for maritime domain awareness of surface traffic. It gives warning and cueing, never engagement quality. If you need to intercept something, the OTH track has to hand off to a microwave sensor or an aircraft. If the requirement is regional rather than continental, a network of conventional radars will be cheaper and far more accurate. Treat ionospheric variability as a planning constraint: coverage maps have to be computed per hour and per season, not stated as a single number.

HF band 3–30 MHz, wavelength 10–100 m · skywave detection 1,000–3,000 km · skip zone from 0 to about 1,000 km with no coverage · track accuracy in tens of kilometers · transmit arrays hundreds of meters to several kilometers long · system cost from hundreds of millions to low billions of dollars.

Ionospheric disturbances from solar activity degrade or black out skywave paths. HF noise jamming is straightforward for a well-resourced opponent, since the band is crowded and receivers are sensitive. Flying under the skip zone or inside coverage gaps avoids detection. The very large fixed installations are also easy to target, which is a strategic vulnerability rather than a technical one.

Australia's Jindalee Operational Radar Network, the US Navy ROTHR sites used for counter-narcotics surveillance, the Russian Container 29B6 and Podsolnukh surface-wave systems, and Chinese OTH installations. The Soviet Duga systems of the 1970s are the well-known early example.

These are national infrastructure projects costing hundreds of millions to low billions, with decades-long service lives and small operator communities. Renewed interest is driven by long-range anti-ship and hypersonic threats, where early warning at 2,000 km is worth a great deal even at low accuracy. Australia's expansion of JORN and its agreement to supply OTH technology to other countries has made this one of the few areas where a middle power holds a genuine technological lead.

17 minUnderstanding Over the Horizon RadarRohde & Schwarz · 10k+ views

20 minUnderstanding HF PropagationRohde & Schwarz · 100k+ views

11 minCobra Mist - The USA's Failed Over The Horizon RadarRingway Manchester · 100k+ views

Waves and Frequency Ranges (Radar Tutorial) · Jindalee Over-the-Horizon Radar (Engineers Australia)

Radars operating in the VHF (30–300 MHz) and UHF (300 MHz–1 GHz) bands use wavelengths of 0.3–10 meters. Stealth aircraft are shaped and coated to defeat centimetric radar, and both techniques lose effectiveness when the wavelength approaches the size of the aircraft's features. At VHF, edges, control surfaces, and fins enter the resonance region where scattering rises sharply, and radar-absorbent coatings tuned for X-band are largely transparent. The result is that a VHF radar detects a low-observable aircraft at a meaningful fraction of the range it would detect a conventional one.

Detection of low-observable targets is the whole point, and it works. VHF radars also propagate well, are relatively cheap per unit of range, and suffer less from rain. The weaknesses are resolution and interference. Long wavelengths need very large antennas for modest angular resolution, so accuracy is poor — typically kilometers, not meters — and that is not good enough to shoot on. The bands are crowded with broadcast and communications users, so interference and spectrum allocation are real problems. Ground clutter and multipath at low elevation angles are worse than at microwave frequencies.

Pick VHF/UHF radar as the cueing layer of an integrated air defense system: it tells you that something low-observable is out there and roughly where, and then a microwave fire-control radar or an infrared search and track system is cued into a small volume to get an engagement-quality track. Do not expect it to close the kill chain by itself. If the threat set does not include low-observable aircraft, the accuracy penalty is not worth accepting and a conventional S-band or L-band surveillance radar is better in every respect.

VHF 30–300 MHz and UHF 300 MHz–1 GHz, wavelengths 0.3–10 m · detection 300–600 km against conventional aircraft on large sets, on vendor figures · track accuracy typically kilometers, not meters · revisit every 5–20 seconds on rotating sets · antenna arrays tens of meters across · system cost in the low tens of millions of dollars.

Poor accuracy means detection does not equal engagement, so the counter is to break the handoff to fire-control sensors — jamming the microwave layer, or attacking it. Standoff jamming in a crowded band is easier than at microwave. Terrain masking still works, and cruise missiles flying low remain difficult targets. Anti-radiation missiles are effective against the large fixed emitters, though wideband seekers covering VHF are less common than X-band ones.

The Russian Nebo-M and Nebo-SVU VHF arrays, the P-18 and its digital successors, Chinese JY-27A and YLC-8B, the Czech VERA-family pairing with VHF sensors, and Serbia's use of a P-18 in the 1999 downing of an F-117, which is the standard case study for this whole category.

VHF radar is comparatively cheap to build and widely exported, which makes it one of the more accessible counters to expensive stealth aircraft — an asymmetry that features heavily in air-defense debates. Costs run in the low tens of millions per system against hundreds of millions for the aircraft they detect. The counter-argument, which is also correct, is that detection is not engagement, and closing the gap between them requires the whole rest of an integrated air defense network.

13 minRADAR Cross Section of Target (Rayleigh Region, Mie or Resonance Region & Optical Region) ExplainedEngineering Funda · 10k+ views

5 minCan Russia and China Detect the F-35 Stealth Aircraft?Covert Cabal · 1m+ views

Waves and Frequency Ranges (Radar Tutorial) · The Radar Game: Understanding Stealth and Aircraft Survivability (Mitchell Institute for Aerospace Studies)

Synthetic aperture radar uses the motion of the platform to synthesize an antenna far longer than the physical one. As the aircraft or satellite flies, it collects returns from the same patch of ground over several seconds and coherently combines them, producing along-track resolution set by the antenna's physical length rather than by its beamwidth. Resolution of 0.3–1 m from a satellite 500 km up is routine. Because it is radar, it images through cloud, smoke, and darkness, which is what makes it complementary to optical imagery rather than a substitute for it.

All-weather, day-night imaging at high resolution is the advantage, and coherent processing enables techniques optical sensors cannot do: interferometry to measure ground displacement to millimeters, and coherent change detection to see that a vehicle moved or soil was disturbed between two passes. Moving-target indication runs alongside imaging. The weaknesses are interpretation and geometry. SAR imagery looks nothing like a photograph and needs trained analysts or trained models. Layover and shadowing distort terrain. Processing is compute-heavy, and the platform's motion must be known precisely, so SAR depends on a high-quality inertial navigation solution.

Pick SAR when you need imagery regardless of weather or daylight, when you need to detect change over time, or when measuring ground deformation matters. If the sky is clear and you need visual identification for a human analyst, electro-optical imagery is cheaper and easier to interpret. If you need persistent stare rather than periodic revisit, neither works well and you want a loitering platform. The strongest current use is coherent change detection over routes and sites, which reveals disturbance that no optical sensor would show.

Resolution 0.3–1 m from a satellite at 500 km · swath 5–10 km in spotlight mode and 100–400 km in wide-swath mode · X-band near 9.6 GHz for resolution, C-band and L-band for penetration · interferometric displacement measurement to a few millimeters · revisit hours to days, set by constellation size · commercial imagery roughly a few hundred to a few thousand dollars per scene.

SAR is coherent, which makes it vulnerable to repeater jamming that injects false scenes or smears the image. Camouflage tuned to radar bands, corner-reflector decoys, and simply parking under structures all work. Because the technique needs a stable collection geometry, evasive platform motion degrades it. Timing the movement of equipment to fall between known satellite passes remains the most widely used counter, and publicly available orbital elements make it easy.

Sentinel-1 (open data, the workhorse of civil SAR), Capella Space and ICEYE small satellite constellations, the US Lacrosse/Onyx satellites, the AN/APG-81's SAR modes on the F-35, Global Hawk's ASARS-2 and the Joint STARS radar's combined SAR and moving-target modes.

SAR was the technology that most clearly changed with smallsats. ICEYE and Capella built sub-meter-capable satellites at a small fraction of traditional cost, and commercial SAR imagery is now bought routinely by defense and insurance customers alike. Prices per image have fallen by an order of magnitude in a decade. The remaining moats are the constellation size that determines revisit rate, and the analytic layer that turns a hard-to-read image into an answer.

5 minSynthetic Aperture Radar (SAR) ExplainedQinetiQ · 50k+ views

16 minSatellites Use 'This Weird Trick' To See More Than They Should - Synthetic Aperture Radar Explained.Scott Manley · 1m+ views

23 minHow Radar Satellites See through Clouds (Synthetic Aperture Radar Explained)Max Lenormand · 100k+ views

Synthetic Aperture Radar (NASA Earthdata) · Spaceborne Synthetic Aperture Radar: Principles, Data Access, and Basic Processing Techniques (NASA SERVIR)

Passive coherent location uses somebody else's transmitters as its illuminator. A receiver picks up the direct signal from an FM radio station, a digital television transmitter, or a cellular base station, and also picks up the same signal scattered off an aircraft. Correlating the two gives bistatic range and Doppler, and several receivers together produce a track. The system transmits nothing at all. Because these transmitters mostly operate in the VHF and UHF bands, passive radar inherits the counter-stealth advantage of long wavelengths as a side effect.

Emitting nothing is the decisive advantage: there is no signal to detect, no signal to home an anti-radiation missile on, and no spectrum allocation to obtain. Equipment is cheap, mostly commercial receivers and computers, and it can be deployed covertly. Bistatic geometry also produces returns from aspects a monostatic radar never sees, which helps against shaping-based stealth. The weaknesses are dependence and geometry: you get the transmitters you get, coverage is dictated by their locations and power, accuracy is worse than an active radar, and processing is demanding because the direct-path signal is far stronger than the echo.

Pick passive radar for covert surveillance, for gap-filling in an integrated air defense network, and for situations where emitting would be dangerous or is not permitted. It is also increasingly used for civil purposes like drone detection near airports, where spectrum licensing for an active radar is difficult. If you need engagement-quality tracks or coverage over ocean where no transmitters exist, an active radar is required. A reasonable architecture uses passive receivers as the always-on layer and turns active radars on only when a track needs refining.

Illuminators in the FM band at 88–108 MHz, DVB-T at 470–790 MHz, and cellular bands · detection typically 50–200 km, set by transmitter power and geometry · track accuracy in the hundreds of meters · update rate around 1 second · zero transmit power · receiver hardware in the thousands to tens of thousands of dollars per node.

Turning off or destroying the illuminators removes the radar, which is one reason broadcast infrastructure is an early target in conflicts. Coverage holes exist wherever transmitters do not. Accuracy is usually insufficient for engagement without handoff. The processing chain can be stressed by clutter and multipath in urban terrain, and low-altitude targets remain hard for the same reasons they are hard for any ground-based sensor.

Hensoldt TwInvis, Lockheed Martin Silent Sentry, Thales Homeland Alerter, the Czech VERA-NG (a related emitter-locating system rather than a true PCL), Chinese DWL002, and several university and startup drone-detection systems using DVB-T illumination.

Hardware costs are low — software-defined receivers and commodity compute — so the barrier is signal processing expertise rather than manufacturing. That has let smaller companies and national research institutes field credible systems, and it makes passive radar unusually proliferation-friendly. The commercial market is growing around counter-drone and airport applications where an active radar's licensing and interference problems are prohibitive. Expect capability to track improvements in general-purpose computing more closely than in radar hardware.

9 minUsing passive radars and satellite signals to detect and identify airborne threatsNATO Science & Technology Organization · 10k+ views

7 minReal time passive radar at home30hours · 100k+ views

13 minFM Radio Passive Radar, WWLI 105.1 MHzJuha Vierinen · 10k+ views

Radar Principle (Radar Tutorial) · DVB-T based Passive Bistatic Radar (Norwegian Defence Research Establishment)

Millimeter-wave radar operates from roughly 30 to 100 GHz, most often at 35 GHz (Ka-band) or 94 GHz (W-band). The short wavelength allows a narrow beam from a small antenna, which is exactly what a missile seeker or a small vehicle radar needs — an aperture a few centimeters across gives an angular resolution that would require a meter-scale dish at X-band. Resolution is good enough to discriminate a tank from a truck by its return profile. Atmospheric absorption is the price: oxygen and water vapor attenuate these frequencies sharply, so useful range is a few kilometers to a few tens of kilometers.

The combination of small aperture, fine resolution, and all-weather operation is what makes mmWave the standard for autonomous terminal seekers and for helicopter fire-control radar. It penetrates dust, smoke, and light foliage that defeat infrared. It also gives direct velocity measurement through Doppler. The weaknesses are range and rain. Attenuation at 94 GHz runs several dB per kilometer in clear air and much worse in heavy rain, so this is inherently a short-range sensor. Components have historically been expensive, though automotive volume at 77 GHz has changed that dramatically.

Pick mmWave radar for terminal seekers on anti-armor weapons, for helicopter mast-mounted fire control, for automotive and robotic sensing, and for short-range counter-drone tracking. If range beyond about 30 km is needed, drop to X-band or lower. If you need to classify targets visually or confirm identity, pair it with an infrared imager — the standard dual-mode seeker does exactly this, using mmWave to find the target through obscurants and imaging infrared to confirm the aimpoint. Automotive silicon has made this pairing affordable in a way it was not fifteen years ago.

30–100 GHz, most often 35 GHz, 77 GHz, or 94 GHz · antenna a few centimeters across for a beamwidth that would need a meter-class dish at X-band · useful range a few kilometers to a few tens of kilometers · clear-air attenuation of several dB per km at 94 GHz, much worse in heavy rain · range resolution from a few centimeters to tens of centimeters, set by bandwidth · automotive radar-on-chip devices at tens of dollars in volume.

Heavy rain cuts range severely. Chaff cut for the band works. Corner reflectors and decoys are effective because the seeker's discrimination logic is simple by necessity. Because the beam is narrow, a small pointing error or a late target maneuver can put the target outside the acquisition basket. Jamming is harder than at lower frequencies mostly because few opponents field mmWave jammers, which is an economic obstacle rather than a physical one.

The Longbow radar on the AH-64D/E and the Longbow Hellfire seeker, the Brimstone dual-mode seeker, SADARM and BONUS submunition sensors, automotive 77 GHz radar from Bosch and Continental, and counter-UAS tracking radars from a number of suppliers.

Automotive demand transformed this segment. Radar-on-chip devices from TI, NXP, and Infineon at 77 GHz cost tens of dollars in volume, and defense applications now ride that supply chain for components even when the system is bespoke. That has cut seeker sensor costs and made mmWave viable in cheap munitions and counter-drone systems. The remaining defense-specific cost is in the antenna, packaging, and the discrimination software, not in the RF silicon.

2 minWhat is mmWave sensing ? | Mouser Electronics | Texas InstrumentsMouser · 50k+ views

7 minWhat is FMCW Radar and why is it useful?Marshall Bruner · 100k+ views

18 minFMCW Radar for Autonomous Vehicles | Understanding Radar PrinciplesMATLAB · 100k+ views

The Fundamentals of Millimeter Wave Sensors (Texas Instruments) · Recommendation ITU-R P.676: Attenuation by atmospheric gases (ITU)

Electro-optical imaging is a visible-band camera with a long lens, stabilized on a gimbal. Resolution is set by the aperture and the diffraction limit: a 1 m telescope on a satellite resolves about 30 cm on the ground from 500 km, which is why the 30 cm-class commercial imagers carry roughly meter-class optics. On an aircraft, a 100–500 mm lens on a stabilized turret reads a license plate at a few kilometers. The technology itself is commodity — the same sensors that go in phones and cinema cameras — and the engineering value sits in the stabilization, the optics, and the processing that turns a shaky long-focal-length image into something usable.

Interpretability is the great advantage. A visible image is immediately understandable by a human and by every off-the-shelf computer vision model, which is not true of radar or infrared. Passive operation means no emissions to detect. Cost is low and falling. The weaknesses are equally obvious: no imagery at night without illumination, and none through cloud, fog, smoke, or dust. Haze cuts contrast at long slant ranges. Atmospheric turbulence limits resolution regardless of aperture on long horizontal paths, which is why ground-to-ground imaging past 10–20 km rarely works well no matter how good the optics are.

Pick electro-optical imaging as the default identification sensor whenever there is daylight and reasonable weather, and for any application where a human has to make a positive identification. Pair it with thermal imaging rather than choosing between them — nearly every modern turret carries both, because the failure conditions barely overlap. If the requirement is all-weather or all-hours coverage, radar has to be in the architecture. For satellite imagery, check revisit rate and cloud statistics for the target area before assuming an optical constellation solves the problem.

Satellite ground resolution about 30 cm from 500 km with a 1 m aperture · airborne turrets with 100–500 mm lenses read a license plate at a few kilometers · ground-to-ground imaging limited by turbulence past 10–20 km · turret mass from under 1 kg on small drones to 40–50 kg on large ISR turrets · turret cost from a few thousand dollars to over $100k.

Darkness, cloud, fog, and smoke are the ordinary defeats and they are cheap to arrange — smoke generators and multispectral obscurants are standard equipment. Camouflage and concealment work well against a visible-band sensor. Laser dazzlers can saturate or damage a sensor, and this is now widely fielded. At long slant ranges, atmospheric turbulence and haze reduce effective resolution well below the diffraction limit.

The MX-series and Wescam turrets on most ISR aircraft, the Sniper and Litening targeting pods, the Maxar and Planet imaging constellations, and the electro-optical sensors on nearly every drone from quadcopters upward.

This is the most commoditized sensing category in defense. Consumer and automotive image sensors set the price, and the defense premium sits almost entirely in the gimbal, the optics, and the export-controlled processing. Small turrets that cost $100k a decade ago now cost a fraction of that, which is what made cheap armed drones practical. The commercial imagery market has followed the same curve, with sub-meter imagery available on subscription and analysis rather than collection increasingly being where value accrues.

19 minThe Terrifying Technology Inside Drone CamerasNew Mind · 1m+ views

20 minI Couldn't Find a Video Explaining Satellite Images, So I Made OneMax Lenormand · 100k+ views

5 minThe Rayleigh Criterion | Diffraction-Limited Vision and Photography | Doc PhysicsDoc Schuster · 50k+ views

Earth observation data basics (NASA Earthdata) · National Security Space: Actions Needed to Better Use Commercial Satellite Imagery and Analytics (GAO)

Mid-wave infrared imaging covers 3–5 µm, the band where hot objects — engine exhausts, gun barrels, vehicle powerpacks — radiate strongly and where the atmosphere has a good transmission window. Detectors are photon detectors, usually mercury cadmium telluride or indium antimonide, and they must be cooled to about 77 K by a Stirling cryocooler to keep thermal noise below the signal. That cooler is the defining engineering burden: it costs money, draws power, takes several minutes to reach temperature, vibrates, and is usually the first thing to fail.

Sensitivity and range are the reasons to accept the cooler. A cooled MWIR system detects a vehicle-sized target at two to five times the range of an uncooled LWIR camera of similar aperture, and its noise-equivalent temperature difference of 20 mK or better resolves detail that uncooled sensors miss. It works in total darkness and sees through most battlefield smoke. The weaknesses are cost, cooldown time, cooler life measured in a few thousand hours, and export control — cooled thermal imagers are among the most tightly restricted items in the ITAR and Wassenaar regimes.

Pick cooled MWIR when detection range or thermal sensitivity is the requirement: targeting pods, long-range surveillance, infrared search and track, and missile seekers. If the application is short-range and cost-sensitive — a soldier's weapon sight, a small drone, a driver's viewer — uncooled LWIR gives adequate performance at a tenth of the price with no cooldown delay. A useful rule of thumb is that cooled sensors are worth it past about 3–5 km against vehicle targets, and rarely below that.

3–5 µm band · detector cooled to about 77 K, with several minutes of cooldown from a Stirling cooler · noise-equivalent temperature difference of 20 mK or better · arrays typically 640×512 to 1280×1024 · vehicle detection at two to five times the range of a comparable uncooled camera, which pays off past roughly 3–5 km · cooler life of a few thousand hours · detector cost in the tens of thousands of dollars.

Thermal signature management works: exhaust cooling, thermal blankets, and simply keeping engines off. Multispectral smoke that blocks both visible and infrared bands is standard issue in mechanized forces. Water vapor and heavy rain absorb strongly in the MWIR band. Flares defeat older non-imaging seekers, and directed infrared countermeasures using modulated lasers defeat many seeker types. Thermal crossover — the period at dawn and dusk when target and background reach the same temperature — degrades contrast for everyone.

The Sniper ATP and Litening targeting pods, Raytheon's third-generation FLIR on the Abrams and Bradley, IRST systems like the Legion pod and the OLS-35, the sensors in Javelin and imaging-infrared air-to-air seekers, and the MX-20 and similar high-end turrets.

Cooled detector arrays are made by a handful of suppliers — Leonardo DRS, Teledyne FLIR, Lynred, Sofradir-derived lines, and a small number of Chinese and Israeli producers — and prices remain in the tens of thousands of dollars per detector at best. Export control keeps the market segmented and margins high. The interesting trend is that improvements in uncooled technology keep pushing the crossover range upward, which slowly erodes the cooled market from below.

2 minWhat's the difference between cooled & uncooled thermal detectors?Teledyne FLIR · 10k+ views

3 minHow Does a Sunpower Stirling Cryocooler Work? Sunpower Free-Piston Stirling Cryocooler AnimationSunpower, Inc. · 100k+ views

5 minTech Talk: MWIR & LWIR OverviewSierra-Olympia Technologies · 10k+ views

Research and science (Teledyne FLIR) · Fundamentals of Infrared Detector Materials (SPIE)

Uncooled long-wave infrared imaging uses a microbolometer array — a grid of tiny thermally isolated resistors whose resistance changes as incident infrared radiation warms them. There is no cryocooler, no cooldown, and no vacuum-encapsulated cold finger to fail. Sensitivity is lower than a cooled photon detector, with noise-equivalent temperature difference typically 30–50 mK against 20 mK or better for cooled MWIR, but the sensor turns on in under a second, draws a watt or two, and costs a few hundred dollars at volume.

Cost, power, size, and instant-on operation are the advantages, and they compound: an uncooled core small enough for a handheld sight is also small enough for a 250 g drone. The 8–14 µm band works well against ambient-temperature targets like people, and it penetrates smoke and haze better than MWIR. The weaknesses are sensitivity and range. Detection range against vehicles is typically a third to a half of a comparable cooled system, resolution is usually 640×512 or below at accessible prices, and the sensor needs periodic flat-field correction, which produces a brief shutter interruption in the image.

Pick uncooled LWIR for anything short-range and cost-constrained: weapon sights, driver's viewers, handheld observation devices, small drone payloads, and industrial or firefighting thermal cameras. If detection past about 3–5 km against vehicles matters, or if you need the sensitivity for a seeker, use cooled MWIR. The most common design mistake here is specifying resolution and ignoring optics — a 640×512 core with a wide lens has poor range performance regardless of the detector, and focal length usually matters more than pixel count.

8–14 µm band · microbolometer arrays typically 320×256 to 640×512 at accessible prices · noise-equivalent temperature difference 30–50 mK · startup under a second, with no cryocooler · power draw of 1–2 W · vehicle detection at a third to a half the range of a comparable cooled system · cores under $500 at volume.

Thermal camouflage and signature management work, and multispectral smoke defeats it. Glass is opaque in the LWIR band, so anything behind a window is invisible. Rain and fog attenuate. The sensor's periodic non-uniformity correction blinks the image for a fraction of a second, which is a minor tactical annoyance. Because sensitivity is modest, thermal contrast between target and background at dawn and dusk can drop below the detection threshold entirely.

Teledyne FLIR Boson and Lepton cores, Lynred and Guide Sensmart microbolometers, thermal sights on infantry weapons, the payloads on DJI and military small drones, and the thermal channels in most automotive night vision systems.

Microbolometer prices have fallen roughly an order of magnitude in fifteen years, and Chinese suppliers now compete hard at the low end, which has both cut prices and created a supply-chain policy problem for Western buyers. Cores under $500 have made thermal imaging standard on small drones, and that in turn is one of the reasons cheap drones became so effective at night. Export control still applies above certain frame rate and resolution thresholds, and those thresholds are increasingly out of step with what is commercially available.

4 minAsk an Expert: What is a Microbolometer?Sierra-Olympia Technologies · 10k+ views

15 min#1486 Micro Bolometer Thermal Camera SensorIMSAI Guy · 5k+ views

3 minHow Do Thermal Imaging Goggles Work?SciShow · 100k+ views

Research and science (Teledyne FLIR) · Fundamentals of Infrared Detector Materials (SPIE)

Short-wave infrared covers roughly 0.9–1.7 µm, between visible light and thermal infrared. It is reflected radiation rather than emitted heat, so a SWIR image looks like a photograph rather than a thermal map, but it exploits nightglow — the faint airglow from the upper atmosphere that is much brighter in SWIR than in the visible band — so it produces useful imagery on a moonless night without any illumination. Detectors are usually indium gallium arsenide arrays, which work at or near room temperature with modest thermoelectric cooling.

SWIR sees through haze, fog, and some smoke far better than visible light, because scattering falls with wavelength. It images the 1,064 nm and 1,550 nm laser wavelengths used by designators and rangefinders, so it can see laser spots that no other imaging band shows. It also distinguishes materials that look identical in the visible band, including wet versus dry surfaces and some camouflage against real foliage. The weaknesses are cost and resolution: InGaAs arrays are expensive, typically $10k–50k, resolution is usually lower than visible sensors, and the band is nearly useless in total darkness indoors where there is no airglow.

Pick SWIR when you need to see laser designation spots, image through haze, or discriminate materials that defeat visible-band camouflage. It is the standard choice for laser spot trackers and increasingly for covert illumination systems, since a 1,550 nm illuminator is invisible to both the naked eye and conventional night vision goggles. If cost matters and the requirement is simply night vision, image-intensified goggles or uncooled thermal are cheaper by an order of magnitude. In practice SWIR usually appears as the third channel in a high-end turret rather than as a sensor on its own.

0.9–1.7 µm band · indium gallium arsenide detectors at or near room temperature with modest thermoelectric cooling · images the 1,064 nm and 1,550 nm laser lines used by designators and rangefinders · arrays typically 320×256 to 1280×1024 · camera cost $10k–50k.

Heavy cloud and rain still block it. It is a reflective band, so a target in genuine darkness with no airglow — indoors, under cover — is invisible. Camouflage designed with SWIR reflectance in mind, which modern multispectral camouflage materials include, works. As with all optical sensors, laser dazzling and damage are possible, and the same InGaAs sensitivity that makes it good at seeing lasers makes it easy to saturate.

Laser spot trackers on targeting pods, the SWIR channels in MX-series and Wescam turrets, Sensors Unlimited and Xenics InGaAs cameras, hyperspectral instruments that use SWIR bands for material identification, and covert 1,550 nm illuminators used with SWIR viewers.

InGaAs detector production is small and specialized, with a handful of suppliers, and prices have come down slowly compared with silicon or microbolometer arrays. There is meaningful non-defense demand from semiconductor inspection, agricultural sorting, and solar-cell testing, which helps sustain the supply base. Newer quantum-dot and germanium-on-silicon approaches promise much cheaper SWIR sensing and are the thing to watch, since a $500 SWIR sensor would change several product categories at once.

5 minInGaAs Cameras - TechBites SeriesHAMAMATSU PHOTONICS · 1k+ views

1 minSWIR Imaging Enhances Fire Zone Visibility: Visible Spectrum, NIR & LWIR Comparison | Sony OfficialSony · 10k+ views

4 minWhen And Why You Should Use SWIRPhotonicsOnline · 1k+ views

Fundamentals of Infrared Detector Materials (SPIE) · NIR and SWIR Questions and Answers (Hamamatsu Photonics)

An infrared search and track system is a wide-field cooled infrared sensor that scans a large volume, detects point-source targets against the sky background, and maintains tracks on them without transmitting anything. Aircraft are detectable in the infrared from their engine exhaust, from skin friction heating at speed, and from the plume, and none of these are addressed by radar-focused stealth design. The system produces very accurate angular data but no range, so a single IRST gives a bearing rather than a firing solution unless it is combined with a second sensor or with kinematic ranging over time.

Passive operation is the point: an IRST detects without revealing the platform, and it is largely unaffected by radar-band stealth measures. Angular accuracy is excellent, better than radar, which makes it a good cueing sensor and a good way to keep a track when radar is jammed. The weaknesses are weather, range ambiguity, and clutter. Cloud and moisture cut range severely, so performance varies enormously with conditions. Getting range requires either triangulation between platforms, a laser rangefinder, or several seconds of kinematic observation. Background clutter over land is far worse than over the sea.

Pick IRST for air-to-air detection of low-observable targets, for passive engagement in a jammed environment, and for any situation where emitting is unacceptable. Use it alongside radar rather than instead of it: the standard doctrine is that radar-band stealth pushes an opponent toward infrared, and infrared's weather dependence pushes them back toward radar, so a capable force carries both. If the theatre is persistently cloudy, discount IRST range claims heavily, since vendor figures usually assume clear conditions and a hot target aspect.

Cooled infrared detector scanning roughly ±60–90° in azimuth · detection 50–100 km against fighter-sized targets in clear air on vendor figures, far less head-on or through cloud · angular accuracy well under 1 mrad, better than radar · no range output, so kinematic ranging needs several seconds of observation · scan revisit of about 1–2 seconds over the search volume · system cost in the millions of dollars.

Weather is the primary defeat, and it is free. Infrared signature reduction on engines and nozzles — a design feature of every recent combat aircraft — cuts detection range. Flying slow and cold reduces skin heating. Cloud layers can be used deliberately for masking. Against the sensor itself, directed infrared countermeasures and dazzling lasers work, though against a wide-field search sensor they also announce your presence.

Lockheed Martin's Legion Pod on the F-15 and F-16, the F-35's Distributed Aperture System and Electro-Optical Targeting System, the OLS-35 on the Su-35 and OLS-UEM on the Su-57, PIRATE on the Eurofighter, and the OSF on the Rafale.

IRST is a high-value, low-volume segment dominated by the same suppliers as cooled thermal imaging. Its strategic interest has risen sharply because it is the most credible passive counter to low-observable aircraft, which makes it a direct hedge against stealth investments. Costs run in the millions per system. The technical race is in detector format and scan rate — bigger arrays covering more sky per second — and in the algorithms that pull point targets out of cluttered backgrounds, which is increasingly a machine-learning problem.

10 minThe Stealth Buster - IRSTCovert Cabal · 100k+ views

16 minThe Other RADAR | Can IRST Infrared Detection Replace Radar?Millennium 7 * HistoryTech · 50k+ views

9 minFLIR & IRST: What is the difference?Millennium 7 * HistoryTech · 10k+ views

Fundamentals of Infrared Detector Materials (SPIE) · Infrared Dim Small Target Detection Networks: A Review (Sensors via PMC)

A hyperspectral imager records a continuous spectrum — typically 100 to 400 contiguous bands from the visible into the shortwave infrared — for every pixel in the scene. Every material has a spectral reflectance signature set by its molecular composition, so the resulting data cube lets an analyst identify what something is made of rather than just what shape it is. Multispectral imagers do the same thing with 4 to 15 broad bands, which is enough for vegetation health and land classification but not enough to separate materials with similar colors.

Material identification is the unique capability. Hyperspectral data separates real vegetation from green paint, identifies disturbed soil, finds specific minerals, detects chemical plumes, and picks camouflage out of a background that looks identical in a photograph. It works passively at long range. The weaknesses are data volume, atmosphere, and resolution. A single scene is gigabytes, so onboard processing or aggressive downlink management is mandatory. Atmospheric correction is required before signatures mean anything. And there is a hard trade between spectral and spatial resolution, so hyperspectral imagery is usually much coarser spatially than a panchromatic image from the same platform.

Pick hyperspectral imaging when the question is what something is made of: camouflage detection, chemical and effluent detection, mineral exploration, crop and environmental monitoring, and detecting disturbed earth. If the question is where something is or what shape it has, ordinary imagery is far cheaper and easier. The realistic operational pattern is tip-and-cue — a wide-area imager finds candidates, hyperspectral confirms material, and a high-resolution optical sensor gets the picture a human can act on.

100–400 contiguous bands from roughly 400 to 2,500 nm, against 4–15 broad bands for multispectral · spectral sampling of about 5–10 nm · spatial resolution around 30 m from current civil satellites and 1–5 m from airborne instruments · data volume of gigabytes per scene · civil hyperspectral satellite missions costing roughly $100M–400M.

Spectrally matched camouflage materials are the direct counter, and they exist, though matching across 200 bands is much harder than matching visible color. Cloud blocks it entirely. Coarse spatial resolution means small objects are subpixel, and their signature gets mixed with the background, which is the practical limit for most tactical uses. Poor atmospheric correction produces false identifications, so the analytic chain is as much a vulnerability as the sensor.

NASA's AVIRIS airborne instrument and the EMIT instrument on the ISS, the Italian PRISMA and German EnMAP satellites, Planet's Tanager and Pixxel's commercial hyperspectral constellations, the ARTEMIS sensor on TacSat-3, and airborne systems used for chemical agent detection.

Commercial hyperspectral constellations are being built now by several small companies, betting that methane monitoring, agriculture, and mineral exploration will fund the capability that defense also wants. Methane detection in particular has become a real revenue driver because of emissions regulation. Sensor cost is high and data handling is harder than imaging, so the value has migrated to analytics — very few customers want a data cube, they want an answer about a specific material at a specific place.

6 minMapping the Invisible: Introduction to Spectral Remote SensingNEON Science · 100k+ views

3 minWhat is hyperspectral imaging - TutorialSpecimSpectral · 50k+ views

3 minHyperspectral and Multispectral Imaging - TRENDING IN OPTICSEdmund Optics · 10k+ views

Earth observation data basics (NASA Earthdata) · EMIT imaging spectrometer (NASA JPL)

A laser rangefinder fires a short pulse, usually at 1,064 nm or the eye-safer 1,550 nm, and times the return to measure range to a meter or better at 10–20 km. A designator does the same thing continuously, marking a target with a coded pulse train that a semi-active laser seeker can home on. The pulse repetition frequency is set by a shared code so that several designators can operate in the same area without confusing each other's weapons. Together with an inertial unit and a GPS fix, a rangefinder converts an angle into target coordinates, which is how most modern targeting works.

Precision is the advantage. Range accuracy of a meter and designation accuracy under a meter give laser-guided weapons their characteristic circular error of one to two meters, which is far better than GPS guidance alone. The equipment is compact and cheap enough for a dismounted team. The weaknesses are weather and exposure. Cloud, fog, smoke, and dust block the beam, and the designator has to keep the spot on the target through the weapon's whole terminal flight, which means someone or something must maintain line of sight and stay put. The beam is also detectable by laser warning receivers.

Pick laser designation when you need one to two meter accuracy against a specific aimpoint, when the target is moving, or when collateral damage constraints are tight. If the weather is poor or the target is stationary and its coordinates are known, GPS or inertial guidance works through cloud and does not require anyone to stay exposed. The common architecture uses both: a dual-mode weapon takes GPS guidance through the clouds and switches to laser terminal homing if a spot is available, which covers both failure modes.

1,064 nm, or 1,550 nm where eye safety matters · range accuracy of about 1 m at 10–20 km · designation pulse repetition frequency around 10–20 Hz, set by a shared code · laser-guided weapon circular error of 1–2 m · dismounted designators weighing roughly 3–6 kg · handheld rangefinders at a few thousand dollars, full designators at tens of thousands.

Smoke and obscurants block designation, and this is the most common practical counter. Laser warning receivers alert the target, which can then move or deploy smoke within seconds. Some vehicles carry systems that detect the designation and slew a countermeasure onto the source. Clouds below the aircraft break the geometry. Because designation requires continuous line of sight, simply driving behind a building can defeat a weapon already in flight.

The AN/PEQ-1 SOFLAM and LTD-2000 ground designators, designator channels in Sniper, Litening, and MX-series turrets, the seeker on every Paveway and Hellfire variant, and vehicle-mounted rangefinders in tank fire-control systems.

Rangefinders are commodity items now, with handheld units at a few thousand dollars, while full designators with the accuracy and coding needed for weapons remain export-controlled and cost tens of thousands. Fiber laser technology has cut size and power draw substantially. The larger economic effect is downstream: cheap accurate designation is what makes a $25k guided munition worth more than a $3k unguided one, and that ratio has driven the precision-weapons market for thirty years.

3 minHow does laser range finder work? Explained in 3 minutes.VROOK Learning · 10k+ views

4 minPrecision Targeting Device TrainingPEO Soldier · 10k+ views

20 minTechnology Explained: How the Paveway II Laser Guided bomb Seeker worksIron Physik · 5k+ views

Defense Primer: U.S. Precision-Guided Munitions (Congressional Research Service) · Laser rangefinder built with APDs and PLDs (Laser Components)

LiDAR builds a three-dimensional point cloud by firing many laser pulses and timing each return. Military LADAR systems extend the idea with two useful modes. Geiger-mode and single-photon detectors let a sensor collect returns fast enough to map large areas from an aircraft, and because a pulse that passes through a gap in foliage still returns from the ground, multiple-return processing produces bare-earth terrain maps under tree cover. Gated-viewing systems use a pulsed illuminator and a shuttered camera to image a specific range slice, cutting through fog and haze that defeat ordinary cameras.

Direct geometric measurement is the advantage. Point clouds give absolute dimensions, which makes them good for terrain mapping, change detection, target dimensioning, and navigation in GPS-denied environments. Foliage penetration is a genuinely unique capability. The weaknesses are range, weather, and cost. Atmospheric scattering limits practical range to a few kilometers for airborne systems, rain and fog cut it further, and shiny or absorbing surfaces return nothing. Automotive-driven cost declines have been real but military-grade long-range units remain expensive, and data volumes are large.

Pick LiDAR for terrain and site mapping, for foliage-penetrating survey, for autonomous navigation where measured geometry beats inferred depth, and for gated imaging through obscurants at short range. If the requirement is wide-area imaging at long range, SAR does it through weather and LiDAR does not. If the platform is a small drone that needs obstacle avoidance, cheap solid-state automotive LiDAR now costs a few hundred dollars and is usually the right answer. For archaeology and terrain intelligence under canopy, nothing else does the job.

Wavelengths of 905 nm or 1,550 nm · airborne mapping range of a few kilometers, automotive units to about 200 m · pulse rates of 100 kHz to a few MHz on airborne mapping systems, far higher in Geiger mode · point density 1–50 points per square meter and vertical accuracy 5–15 cm on survey work · automotive solid-state units at a few hundred dollars, airborne mapping systems at hundreds of thousands.

Fog, rain, dust, and smoke scatter the beam and limit range. Absorbing and specular surfaces return nothing. Because it is an active emitter at optical wavelengths, a laser warning receiver detects it. Camouflage netting that presents a broken surface degrades point-cloud interpretation. The practical limit in most military uses is simply range: a sensor that works well at 1 km and poorly at 5 km constrains the platform's standoff.

Airborne bathymetric and terrain LiDAR used for hydrographic survey, the ALIRT and Jigsaw foliage-penetrating systems, automotive units from Hesai, Luminar, and Ouster now appearing on ground robots, and the LiDAR mapping that has transformed archaeological survey in forested regions.

The automotive market drove LiDAR costs down by more than an order of magnitude over a decade and consolidated the supplier base around a few Chinese and US manufacturers. Defense buyers benefit from the component cost curve but face a supply-chain concentration problem, since the volume producers are mostly Chinese. High-end airborne mapping systems remain a separate, small, expensive market with different suppliers and no comparable cost decline.

4 minWhat is Lidar? How does Lidar work? Know all about LiDARGeospatial World · 500k+ views

7 minHow a LiDAR Drone Maps the Bare Earth through trees!Indiana Drones · 10k+ views

6 minLost beneath the leaves: Lasers reveal an ancient Amazonian civilisationnature video · 1m+ views

What is lidar? (NOAA National Ocean Service) · 3D Elevation Program (US Geological Survey)

A radar warning receiver listens across the radar bands, compares detected pulse trains against a stored threat library, and tells the crew which radars are looking at them and roughly where those radars are. Because radar energy travels out and back to the radar but only one way to the receiver, the echo the radar works with falls off as the fourth power of range while the signal reaching the RWR falls off only as the square. An RWR therefore detects a radar well before that radar detects the aircraft — a radar that sees you at 100 km is usually audible at 200 km or more. That asymmetry is the fundamental reason passive warning works.

The one-way advantage means warning generally arrives before engagement, which is the entire point. RWRs are passive, cheap relative to the aircraft, and directly drive the survivability chain: warn, maneuver, jam, dispense countermeasures. Weaknesses come from the threat library and from waveform trends. An unknown or reprogrammed emitter shows as ambiguous, and library updates are a continuous operational burden measured in days, not years. Low-probability-of-intercept radars using low peak power and spread waveforms are much harder to detect, and modern AESAs deliberately exploit this. Angular accuracy from a few small antennas is coarse, typically 10–30°.

An RWR is standard equipment on any aircraft expected to operate near threats, and the real decision is how much capability to buy. If the aircraft flies in a well-characterized threat environment, a basic receiver with a good library suffices. If it will face modern AESA and low-probability-of-intercept emitters, you need a digital receiver with wideband channelized processing, which costs several times as much. Treat the mission data file update process as part of the system: an excellent receiver with a stale library is a liability, because it will report a modern threat as unknown.

Warning range typically twice the threat radar's detection range or more, so 200 km-plus against a radar that sees you at 100 km · angular accuracy typically 10–30° · frequency coverage roughly 2–18 GHz on legacy sets, 0.5–40 GHz on wideband digital receivers · instantaneous bandwidth of hundreds of MHz to a few GHz on a digital receiver · mission data file update cycle measured in days.

Low-probability-of-intercept waveforms — low peak power, wide bandwidth, long integration — can put a radar below the RWR's detection threshold. Emitters not in the library produce ambiguous warnings. Dense signal environments cause processing overload and false alarms, which trains crews to distrust the display. Passive sensors and infrared-guided weapons trigger no warning at all, which is a growing gap as infrared search and track systems become common.

The AN/ALR-56M and ALR-69A on US fighters, the ALR-67(V)3 on the Super Hornet, Praetorian on the Eurofighter, SPECTRA on the Rafale, and the digital receivers integrated into the F-35's AN/ASQ-239 electronic warfare system.

RWRs sit inside larger electronic warfare suites and are increasingly indistinguishable from them, since the same wideband digital receivers serve warning, electronic support, and targeting. The technology has shifted decisively toward software-defined receivers, which moves the value from hardware to the signal-processing and library-generation pipeline. That pipeline — collecting emitter parameters, characterizing them, and pushing updates to the fleet quickly — is now a defining national capability and is largely invisible in procurement budgets.

3 minHENSOLDT Kalaetron RWR – Radar Warning ReceiverHENSOLDT · 10k+ views

4 minRadar Warning Receiver - RWR | Fighter Pilot Course | PilotCourse99Bemlish · 1k+ views

Introduction to Radar Warning Receivers (Georgia Tech Research Institute) · Defense Primer: Electronic Warfare (Congressional Research Service)

Electronic support measures detect, characterize, and locate emitters. Direction finding gives a bearing, and a bearing from two or more receivers gives a fix. Modern systems use time difference of arrival and frequency difference of arrival between spatially separated receivers, which produces geolocation accuracy far better than amplitude comparison — hundreds of meters rather than tens of kilometers, given adequate baselines and precise time synchronization. Everything is passive, so the emitting side gets no indication it has been located.

Passive geolocation of emitters is enormously valuable: it builds the electronic order of battle, cues other sensors, and produces targets for anti-radiation weapons. Range is limited only by line of sight and receiver sensitivity, so an airborne or space-based receiver covers a very large area. The weaknesses are that the target has to transmit — emission control defeats the whole approach — and that accuracy depends on geometry, requiring several well-separated platforms with precise time and position knowledge. Dense emitter environments make sorting and association hard, and modern low-probability-of-intercept waveforms are difficult to detect at all.

Use electronic support as the persistent, always-on layer of any surveillance architecture, because it costs nothing to leave running and it reveals what the opponent is doing. It is the standard cueing sensor for suppression of enemy air defenses. If the target set is deliberately silent, this approach contributes nothing and you need active or optical sensing. Multi-platform geolocation is where the accuracy comes from, so budget for the constellation and the time synchronization rather than for a single very good receiver.

Detection range limited by line of sight, roughly 400 km from an airborne or elevated receiver · amplitude-comparison bearing accuracy 10–30°, interferometer arrays 1–3° · TDOA and FDOA geolocation to hundreds of meters against tens of kilometers for single-platform methods · three or more receivers with time synchronization to a few nanoseconds · national system cost in the hundreds of millions.

Emission control is the complete counter: a radar that is off cannot be located. Operating briefly and shutting down before a fix converges works, which is why modern air-defense doctrine emphasizes short illumination periods and rapid displacement. Decoy emitters draw fixes onto empty ground cheaply. Low-probability-of-intercept and spread-spectrum waveforms reduce detection range substantially, and networked systems that share tracks over directional datalinks emit far less than legacy ones.

The Czech VERA-NG and Ukrainian Kolchuga passive tracking systems, the RC-135V/W Rivet Joint and EP-3E aircraft, the US Navy's Ocean Surveillance satellites and their successors, HawkEye 360's commercial RF geolocation constellation, and the ESM suites on most warships.

The most significant recent change is commercial: HawkEye 360, Kleos, Unseenlabs, and others sell RF geolocation from smallsat clusters, which puts a capability that was strictly national into a subscription product. Maritime applications — finding vessels that have turned off their transponders — drive most of the commercial revenue. Costs for a national system remain in the hundreds of millions, but the commercial layer now provides a useful and much cheaper baseline that many governments buy alongside their own.

19 minStefan Scholl, DC9ST: Introduction and Experiments on Transmitter Localization with TDOASoftware Defined Radio Academy · 10k+ views

14 minTrack Down Radio Transmitters / KrakenSDRsn0ren · 100k+ views

3 minIntroduction to TDOA geolocation with RFeye SiteCRFS - Extraordinary RF Technology · 5k+ views

An Introduction to Direction Finding Methodologies (Rohde & Schwarz) · Emitter Location with Azimuth and Elevation Measurements Using a Single Aerial Platform for Electronic Support Missions (Sensors via PMC)

Electronic intelligence is the collection and analysis of non-communications emissions — radars, beacons, telemetry, guidance links — to characterize what equipment an opponent has and how it operates. The output is a parametric description of each emitter: frequency agility pattern, pulse repetition interval, scan rate, modulation, and the intrapulse features that distinguish one manufacturer's set from another. That parametric library is what makes radar warning receivers, jammers, and anti-radiation seekers work, so ELINT is upstream of most of electronic warfare.

ELINT is entirely passive and produces information available no other way: you cannot deduce a radar's pulse agility pattern from a photograph of its antenna. Collection at long range from aircraft, ships, and satellites gives broad coverage. The weaknesses are that it depends on the opponent transmitting, ideally in an operationally realistic mode, and opponents know this and reserve their real modes for wartime. That produces a persistent intelligence gap: the peacetime library may describe training modes rather than combat modes, and discovering the difference during a conflict is the classic failure case.

ELINT is a continuous national activity rather than a system you select for a mission, but the relevant planning question is how much collection to task against a given emitter set and how quickly the library can be updated and distributed. If an opponent fields a new radar, the value of getting its parameters into fleet mission data files within days rather than months is very high. For a program planner, invest in the analysis and distribution pipeline at least as much as in collection platforms, since collection has rarely been the bottleneck.

Collection range set by the radio horizon, roughly 400 km from a 10 km altitude aircraft and global from satellites · frequency coverage typically 0.5–40 GHz · pulse timing measured to nanoseconds and frequency to well under 1 MHz · library update target of days rather than months after a new emitter appears · platform and ground-segment cost in the hundreds of millions.

Emission control and mode discipline are the standard counters, and they are effective. Reserving wartime modes means the collected library is incomplete by design. Modern software-defined radars can change waveform parameters with a software load, which devalues a library much faster than hardware-defined emitters did. Deliberate false emissions from decoys can pollute a library. Low-probability-of-intercept waveforms reduce collection range.

RC-135V/W Rivet Joint, the RC-135S Cobra Ball for missile telemetry, the EP-3E and its successors, the US Navy's ocean surveillance satellites, the Russian Lotos-S and Chinese Yaogan ELINT constellations, and the ground stations that support all of them.

This is national-capability spending with no commercial market and very limited visibility. The costs are in platforms, ground processing, and analyst time. The important trend is that software-defined radar shortens the useful life of an ELINT library, which raises the required collection and update tempo. That favors organizations with automated parametric analysis over those relying on manual characterization, and it is one of the clearer places where machine learning has produced real operational value.

6 minSCEPTRE: RADAR Analysis3dB Labs · 10k+ views

6 minPulse Analysis with VSA 2020 Release #03: Deinterleaving for Multi-emittersKeysight Design Software · 1k+ views

Intelligence Collection Programs and Systems (Federation of American Scientists) · The Intelligent Evolution of Radar Signal Deinterleaving: A Systematic Review from Foundational Algorithms to Cognitive AI Frontiers (Sensors via PMC)

Communications intelligence collects and exploits communications signals. Modern practice divides into content — what was said, which usually requires breaking or bypassing encryption — and metadata and traffic analysis, which needs neither. Who transmits, when, from where, to whom, and how often reveals organizational structure, readiness state, and movement even when every message is encrypted. Direction finding on the transmissions produces locations. For most tactical purposes traffic analysis and geolocation matter more than content, because they are available immediately and reliably.

COMINT is passive and gives insight into intent and organization that no other sensor provides. Metadata analysis is robust to encryption, which is what keeps the discipline valuable as strong crypto has become universal. The weaknesses are volume and law. Collection produces far more data than can be processed, so selection and automation are the binding constraints. Widespread end-to-end encryption has cut content access sharply. Domestic collection is legally constrained in most democracies, and the boundary is a recurring subject of oversight and litigation.

Tactical COMINT is worth deploying wherever an opponent uses radio: it supports targeting through geolocation, provides warning, and reveals unit boundaries and command relationships. The Ukraine conflict has shown how much can be derived from undisciplined use of unencrypted or commercially encrypted radios and phones. If the opponent has good communications security discipline and uses low-probability-of-intercept datalinks, expect little content and plan around metadata and geolocation instead. Cheap software-defined receivers make basic tactical collection accessible to almost anyone now.

Tactical VHF and UHF intercept range roughly 30–50 km from a ground mast, several hundred kilometers from an aircraft, intercontinental on HF skywave · direction-finding accuracy typically 2–5° RMS, giving fixes of hundreds of meters to a few kilometers · frequency coverage typically 20 MHz to 6 GHz on tactical systems · wideband software-defined receiver hardware from a few thousand dollars.

Communications security discipline, strong encryption, frequency hopping, and directional or low-probability-of-intercept links defeat most of it. Radio silence defeats all of it. Traffic padding and dummy transmissions frustrate traffic analysis. At the tactical level, using wired links and couriers instead of radio is a low-technology counter that continues to work well against a technically superior opponent.

National agency collection systems, tactical ground stations like the US Army's Prophet, airborne collection on Rivet Joint and Guardrail, shipborne SIGINT suites, and the widely reported use of commercial software-defined radios for battlefield collection by both regular and irregular forces.

Software-defined radio has commoditized the collection hardware — a capable wideband receiver now costs a few thousand dollars — which has pushed collection capability far down the sophistication ladder and into non-state hands. The remaining barriers are processing at scale, language and analytic capacity, and legal authority. Value in this discipline has moved almost entirely to the software and analytic layer, and commercial vendors selling collection and analysis tools have become a policy concern in their own right.

6 minSIGINT (Explained Simply)Matt Luskus · 5k+ views

11 minIntercept Radio Signals For Intelligence Gathering With An RTL SDRTech Minds · 100k+ views

Intelligence Collection Programs and Systems (Federation of American Scientists) · Defense Primer: Electronic Warfare (Congressional Research Service)

Specific emitter identification fingerprints an individual transmitter rather than a transmitter type. Every radio and radar carries small unintentional modulations from its own components — oscillator drift, amplifier nonlinearity, turn-on transients, tiny timing irregularities — and these vary between units of the same model. Measuring them precisely enough produces a fingerprint that identifies a particular ship, aircraft, or vehicle across time and location. The technique moves electronic support from "there is an X-band radar of this type" to "that is the same emitter we saw in the Gulf last month."

Persistent tracking of specific platforms is the payoff, and it works through call-sign changes and encryption because the fingerprint lives in the physical layer. It is entirely passive. The weaknesses are that fingerprints drift as equipment ages, is repaired, or is temperature-cycled, so the reference database needs continual maintenance. Very high receiver fidelity is required, which drives cost. Machine-learning classifiers trained on a fingerprint set often degrade badly when the collection receiver or the propagation conditions change, and validating a classifier's accuracy in the field is genuinely difficult.

Use specific emitter identification when identity, not detection, is the requirement: tracking particular vessels in a crowded maritime picture, confirming that a specific air-defense unit has moved, or associating a new emission with a previously observed platform. If you only need to know a threat type is present, ordinary electronic support is far cheaper. Treat database currency as an operational task with its own tempo — the fingerprint library ages continuously, and its accuracy is the system's real performance metric rather than the receiver's specifications.

Usable range shorter than detection range, since fingerprinting needs roughly 20 dB signal-to-noise or better · receiver fidelity of 14–16 bit digitizers with very low phase noise · published classifier accuracy above 90% under matched conditions and well below that when the receiver or channel changes · fingerprint libraries refreshed on a months timescale as equipment ages and is repaired.

Emission control defeats it. Replacing or repairing components changes the fingerprint, sometimes accidentally and sometimes deliberately. Deliberate transmitter-level randomization has been demonstrated in research and would degrade the technique substantially if fielded. Distance and poor signal-to-noise ratio wash out the subtle features the method depends on, so it generally needs closer or higher-quality collection than simple detection does.

Maritime domain awareness systems that track vessels through transponder-off periods, national SIGINT programs, and a growing body of commercial research applying deep learning to RF fingerprinting for wireless-network security, which shares most of the underlying technique.

The capability has been a closely held national one, but wideband software-defined receivers plus machine learning have made research versions accessible to universities and startups. Commercial applications in wireless device authentication and drone identification are emerging and share the technology base. The defense value is disproportionate to the hardware cost because it turns a stream of detections into a persistent track history, which is the kind of intelligence product that is hard to produce any other way.

23 minArtificial Intelligence Colloquium: Radio Frequency Machine Learning SystemsDARPAtv · 5k+ views

A Comprehensive Survey on Radio Frequency (RF) Fingerprinting: Traditional Approaches, Deep Learning, and Open Challenges (arXiv) · Feature Analysis and Extraction for Specific Emitter Identification Based on the Signal Generation Mechanisms of Radar Transmitters (Sensors via PMC)

Passive sonar listens. Hull-mounted arrays, flank arrays, and towed arrays of hydrophones detect the noise a target makes — machinery, propeller cavitation, flow over the hull — and beamform to give a bearing. Towed arrays can be hundreds of meters long, which gives the aperture needed for low-frequency directionality, and being towed astern puts them away from the towing ship's own noise. Sound in the ocean travels far but not simply: temperature and pressure gradients bend rays, creating shadow zones and convergence zones where detection is impossible or unusually good, sometimes tens of kilometers apart.

Listening reveals nothing about the listener, which in submarine warfare is decisive. Detection ranges under favorable conditions reach tens to hundreds of kilometers in the deep ocean via convergence zones or the deep sound channel. Narrowband analysis identifies specific machinery and often the specific class or hull. The weaknesses: it only works if the target makes noise, and modern submarines are extraordinarily quiet — rafted machinery, anechoic tiles, and slow speeds put some boats near the ambient noise floor. Performance varies by an order of magnitude with the sound speed profile, so predicting your own detection range is itself a specialized task.

Passive sonar is the primary sensor for submarine detection and for submarines' own situational awareness, and it is the right default in any undersea problem where remaining undetected matters. If the target is genuinely quiet or the environment is noisy and shallow, active sonar may be the only way to detect at all, at the cost of announcing yourself. Shallow water inverts many of the assumptions: multipath, boundary reflections, and heavy shipping noise make littoral passive detection much harder than deep-ocean detection.

Detection range from a few kilometers in noisy shallow water to hundreds of kilometers through convergence zones or the deep sound channel · convergence zone annuli spaced roughly 55–65 km apart in deep water · towed arrays 100 m to over 1,000 m long, working from about 10 Hz to a few kHz · bearing accuracy around 1° broadside and much worse near endfire · suite cost in the tens of millions per submarine.

Quieting is the counter, and it has been extremely effective — the noise reduction achieved on modern submarines is one of the largest sustained engineering efforts in naval history. Operating slowly reduces flow and cavitation noise. Hiding under a layer or in a shadow zone exploits the sound speed profile. High ambient noise from shipping, weather, and biology masks weak signals, and littoral waters supply this for free.

The AN/BQQ-10 sonar suite on US submarines, TB-29 and TB-34 towed arrays, the SURTASS arrays on ocean surveillance ships, the SOSUS fixed hydrophone network of the Cold War, and the sonar suites on the Astute, Virginia, and Type 212 classes.

Undersea sensing is expensive, concentrated among a few national suppliers, and rising in importance as undersea infrastructure protection has become a public policy issue. The notable change is a commercial layer: distributed low-cost hydrophone networks, uncrewed vehicles carrying arrays, and fiber-optic distributed acoustic sensing on existing seabed cables all promise wide-area coverage at a fraction of traditional cost. Whether cheap distributed sensing can match a large towed array's sensitivity is the open question, and it is currently the most interesting one in the field.

18 minHow the Navy Hears Subs 3,000 Miles Away. (It's Genius)Second Order Science · 100k+ views

11 minHow Submarines Can 'See' Underwater - Sonar OverviewCovert Cabal · 100k+ views

3 minAPL Defining Innovations: Advanced Sonar ArraysJohns Hopkins Applied Physics Laboratory · 10k+ views

The Sound Surveillance System (SOSUS) (Federation of American Scientists) · TB-29 Thin Line Towed Array (Federation of American Scientists)

Active sonar transmits a pulse and listens for the echo, which gives range as well as bearing — something passive sonar cannot do from a single platform. Low-frequency active sonar in the hundreds of hertz travels far and is used for wide-area search; mid-frequency around 1–10 kHz is the standard shipborne and dipping-sonar band; high-frequency above 100 kHz gives fine resolution for mine hunting and imaging at short range. Multistatic operation, where one platform transmits and several others listen, keeps the receivers covert and complicates the target's counter-detection problem.

Active sonar detects quiet targets that passive sonar cannot hear, and it produces range immediately, which is what a firing solution needs. Against a modern quiet submarine it is often the only option. The costs are severe: transmitting announces your position to every passive listener within a much greater radius than your own detection range, so going active in a submarine-threat environment is a deliberate tactical decision. Reverberation from the seabed and surface clutters returns in shallow water. Environmental restrictions on high-power low-frequency sonar are significant in many waters because of marine mammal effects.

Go active when detection matters more than concealment: a surface task group screening against a suspected submarine, a helicopter localizing a contact for attack, or a mine countermeasures vehicle imaging the seabed. Submarines generally do not, except for a single ranging ping in extremis. Multistatic architectures — a noisy expendable source with quiet distributed receivers — get much of the benefit without exposing the valuable platform, and they are where most of the current development effort is going.

Low-frequency search in the hundreds of hertz, mid-frequency hull and dipping sonar at 1–10 kHz, mine-hunting sonar above 100 kHz · source levels around 220–235 dB re 1 micropascal at 1 m on hull sets · detection range typically 10–20 km at mid frequency in good conditions · range resolution from tens of meters at low frequency to centimeters above 100 kHz · counter-detection range for a listening submarine several times the sonar's own detection range · shipborne suite cost in the tens of millions.

Anechoic coatings absorb the incident pulse and cut target strength substantially. Bubble decoys and mobile acoustic countermeasures generate false echoes. Bottom-following and hiding in reverberation-heavy shallow water frustrate returns. And the fundamental counter is simply listening: a submarine detects an active sonar at far greater range than the sonar detects the submarine, so it can leave the area before the searcher knows it was there.

The AN/SQS-53C hull sonar on US destroyers, the AN/AQS-22 dipping sonar on MH-60R helicopters, SURTASS-LFA low-frequency active systems, variable depth sonars like CAPTAS on European frigates, and high-frequency mine-hunting sonars on uncrewed underwater vehicles.

Full shipborne active sonar suites cost tens of millions and are built by a small number of suppliers. The direction of travel is toward distributed multistatic systems using expendable or uncrewed sources, which spreads cost and reduces the exposure of high-value platforms. Environmental regulation is a genuine constraint on where and how systems can be used and tested, and it shapes procurement in a way that has few parallels in other sensing domains.

27 minHow Sonar Works (Submarine Shadow Zone) - Smarter Every Day 249SmarterEveryDay · 1m+ views

4 minUnrivalled Anti-submarine warfare with CAPTAS-4 VDS and FLASH dipping sonar - ThalesThales · 10k+ views

3 minInitial In-Water Testing of Northrop Grumman's AQS-24 Mine Hunting SonarNorthrop Grumman · 10k+ views

AN/SQS-53 Sonar (Federation of American Scientists) · Incidental Take Authorizations for Military Readiness Activities (NOAA Fisheries)

A sonobuoy is an expendable acoustic sensor dropped from an aircraft. It deploys a hydrophone to a set depth, floats on the surface with a radio antenna, and relays what it hears to the aircraft. Passive types listen omnidirectionally or with directional processing; active types transmit a pulse; specialized types measure the water temperature profile so the crew can predict sound propagation. A maritime patrol aircraft carries 80–200 of them and lays patterns over an area, converting a single aircraft into a temporary distributed acoustic array covering hundreds of square kilometers.

Coverage per aircraft sortie is the advantage, and the sensor is expendable so it can be placed anywhere without risking a platform. Cost per unit is a few hundred to a few thousand dollars. Patterns can be tailored to the environment, and the temperature-profile buoys let the crew adapt in real time. The weaknesses are life and consumption: a buoy runs for one to eight hours and then scuttles itself, sea state limits performance, and the consumption rate in a real search is high enough that magazine depth and production capacity become the binding constraint rather than aircraft availability.

Use sonobuoys for airborne anti-submarine search and localization, which is what they exist for, and increasingly for wide-area acoustic monitoring from uncrewed aircraft. If persistent monitoring of a fixed area is the requirement, seabed arrays or moored sensors are far more economical over time. If the search area is small and a helicopter is available, a dipping sonar can be repositioned repeatedly and is not expended. The planning number that matters is buoys per contact prosecuted, and it is usually larger than peacetime exercises suggest.

Unit cost a few hundred to a few thousand dollars · life 1 to 8 hours before the buoy scuttles itself · hydrophone deployed to depths of roughly 30–300 m · 80–200 buoys per maritime patrol aircraft, covering hundreds of square kilometers per pattern · VHF relay to the aircraft over tens of kilometers.

Submarine quieting reduces detection range, as with all passive acoustics. High sea state raises ambient noise and degrades the surface radio link. Buoy life limits how long a pattern holds. Deep or shallow acoustic conditions can put a target outside the buoy's depth coverage. And because production rates are modest, a sustained campaign can simply exhaust the inventory — a supply problem, not a sensing one.

The AN/SSQ-53 DIFAR passive directional buoy, the SSQ-62 DICASS active buoy, the SSQ-36 bathythermograph, multistatic active coherent buoy sets, and the loads carried by P-8A Poseidon, P-1, and MQ-4C Triton aircraft.

Sonobuoys are one of the clearer cases of a consumable driving strategic capability. Unit costs are low but annual consumption in a serious campaign would run into six figures of units, and the production base is narrow — a handful of suppliers in the US, UK, and elsewhere. Multi-year procurement has been used specifically to justify capacity investment. Uncrewed aircraft dispensing buoys is an active development area, because it decouples buoy delivery from expensive crewed platforms.

4 minSonobuoy: How Does This Portable SONAR System Bust Enemy Submarines?ScienceABC II · 1k+ views

18 minU.S. NAVY SONOBUOY INDICATOR GROUP AN/AQA-1 ANTI-SUBMARINE WARFARE FILM 51114PeriscopeFilm · 10k+ views

4 minThales SonoFlash Active Passive SonobuoyNaval News · 5k+ views

Sonobuoys (Federation of American Scientists) · AN/SSQ-53 Directional Frequency Analysis and Recording Sonobuoy (Federation of American Scientists)

A magnetic anomaly detector measures tiny distortions in the Earth's magnetic field caused by the several thousand tonnes of ferrous metal in a submarine hull. The Earth's field is about 50,000 nanotesla; a submarine at a few hundred meters produces a perturbation of a few nanotesla, so the instrument must resolve roughly one part in ten thousand while being carried on a moving aircraft made of metal. Sensors are optically pumped cesium or helium magnetometers, mounted on a boom or towed in a bird well away from the aircraft, with careful compensation for the platform's own magnetic signature.

A magnetic detection is nearly unambiguous — very little else produces that signature — so it confirms a contact that acoustics only suggested, and it is completely passive and unjammable. The weakness is range. Detection is limited to a few hundred meters to about a kilometer, so MAD is a confirmation and localization sensor, not a search sensor: the aircraft must already be nearly on top of the submarine. That forces low, slow flight, which is why the sensor has been dropped from some modern patrol aircraft in favor of remaining at altitude and using acoustics and radar.

Use MAD for final localization and attack confirmation once acoustics have narrowed a contact to a small area, and for shallow-water work where acoustic conditions are poor but a submarine cannot be far below. If the platform cannot safely fly low and slow, or if the contact area is still large, MAD adds nothing. The P-8A's omission of a MAD boom in favor of high-altitude acoustic search is the clearest statement of the current trade — the sensor is genuinely useful but it constrains the aircraft's whole employment concept.

Earth's field about 50,000 nanotesla, submarine anomaly a few nanotesla at a few hundred meters · detection range a few hundred meters to about 1 km · signal falling off with the cube of distance · sensor noise floor of a few picotesla per root hertz on optically pumped magnetometers · search altitude of roughly 60–100 m, which forces low and slow flight.

Degaussing reduces a submarine's permanent magnetic signature, and every navy does it as routine maintenance. Non-magnetic hull materials — titanium on some Soviet designs, glass-reinforced plastic on mine countermeasures vessels — reduce or eliminate the signature. Depth helps, since the anomaly falls off with the cube of distance. Geological magnetic variation and solar activity raise the noise floor and produce false contacts in some areas.

The AN/ASQ-81 and ASQ-508 MAD sets on P-3C, S-3, and various helicopters, the MAD-XR compact system offered for uncrewed platforms, and the magnetic sensors used in mine countermeasures and unexploded-ordnance survey.

MAD is a small, mature market, and the technology overlaps heavily with geophysical survey magnetometers used in mineral exploration, which sustains the supplier base. The interesting development is miniaturization: compact magnetometers light enough for uncrewed aircraft change the employment concept entirely, because a cheap drone can fly the low, slow profile that a crewed patrol aircraft should not. Quantum magnetometers using nitrogen-vacancy diamond or SQUID technology promise better sensitivity and are an active research area with a plausible path to fielding.

3 minCAE MAD-XR - (Magnetic Anomaly Detection-Extended Role)CAEvideogallery · 5k+ views

10 minU.S. NAVY TOWED MAGNETIC AIRBORNE DETECTION SYSTEM ANTI-SUBMARINE WARFARE VEHICLE GG11485PeriscopeFilm · 5k+ views

21 minPart II : Optically Pumped Magnetometer System Example: ASQ-81Andrew Ochadlick · 1k+ views

Magnetic Anomaly Detection (Bartington Instruments) · Determinants of Maximum Magnetic Anomaly Detection Distance (Sensors via PMC)

A GNSS receiver measures the travel time of signals from four or more satellites and solves for position and time. GPS, Galileo, GLONASS, and BeiDou each provide global coverage; a modern multi-constellation, multi-frequency receiver uses all of them, which improves accuracy and availability. Standard positioning gives 3–5 m accuracy, dual-frequency corrections give under a meter, and differential or real-time kinematic techniques give centimeters. The receiver also provides precise time, and that timing function underpins power grids, financial systems, and cellular networks far more than most people realize.

Accuracy, global coverage, and near-zero cost are the reasons GNSS is in everything: a chipset costs a couple of dollars and needs no infrastructure from the user. The weakness is the signal, which arrives from 20,000 km away at about −160 dBW — roughly the power of a 25 W light bulb seen from 20,000 km. That is far below the thermal noise floor and recoverable only through spread-spectrum processing gain. A few watts of jamming denies it over tens of kilometers. Spoofing, which transmits false signals rather than noise, is worse: the receiver reports a confident, wrong position.

Use GNSS as the primary position and time reference in any benign environment, which covers nearly all civil use. In contested environments, treat it as a bonus rather than a foundation: assume it will be unavailable, design the system to work on inertial and other references, and use GNSS to bound inertial drift when it happens to be available. That inversion — GNSS-aided inertial rather than inertial-aided GNSS — is the single most important architectural change in military navigation over the past decade, and it is being repeated now in civil aviation as jamming near conflict zones has become routine.

Position accuracy 3–5 m standard, under 1 m dual-frequency, centimeters with real-time kinematic corrections · timing to tens of nanoseconds · received signal power about −160 dBW, below the thermal noise floor · over 100 satellites across four constellations, 30-plus in view at a time · chipset cost a couple of dollars · a few watts of jamming denying it over tens of kilometers.

Barrage jamming with a few watts denies a wide area, and commercial jammers cost tens of dollars. Spoofing with a software-defined radio produces false positions, and it has been observed at scale near conflict zones and around some ports and airports. Urban canyons and foliage attenuate the signal without any adversary involved. Ionospheric scintillation degrades it at high latitudes and near the equator. Signal outages also break timing in infrastructure that has no backup, which is a systemic civil vulnerability.

Every smartphone, vehicle, and precision-guided munition; u-blox and Trimble receiver modules; survey-grade RTK systems; the timing receivers in cellular base stations and power substations; and the widely documented GPS interference around the Baltic, the eastern Mediterranean, and the Black Sea.

GNSS is free infrastructure paid for by four governments and consumed by a global commercial industry worth hundreds of billions. Receiver chipsets are commodities. The economically interesting layer now is resilience: anti-jam antennas, alternative timing sources, and commercial low-Earth-orbit navigation services that transmit at far higher power than medium-orbit constellations. Several startups are building exactly that, and the demand signal from aviation and critical infrastructure has become substantial since 2022.

4 minHow GPS works? Trilateration explainedunfa🇺🇦 · 100k+ views

14 minUnderstanding GPS Links and CodesRohde & Schwarz · 10k+ views

20 minGPS Jamming & Spoofing - How Does It Work, And Who's Doing It?Scott Manley · 100k+ views

GPS.gov: official U.S. government information about GPS · GNSS Receivers General Introduction (ESA Navipedia)

Anti-jam GNSS combines three separate defenses. Controlled reception pattern antennas use an array of four to sixteen elements to steer nulls toward jammers and beams toward satellites, typically providing 20–50 dB of rejection and handling as many jammers as the array has elements minus one. Encrypted military signals — GPS M-code, Galileo PRS — resist spoofing because the receiver can authenticate them, and M-code's higher power and better structure also improves jam resistance. And tight coupling with an inertial unit lets the navigation solution coast through outages and reject implausible position jumps.

A well-designed anti-jam installation raises the jammer power needed by two to five orders of magnitude, which turns a $50 jammer problem into a serious engineering problem for the opponent. Authenticated signals largely close the spoofing hole. The weaknesses are size, cost, and rollout. A CRPA array needs a clear aperture roughly 20–40 cm across, which does not fit a small munition, and it costs thousands to tens of thousands of dollars. M-code receiver fielding has been slow and expensive across every military that has attempted it, and legacy platforms often cannot take the new hardware without significant modification.

Fit anti-jam GNSS on anything that must navigate accurately in contested airspace and can carry the antenna: aircraft, ships, ground vehicles, and larger munitions. For small munitions and cheap drones the antenna does not fit, and the answer is a good inertial unit plus terrain or visual navigation instead. As a rule of thumb, if the platform costs more than about $100k, anti-jam GNSS is worth it; below that, design for GNSS-denied operation from the start rather than trying to protect a signal you cannot adequately defend.

CRPA arrays of 4–16 elements giving 20–50 dB of jammer rejection and nulling N−1 jammers · antenna aperture roughly 20–40 cm across · required jammer power raised by two to five orders of magnitude · antenna and receiver cost thousands to tens of thousands of dollars · worth fitting above roughly $100k of platform value.

Numbers beat nulls: an array with N elements can null N−1 jammers, so enough distributed jammers saturate it. Very high power jamming close in overwhelms the front end. Spoofing that mimics authenticated signals is hard, but replay and meaconing attacks still cause problems for receivers that do not check timing consistency. And the practical defeat is simply that most platforms in most militaries still do not have anti-jam fitted, so the capability exists on paper more than in the field.

The GPS Directorate's M-code receivers (MGUE), Raytheon and BAE anti-jam antenna systems, the Digital Anti-Jam Receiver family, Galileo PRS receivers in European programs, and the retrofit programs that most air forces are currently running on their fleets.

This is a large, slow, expensive modernization market. M-code fielding alone is a multi-billion-dollar effort across US services, and progress has repeatedly slipped. Antenna and receiver suppliers are a small group with strong export controls. The parallel commercial development — low-Earth-orbit navigation and timing services from companies like Xona and TrustPoint, plus fiber and atomic timing for infrastructure — is growing quickly because civil operators have concluded they cannot wait for military solutions.

10 minTesting CRPA ReceiversRohde & Schwarz · 1k+ views

3 minHow to Protect from GPS and GNSS Jamming | GAJT® Anti-Jam Solutions | NovAtel® Defense SolutionsNovAtel · 5k+ views

10 minNull Steering: IntroductionJohn Buck · 1k+ views

GPS.gov: official U.S. government information about GPS · GPS Modernization: Delays Continue in Delivering More Secure Capability for the Warfighter (GAO)

Terrain contour matching compares the ground profile measured by a radar altimeter along the flight path against a stored digital elevation map, and slides the measured profile over the map until it fits. The result is an absolute position fix that requires no external signal and cannot be jammed. TERCOM was developed for cruise missiles in the 1970s and gave Tomahawk its accuracy over intercontinental ranges long before GPS existed. TERPROM, the terrain-referenced navigation used on combat aircraft, applies the same principle continuously for both navigation and ground-collision avoidance.

Complete immunity to jamming and spoofing is the defining advantage, and accuracy is good — tens of meters, improving to meters in distinctive terrain. It is entirely self-contained apart from the stored map. The weaknesses are terrain and data. Flat terrain, water, and desert have no contours to match, so the technique fails exactly where a low-flying missile might most want it. Fresh snow and seasonal change alter measured profiles. And the map has to be made in advance from elevation data of sufficient resolution, which is a planning burden and historically a limit on where the weapon could be used.

Use terrain-referenced navigation for cruise missiles and low-flying aircraft over terrain with relief, as the jam-proof backbone of a navigation solution that GNSS merely refines when available. If the route crosses ocean, flat plains, or featureless desert, it will not work and you need inertial coasting or scene matching against distinctive features instead. Modern practice blends it: inertial for continuity, terrain matching for absolute fixes over land, GNSS opportunistically, and scene matching for the terminal phase.

Fix accuracy tens of meters, a few meters over distinctive relief · accuracy constant with flight time, unlike inertial drift · reference maps built from 30 m SRTM-class elevation data or better · radar altimeter working in the 4.2–4.4 GHz band · fixes at discrete map patches en route, continuous in TERPROM-style aircraft systems.

Featureless terrain defeats it passively. Large-scale terrain modification is impractical, which is part of the technique's appeal, but seasonal change, flooding, and heavy snow degrade matching. The dependence on pre-mission elevation data means denying or corrupting that data is an attack path, and global elevation datasets from radar interferometry have made this far harder to control than it was in the 1980s. Radar altimeter emissions are also detectable, so the aircraft is not entirely passive.

TERCOM on Tomahawk and the AGM-86 ALCM, TERPROM on the F-16, Tornado, Typhoon, and Gripen, the Russian equivalents on Kh-55 and Kalibr, and terrain-referenced navigation systems increasingly fitted to uncrewed aircraft as a GPS backup.

The technology itself is old and inexpensive; the value sits in the elevation data and the mission-planning system that turns it into usable maps. Global elevation datasets from SRTM and commercial radar satellites removed much of the former data barrier, which has made terrain-referenced navigation broadly accessible. That is a meaningful proliferation effect: a jam-proof navigation technique that once required national mapping resources now needs an open dataset and reasonable software.

16 minTomahawk Missile – How is it Nearly 50 Years Old and Still Going Strong?Curious Droid · 500k+ views

10 minHow Does a Cruise Missile Work?History of Simple Things · 500k+ views

3 minSOFEX 2014 UTC Aerospace Systems TERPROM Digital Terrain SystemJanes · 1k+ views

Intelligence Collection Programs and Systems (Federation of American Scientists) · 3D Elevation Program (US Geological Survey)

Scene matching compares a live image from a downward or forward-looking camera against stored reference imagery and computes a position fix from the correlation. The original implementation, DSMAC on Tomahawk, correlated digitized scenes near the target to refine terminal accuracy to a few meters. Modern versions use learned features rather than raw correlation, match against satellite imagery basemaps, and run on hardware costing a few hundred dollars, which is why visual navigation has appeared rapidly on cheap drones as a GPS-denied fallback.

It gives an absolute fix with no emissions and no external signal, and modern implementations achieve tens of meters accuracy over long flights and meter-level accuracy against distinctive scenes. Hardware is cheap and light. The weaknesses are all about the imagery: night, cloud, snow, and seasonal change alter the scene enough to break naive matching, though learned methods are considerably more robust than correlation was. Featureless terrain — open water, desert, uniform forest — provides nothing to match. And the reference imagery must be current enough that the world still looks like the map.

Use visual navigation as the GPS-denied fallback on any small air vehicle that carries a camera anyway, which is nearly all of them. It is the cheapest jam-resistant absolute navigation available, and it is what most low-cost strike drones now use when GNSS is denied. If the route is over water or featureless ground, or the mission is at night without a thermal camera, plan on inertial coasting instead. Pair it with a good MEMS inertial unit: the camera supplies absolute fixes and the inertial unit bridges the gaps and handles fast motion.

Fix accuracy tens of meters over a long flight, meter-level against a distinctive scene, a few meters for DSMAC-style terminal matching · camera and compute module costing tens to a few hundred dollars per aircraft · reference imagery at sub-meter to a few meters per pixel from commercial satellite basemaps · matching at camera frame rate, typically 10–30 Hz · no usable fix over open water, desert, or uniform forest.

Darkness and cloud stop visible-band matching, though thermal cameras and radar altimetry partially cover that. Featureless terrain gives nothing to correlate. Seasonal and structural change degrades matching against stale reference imagery. Deliberate scene alteration is possible at small scale but impractical over a route. Smoke and dust in an active battle area degrade it exactly when it matters most, which is a real limitation observed in current conflicts.

DSMAC on Tomahawk, the visual terminal guidance on several loitering munitions, commercial visual positioning systems from companies like Sightec and VisionNav, ArduPilot and PX4 optical-flow and visual-odometry modes, and the widely reported use of visual navigation on Ukrainian and Russian strike drones.

This is the cheapest anti-jam navigation technology available and it is diffusing extremely fast, because the hardware is a camera and a small compute module and the reference imagery is commercially available. Cost per aircraft is tens to hundreds of dollars. That combination has significant consequences: GNSS jamming, which was a reliable counter to cheap drones two years ago, is much less effective against a drone that navigates visually, and defensive planning has had to adjust accordingly.

6 minPalantir | Visual Navigation for DronesPalantir · 10k+ views

5 minGPS-Denied Navigation Anywhere in the WorldVantor · 5k+ views

2 min150km GPS-denied UAV Localization with one single Satellite ImageMetaSLAM · 1k+ views

A digital scene matching technique for geometric image correction and autonomous navigation (NASA Technical Reports Server) · Cross-view geo-localization: a survey (arXiv)

Automated celestial navigation measures the angular position of stars with a small telescope and computes position from the known positions of those stars and a precise time reference. Modern star trackers work in daylight, using narrowband filters and sensitive detectors to pull a star out of the bright sky background. Combined with an inertial unit, the result is a stellar-inertial system: the inertial unit provides continuity and attitude, and periodic star fixes bound its drift. Accuracy of tens to hundreds of meters over long flights is achievable with no external signal whatsoever.

It is unjammable and unspoofable in any practical sense — you cannot fake the fixed stars — and it requires no infrastructure at all. Accuracy is stable regardless of flight duration, which is what makes it valuable for very long-range systems where inertial drift would otherwise accumulate. The weaknesses are clouds, cost, and mechanical complexity. A star tracker needs a clear view of the sky, which rules out low-altitude flight under overcast. The optics and the stabilization to point them from a moving vehicle are expensive, and the whole system requires very accurate time and attitude to convert a star sighting into a position.

Use stellar-inertial navigation for systems that must be accurate over very long distances with no external reference: ballistic missiles, strategic bombers, high-altitude reconnaissance aircraft, and spacecraft. If the platform flies low or in weather, celestial navigation will not have sky access and terrain or visual matching is the answer. The renewed interest in this old technology is entirely driven by GNSS vulnerability, and daytime star trackers small enough for tactical aircraft are now being marketed on exactly that argument.

Position accuracy tens to hundreds of meters, stable regardless of flight duration · star tracker attitude accuracy of a few arcseconds · time reference good to tens of milliseconds, since 1 s of time error is about 460 m of position error at the equator · smallsat-class star tracker cost in the tens of thousands of dollars · no fixes at all below overcast.

Cloud cover is the complete counter for any platform below it. The systems are expensive and heavy enough that they will never be fitted broadly. Accurate time is required, and if the time reference is derived from GNSS the system inherits a GNSS dependency that partly defeats the purpose — good implementations carry an independent atomic or oven-controlled oscillator. Beyond that, there is no known way to attack the technique remotely, which is the reason it persists.

The stellar-inertial guidance on Trident II D5 and on Soviet and Russian SLBMs, the astro-tracker on the SR-71 and the B-2, the star trackers on essentially every spacecraft, and modern compact daytime star trackers being offered for aircraft and munitions as GNSS alternatives.

Historically this was a strategic-systems technology with very small production runs and high unit costs. The current commercial angle is that spacecraft star trackers have become cheap and small — a smallsat unit costs tens of thousands of dollars — and several companies are adapting that supply chain to atmospheric use. Whether daytime performance at tactical aircraft altitudes justifies the cost against cheaper terrain and visual methods is the open commercial question.

26 minB-52 Bomber Astro Tracker - Part 1: Grand Opening (ft. Le Labo de Michel)CuriousMarc · 100k+ views

20 minB-52 Bomber Astro Tracker - Part 3: Figuring out the optical systemCuriousMarc · 50k+ views

20 minCelestial Navigation: Celestial Position FixRefresh Maritime · 100k+ views

Guidance, Navigation, and Control: State of the Art of Small Spacecraft Technology (NASA) · Celestial Navigation Resources (US Naval Observatory)

The Earth's crust produces small, spatially varying anomalies in both the magnetic and gravitational fields, and those anomalies are stable over decades. A vehicle carrying a sensitive magnetometer or gravity gradiometer can measure the local field, compare it against a stored anomaly map, and derive a position fix by the same correlation logic that terrain matching uses — except that it works over water, under water, in cloud, and at night, because the fields pass through everything. Accuracy demonstrated in flight testing is in the hundreds of meters to a few kilometers, improving as maps and sensors improve.

It is passive, unjammable, and works in every medium and every weather condition, including underwater where nothing else absolute is available. Field anomalies are stable, so maps stay valid for decades. The weaknesses are sensitivity and interference. Magnetic anomalies are a few hundred nanotesla against a 50,000 nT background, and the vehicle's own electrical systems produce comparable disturbances, so platform compensation is the central engineering problem. Solar activity perturbs the magnetic field on hourly timescales. Gravity gradiometers are expensive and slow to average. And large areas of the world are not mapped at useful resolution.

Use field-based navigation where GNSS is denied and no other absolute reference is available: submarines, long-range aircraft over ocean, and any platform that must navigate under cloud with no terrain relief. It complements terrain matching rather than replacing it, since terrain works well over land and field navigation works over water. Treat map availability as the gating question — the technique's accuracy is set by survey coverage and resolution more than by the sensor, and the relevant maps are national assets in many regions.

Fix accuracy hundreds of meters to a few kilometers in flight testing · crustal anomalies of a few hundred nanotesla against a 50,000 nanotesla background · platform magnetic interference comparable in size to the signal, so compensation sets the achievable accuracy · global anomaly grids at roughly 2 arc-minute resolution, about 4 km, with national surveys much finer · maps stable over decades · solar storms degrading fixes for hours.

Poor or absent map coverage limits it geographically. Solar storms disturb the magnetic field enough to degrade fixes for hours. Platform magnetic interference must be compensated continuously, and a change in the vehicle's electrical configuration can invalidate the calibration. Because accuracy is coarse compared with GNSS, it is a bounding technique for inertial drift rather than a precision fix.

The MagNav flight demonstrations by the US Air Force and MIT Lincoln Laboratory, SandboxAQ's quantum magnetometer navigation trials, gravity-aided inertial navigation on submarines, and the underlying national magnetic and gravity anomaly maps produced by geological surveys.

Interest is recent and driven by GNSS denial. Quantum magnetometers — optically pumped alkali vapor and nitrogen-vacancy diamond sensors — are the technology enabler, since they offer high sensitivity in a small package without cryogenics. Several startups and national laboratories are working on it, funding is growing, and flight demonstrations have shown real navigation performance. The pacing items are platform compensation algorithms and map coverage, neither of which is a sensor problem.

3 minThe Science Behind Magnetic Mapping and GPS-Free NavigationSandboxAQ · 5k+ views

9 minSignal Enhancement for Magnetic Navigation Scientific Machine Learning Challenge ProblemThe Julia Programming Language · 1k+ views

2 minGravity Gradiometer for Navigation.wmvANT Center · 1k+ views

MagNav project successfully demonstrates real-time magnetic navigation (MIT Lincoln Laboratory) · Earth Magnetic Anomaly Grid (EMAG) 2 (NOAA National Centers for Environmental Information)

Terrestrial radio navigation uses ground transmitters instead of satellites. TACAN and VOR/DME give aircraft bearing and distance to a station within a few hundred kilometers. Enhanced Loran transmits at 100 kHz and about 250 kW from a few large stations, giving 10–20 m accuracy over a thousand kilometers and precise timing as a backup to GNSS. Newer approaches use signals of opportunity — cellular base stations, broadcast towers, and low-Earth-orbit satellite downlinks — where the receiver derives position from signals never designed for navigation.

Ground transmitters deliver signals many orders of magnitude stronger than GNSS at the receiver, so jamming them requires vastly more power and far closer proximity. Low-frequency signals like eLoran propagate around terrain and into buildings and underground. Receivers are cheap. The weaknesses are coverage and infrastructure: someone has to build, power, and defend the transmitters, coverage is regional rather than global, and accuracy is worse than GNSS. Skywave interference at night degrades low-frequency systems, and cross-rate interference between stations requires careful management.

Use terrestrial radio navigation as the resilient backup layer for critical infrastructure timing and for aviation and maritime navigation in regions that fund it. It is the most practical answer to the systemic problem that power grids, financial timestamps, and telecom networks all depend on GNSS timing with no fallback. If you need global coverage or centimeter accuracy, it does not deliver either. The policy question — whether to fund eLoran — has been debated for two decades in the US and UK, and the answer has changed several times.

eLoran carrier 100 kHz at roughly 250 kW per station · fix accuracy 10–20 m out to about 1,000 km · VOR/DME and TACAN service radius of a few hundred kilometers per station · VOR bearing accuracy typically 1–2° · received signal many orders of magnitude stronger than GNSS · receivers cheap, transmitters the expensive part.

High-power jamming close to the receiver still works, it is just much harder than jamming GNSS. Destroying or disabling the small number of high-power transmitters removes regional coverage, and they are large fixed targets. Skywave interference degrades accuracy at night. The deeper vulnerability is political rather than technical: these systems get switched off when budgets tighten, and Loran-C was decommissioned in the US in 2010 precisely because GPS seemed to have made it redundant.

VOR/DME and TACAN networks worldwide, eLoran services in South Korea and the UK's trials, the Russian Chayka system, cellular-based positioning used in phones, and commercial LEO signals-of-opportunity navigation from companies working with Iridium and Starlink downlinks.

The economics are public-infrastructure economics: high fixed cost, no direct revenue, and a benefit that only becomes visible when GNSS fails. That has made funding chronically unstable. The commercial alternative gaining traction is LEO-based navigation and timing from dedicated smallsat constellations, which transmit far more power than medium-orbit GNSS and can be funded as a subscription service. Several such ventures are now funded, and critical-infrastructure demand for GNSS-independent timing is the clearest driver.

2 minHow LORAN WorksSmithsonian National Air and Space Museum · 100k+ views

7 minUnderstanding VORRohde & Schwarz · 50k+ views

4 minI found a cool GPS alternative nobody knows aboutLevel 2 Jeff · 100k+ views

GPS.gov: official U.S. government information about GPS · 2021 Federal Radionavigation Plan (Departments of Defense, Transportation, and Homeland Security)

A ring laser gyro sends two laser beams around a closed triangular or square cavity in opposite directions. Rotation makes one path effectively longer than the other, and the resulting beat frequency between the counter-propagating beams is directly proportional to rotation rate. Three of these plus three accelerometers form a navigation-grade inertial measurement unit. There are no moving parts except a small dither mechanism that keeps the beams from locking together at very low rotation rates, and the output is inherently digital, which is why RLGs displaced spinning-mass gyros in aircraft from the 1980s onward.

Navigation-grade performance is the point: bias stability around 0.001–0.01 degrees per hour gives position drift of roughly one nautical mile per hour of free inertial flight, which is good enough to cross an ocean. RLGs turn on quickly, have excellent scale-factor stability and a very wide dynamic range, and last tens of thousands of hours. The weaknesses are size, cost, and export control. A navigation-grade unit is the size of a shoebox and costs $50k–200k, the optical cavity requires precision machining and ultra-clean assembly, and the whole category is tightly controlled under Wassenaar and ITAR.

Pick an RLG inertial unit for aircraft, ships, and vehicles that need to navigate accurately for hours without external references. Below that requirement, a fiber-optic gyro unit is cheaper and quieter, and a MEMS unit is cheaper again by two orders of magnitude with correspondingly worse drift. The design rule is to work backwards from allowed position error at the end of the GNSS-denied period: one nautical mile per hour means a four-hour denied flight ends four miles off, which may be fine for an aircraft and useless for a weapon.

Gyro bias stability 0.001–0.01 °/hr · free-inertial position drift roughly 1 nautical mile per hour · navigation-grade unit about the size of a shoebox · $50k–200k per unit · service life in the tens of thousands of hours · export-controlled under Wassenaar and ITAR.

There is no remote attack on an inertial unit — it senses only its own motion, which is why it anchors every jam-resistant navigation architecture. The vulnerability is intrinsic: drift accumulates without bound, so a long enough denied period defeats it regardless of quality. Temperature changes and vibration degrade bias stability from the specification values. And an inertial unit initialized with a spoofed GNSS position starts from the wrong place and stays wrong, which is why initialization integrity matters as much as gyro quality.

Honeywell's HG9900 and the LN-100 family from Northrop Grumman, the inertial reference systems on essentially every airliner since the 757, the navigation units on most Western combat aircraft, and shipboard inertial navigation on submarines and surface combatants.

The market is mature and concentrated among Honeywell, Northrop Grumman, Safran, and a few others, with high barriers from both manufacturing precision and export control. Prices have not fallen much, because the technology is at its practical limit and volumes are modest. The competitive pressure comes from below: fiber-optic gyros have taken the mid-range, and MEMS units keep improving. RLGs remain dominant where the highest performance in a rugged package is required and cost is secondary.

5 minHow does a Ring Laser Gyroscope function? | What is SAGNAC Effect? | Inertial Guidance SystemJxJ AVIATION · 10k+ views

14 minHow Optical Gyroscopes Work? (Sagnac Effect)Engineering Educator Academy · 5k+ views

11 minInertial Reference System Rising Laser Gyroscope And IRS Of Aircraft | Video 38Airplane Tech Talk · 50k+ views

On the Calculation of an Inertial Navigation System (NASA Technical Reports Server) · Gyroscope Technology and Applications: A Review in the Industrial Perspective (Sensors via PMC)

A fiber-optic gyro uses the same Sagnac effect as a ring laser gyro, but instead of a laser cavity it winds several hundred to several thousand meters of optical fiber into a coil and measures the phase shift between light traveling both ways around it. Sensitivity scales with the total fiber length times the enclosed area, so performance is bought with fiber rather than with precision optical machining. There is no lock-in problem and therefore no dither mechanism, which makes FOGs quieter and mechanically simpler than RLGs.

FOGs cover an unusually wide performance range — from tactical grade at a few degrees per hour up to navigation grade at 0.001 degrees per hour — by changing coil length, so one technology serves many price points. They have no moving parts, tolerate shock well, and cost less than RLGs at comparable performance. The weaknesses are size at the high end, since a navigation-grade coil is physically large, and temperature sensitivity: thermal gradients across the coil produce a spurious phase shift known as the Shupe effect, which requires careful winding patterns and thermal design to suppress.

Pick a FOG when you need good inertial performance in a package that tolerates shock and vibration, at a cost below RLG: missiles, ground vehicles, gimbal stabilization, uncrewed vehicles, and underwater systems where the combination of quiet operation and long-term stability matters. If the application is a small drone or a consumer product, MEMS is two orders of magnitude cheaper. If it is an airliner or a submarine, RLG or a hemispherical resonator gyro is usually chosen for the last increment of performance and long-term stability.

Bias stability from a few °/hr at tactical grade down to 0.001 °/hr at navigation grade · fiber coil several hundred to several thousand meters long · output rates typically 100–1,000 Hz · a few thousand dollars for a tactical unit, above $100k for navigation grade · no moving parts and no dither mechanism.

As with all inertial sensing, it cannot be attacked remotely — only outlasted, since drift grows without bound. Thermal transients degrade real-world performance relative to the specification, and platform vibration at specific frequencies can couple into the coil. Initialization from a corrupted external fix propagates that error indefinitely.

KVH and Emcore FOG products across many vehicle applications, Northrop Grumman's LN-200 tactical unit used on huge numbers of platforms, Safran and iXblue marine and subsea systems, and the inertial units in a wide range of missiles and uncrewed vehicles.

FOGs occupy the broadest slice of the inertial market by volume and value, and the supplier base is more diverse than for RLGs, including significant Chinese, Japanese, and European production. Prices span from a few thousand dollars for tactical units to over $100k for navigation grade. Export control applies above defined performance thresholds, and those thresholds are a recurring point of friction because commercial autonomous-vehicle demand sits close to the controlled boundary.

3 minWhat is a Fiber Optic Gyroscope (FOG)? | GuideNavGuideNav Official | Inertial Navigation Solutions · 50k+ views

5 miniXblue insights - Fiber-Optic GyroscopesiXblue · 50k+ views

6 minFiber Optical & Ring Laser Gyro working principle. What is Sagnac effect ?Aviation.Animation · 10k+ views

Customer support library (KVH Industries) · Initial Alignment of a Strapdown Inertial Reference and Navigation System (NASA Technical Reports Server)

A MEMS inertial measurement unit puts three vibrating-structure gyroscopes and three accelerometers on silicon dies a few millimeters across. Gyros work by sensing the Coriolis force on a vibrating proof mass; accelerometers sense the displacement of a suspended mass. Consumer units cost under a dollar and drift tens of degrees per hour. Tactical-grade industrial units at $500–5,000 reach 0.1–3 degrees per hour, and the best available MEMS devices are now approaching 0.01 degrees per hour, which begins to overlap the bottom of the FOG range.

Size, cost, and power are the advantages, and they are decisive for anything small: an entire IMU fits in a package smaller than a fingernail and draws milliwatts. The whole class of small drones and guided munitions exists because of it. The weakness is drift. Bias instability and random walk mean a MEMS-only navigation solution degrades in minutes rather than hours, so it must be aided by GNSS, vision, magnetometers, or an air-data or wheel-odometry reference. Performance also varies significantly with temperature, and calibration quality separates a good product from a bad one more than the raw die does.

Use MEMS inertial sensing on anything small, cheap, or numerous, and design the architecture to correct it with something else. For a drone that navigates visually, a MEMS unit bridges between camera fixes perfectly well. For a munition with a 60-second flight, a good tactical MEMS unit gives adequate accuracy on its own. If the requirement is unaided navigation for more than a few minutes, MEMS will not do it and you need FOG or better. The practical selection question is usually bias instability and the quality of the vendor's temperature calibration, not the headline noise figure.

Consumer parts under $1, drifting tens of °/hr · tactical-grade industrial units $500–5,000 at 0.1–3 °/hr · the best available MEMS approaching 0.01 °/hr · package smaller than a fingernail, drawing milliwatts · unaided navigation degrades in minutes rather than hours · performance improving roughly an order of magnitude per decade.

Drift is the intrinsic limit and it is fast. Acoustic attacks are a real vulnerability at this scale — high-intensity sound at the resonant frequency of the proof mass can inject false readings, and this has been demonstrated against drone IMUs in laboratory conditions. Temperature swings and mechanical shock degrade calibration. As with all inertial units, a corrupted initial position propagates indefinitely.

Bosch, STMicroelectronics, and InvenSense consumer parts in every phone and drone, Analog Devices' ADIS industrial family, Honeywell's HG4930 tactical unit, and the inertial measurement units in guided artillery, small munitions, and virtually all commercial autonomous systems.

This is the most commoditized inertial technology by far, driven by phone and automotive volumes that dwarf all defense demand. Prices continue to fall while performance improves roughly an order of magnitude per decade, which is steadily eroding the FOG market from below. The strategic consequence is significant: navigation-capable inertial sensing has become cheap and globally available, which removes what used to be a meaningful barrier to building accurate guided weapons.

14 minHow do MEMS gyroscopes work ?nanolearning · 50k+ views

6 minHow does an Accelerometer work? | 3D AnimationCircuitBread · 100k+ views

12 minIntro to Inertial Measurement Units (IMU)MicWro Engr · 10k+ views

SmallSat Precision Navigation with Low-Cost MEMS IMU Swarms (NASA Technical Reports Server) · Reliability of MEMS inertial devices in mechanical and thermal environments: A review (Heliyon via PMC)

A cold-atom sensor laser-cools a cloud of atoms to microkelvin temperatures, splits their quantum wavefunctions with laser pulses, lets the two paths follow slightly different trajectories, and recombines them. The interference pattern depends on acceleration and rotation along those paths, and it is referenced to the atoms' internal transition frequency and the laser wavelength — both fundamental constants. That is the key property: the scale factor does not drift, because it is set by physics rather than by a manufactured component. Demonstrated bias stability is 100 to 1,000 times better than the best conventional gyros.

Drift-free operation is the promise, and if realized in a deployable package it would let a vehicle navigate for weeks without any external fix. Laboratory instruments have already demonstrated accelerometer and gyro performance well beyond conventional sensors, and gravimeters based on the same principle are commercially sold. The weaknesses are engineering: current systems are large, need vacuum chambers, multiple stabilized lasers, and magnetic shielding, and they are sensitive to vibration and stray fields. Measurement rates are low, typically a few hertz, so they must be combined with a conventional IMU that handles fast dynamics while the atomic sensor corrects its drift.

Consider cold-atom inertial sensing where very long GNSS-denied navigation is required and size is not critical: submarines first, then large ships and aircraft. Sea trials of quantum inertial navigation have been conducted by the UK and France, and submarines are the natural first application because they already carry large navigation systems and cannot receive GNSS while submerged. For anything small or cost-sensitive, conventional inertial sensing plus another absolute reference is the practical answer and will remain so for a long time.

Atoms laser-cooled to microkelvin temperatures · demonstrated bias stability 100–1,000× better than the best conventional gyros · measurement rate a few hertz, so a conventional IMU still covers fast dynamics · hardware still rack-scale: vacuum chamber, several stabilized lasers, magnetic shielding · scale factor set by atomic transition frequency and laser wavelength, so it does not drift · deployable navigation units years to a decade away.

Nothing attacks it remotely. The practical limits are engineering ones: vibration and platform dynamics degrade the interferometer, magnetic and thermal environments must be controlled, and low update rates mean the hybrid conventional sensor still contributes error between atomic measurements. Cost and size will keep it off most platforms regardless of performance.

UK Ministry of Defence and Imperial College quantum navigation sea trials, France's ONERA and iXblue programs, DARPA's atomic sensing efforts, commercial cold-atom gravimeters from Muquans and AOSense, and the atomic clocks that use closely related physics and are already deployed widely.

Funding is heavily governmental, with national quantum technology programs in the UK, France, US, China, and Australia treating inertial navigation as a flagship application. Commercial revenue today comes mostly from gravimeters for mineral exploration and geophysics rather than from navigation. The realistic timeline for deployable navigation systems is measured in years to a decade, and the pacing items are laser packaging, vibration isolation, and power — none of which are physics problems.

4 minQuantum ‘compass’ could allow navigation without relying on satellitesImperial College London · 50k+ views

13 minAtom Interferometry - David PritchardSerious Science · 5k+ views

21 minThe Genius Behind the Quantum Navigation BreakthroughDr Ben Miles · 1m+ views

UK National Quantum Technologies Programme (UKRI) · Atom Interferometry for Fundamental Physics and Gravity Measurements in Space (NASA Technical Reports Server)

Visual-inertial odometry tracks features across camera frames and fuses that with inertial measurements to estimate the platform's motion. The camera supplies drift-free orientation and scaled translation as long as it can see textured surroundings; the inertial unit fills gaps, handles fast rotation, and resolves scale. The result is a relative position estimate that drifts far more slowly than inertial alone — typically 0.1–2% of distance travelled — and needs no external signal. It has become the standard indoor and GPS-denied navigation method for drones and ground robots.

Cost and accuracy are both good: a camera and a MEMS IMU cost tens of dollars and deliver drift rates that would otherwise require a far more expensive inertial unit. It works indoors, underground, and under jamming. The weaknesses are visual: textureless walls, darkness, fog, smoke, and repetitive patterns break feature tracking, and fast motion causes blur. Drift is relative rather than absolute, so error accumulates over distance with no bound unless the system revisits a known place (loop closure) or gets an absolute fix from scene matching or GNSS.

Use visual-inertial odometry on any small autonomous vehicle that needs to navigate without GNSS over short to medium distances: indoor drones, warehouse robots, subterranean vehicles, and the terminal phase of strike drones. Combine it with an absolute reference — scene matching against satellite imagery, or occasional GNSS — for anything flying more than a few kilometers, because unbounded relative drift will otherwise dominate. In smoke, dust, or darkness, add a thermal camera or a radar-based alternative, since visual tracking is the first thing to fail in those conditions.

Drift typically 0.1–2% of distance travelled · camera plus MEMS IMU costs tens of dollars · camera frames at 20–60 Hz against IMU samples at 100–1,000 Hz · relative position only, with unbounded drift unless loop closure or an absolute fix bounds it · no useful output in darkness, smoke, or against featureless surfaces.

Darkness without a thermal or illuminated camera stops it. Smoke, dust, and fog break feature tracking, and these are common in exactly the environments where GNSS is also denied. Featureless surfaces and repetitive patterns cause tracking failures or wrong associations. Rapid motion blurs frames. Deliberate laser dazzling of the camera works. Drift accumulates without bound, so a long enough route degrades the solution regardless of conditions.

Skydio's autonomy stack, DJI's vision positioning systems, the navigation on Ingenuity, the Mars helicopter, ArduPilot and PX4 visual-odometry integrations, Intel RealSense and similar modules, and the subterranean robots developed under the DARPA Subterranean Challenge.

The hardware is commodity and the value is entirely in the software, which has been substantially democratized by open-source implementations like ORB-SLAM and VINS-Mono and by well-supported commercial SDKs. That makes the capability cheap and widely available, and it is a major reason that GNSS jamming is a less complete counter to small drones than it was a few years ago. Differentiation now comes from robustness in degraded conditions rather than from basic capability.

7 minWhat is Visual Inertial Odometry (VIO)?ModalAI · 5k+ views

3 minVisual Slam vs. Visual Odometry – How Does Visual Odometry Work?Inertial Sense, Inc. · 10k+ views

12 minVisual Odometry Series - Part 1 (Concept and Math)Hummingbird · 5k+ views

Machine Learning Based Crater Detection for Terrain Relative Navigation (NASA Technical Reports Server) · Ingenuity Mars Helicopter (NASA JPL)

A Doppler velocity log points four acoustic beams at the seabed and measures the Doppler shift of the returns to derive the vehicle's velocity over ground in three axes. Integrating that velocity gives position with drift proportional to distance travelled rather than to time, which is a much better error growth law than free inertial navigation. Coupled with an inertial unit, a DVL is the backbone of underwater navigation: the inertial unit provides attitude and short-term dynamics, and the DVL bounds its velocity error continuously.

It is the only practical way to get accurate velocity over ground underwater, where GNSS does not penetrate and dead reckoning from a propeller log is badly affected by currents. Accuracy of 0.1–0.3% of distance travelled when bottom-locked is achievable. The weaknesses are altitude and bottom conditions: the beams must reach the seabed, so bottom lock is lost above roughly 30–200 m altitude depending on frequency, and in deep water the vehicle must either fly low or switch to water-track mode, which measures speed relative to the water and therefore inherits the current error the DVL was meant to remove.

Fit a DVL on any uncrewed underwater vehicle or submarine that needs accurate navigation, and design the mission so the vehicle can maintain bottom lock during the portions where accuracy matters. If the vehicle transits deep water far from the bottom, plan for degraded water-track performance and periodic surface GNSS fixes or acoustic positioning updates. For very long missions, pair it with a high-grade inertial unit and consider gravity or magnetic map navigation for absolute bounding, since DVL error still grows with distance.

Velocity-over-ground accuracy 0.1–0.3% of distance travelled when bottom-locked · bottom lock out to roughly 30–200 m altitude depending on frequency · acoustic frequencies typically 300 kHz to 1.2 MHz, with lower frequency buying altitude at the cost of resolution · four beams in a Janus configuration · about $15k for a compact unit, well over $100k for deep-rated long-range systems.

Loss of bottom lock in deep water is the main practical limitation, and it is a physical constraint rather than an attack. Soft sediment and steep terrain degrade returns. Water-track mode is contaminated by currents. The acoustic emission is detectable by a listening opponent, which matters for covert operations. Deliberate acoustic interference is possible but rarely relevant, since the geometry required is demanding.

Teledyne RDI Workhorse and Pathfinder DVLs, Nortek and Sonardyne units on commercial and military uncrewed underwater vehicles, the navigation suites on survey AUVs and on large military UUVs, and DVL-aided inertial systems on submarines.

DVLs are a stable, moderately concentrated market serving offshore energy, hydrographic survey, and defense. Prices run from about $15k for compact units suitable for small vehicles up to well over $100k for long-range deep-rated systems. The growth driver is the expansion of uncrewed underwater vehicles for survey, infrastructure inspection, and military use, and compact low-cost DVLs aimed at small vehicles have been the main product development of the past few years.

6 min5-Minute Guide to a Sonar Series: What is a Doppler Velocity Log?China Sonar · under 1k views

3 minDoppler Speed LogIf I were a MARINER · 1k+ views

3 minThe Reef: Water Linked DVL A50Blue Robotics · 5k+ views

Marine technology solutions (Teledyne Marine) · Science and technology for exploration (NOAA Ocean Exploration)

An infrared homing seeker detects the heat a target radiates and steers the missile to keep the target centered. Early generations used a single detector behind a spinning reticle that chopped the incoming energy into a modulated signal encoding the target's position — simple, cheap, and easily fooled by a flare. Later all-aspect seekers moved to cooled detectors sensitive in the 3–5 µm band, which detect airframe heating and plume emission from the front as well as the exhaust from behind. The seeker itself is a small gimbal, a detector, a cooler, and processing that fits in a 70–130 mm diameter missile nose.

Fire-and-forget operation with no emissions is the advantage: the launching platform can turn away immediately, and the target gets no radar warning. Seekers are compact and comparatively cheap. The weaknesses are weather and countermeasures. Cloud, rain, and humidity cut acquisition range substantially, and the whole engagement depends on thermal contrast that varies with aspect, altitude, and engine setting. Older reticle seekers are defeated by simple flares; that arms race has run for fifty years and each generation of seeker discrimination has been followed by a new countermeasure technique.

Use infrared homing for short and medium-range air-to-air engagements, man-portable air defense, and any application where the launcher must not emit. If the engagement is beyond visual range or in poor weather, active radar homing is the answer. For helicopter and slow-aircraft targets in clear conditions, infrared remains extremely effective, and its passive nature means the first warning a target gets is often the missile itself. Assume the target has some form of infrared countermeasure and select a seeker generation accordingly, since the capability gap between reticle and imaging seekers is very large.

Cooled all-aspect seekers work in the 3–5 µm band · seeker fits a 70–130 mm missile nose · engagement ranges roughly 4–6 km for man-portable systems and 20–35 km for short-range air-to-air · seeker cost from tens of thousands of dollars for MANPADS to a few hundred thousand for an imaging air-to-air unit · zero emissions from the launching platform · acquisition range down substantially in cloud, rain, and humidity.

Flares defeat older seekers, and modern flare formulations tailored to specific seeker bands are effective against many newer ones. Directed infrared countermeasures use a modulated laser to inject false guidance commands into the seeker, and these are now fitted to many transport and rotary aircraft. Engine exhaust suppression and cooling reduce signature. Cloud and heavy rain limit acquisition. Aspect matters: a head-on shot against a subsonic aircraft has much less signal than a tail chase.

The AIM-9 Sidewinder family from the original reticle designs through the imaging AIM-9X, the Russian R-73, the FIM-92 Stinger and its rosette-scan seeker, and the many man-portable systems derived from these lineages.

Infrared seekers are the largest-volume guided weapon seeker category and among the most widely proliferated, with production in the US, Europe, Russia, China, and several other countries. Costs run from tens of thousands of dollars for a MANPADS seeker to a few hundred thousand for an imaging air-to-air seeker. The proliferation of man-portable systems into non-state hands is a standing policy concern, and the countermeasure market — directed infrared countermeasures for civil and military aircraft — is a substantial business in its own right.

8 minHow Heat Seeking Missile Works I Aim 9 SidewinderAiTelly · 500k+ views

12 minThe Genius Engineering of the AIM-9 SidewinderReuben T · 100k+ views

8 minHow Do Heat-Seeking Missiles Find Their Target?History of Simple Things · 10k+ views

Homeland Security: Protecting Airliners from Terrorist Missiles (Congressional Research Service) · Fundamentals of Naval Weapons Systems, Chapter 10: Visible and Infrared Spectrum (US Naval Academy)

An imaging infrared seeker carries a focal plane array rather than a single detector, so it sees a picture rather than a modulated brightness. That changes the problem from tracking a hot spot to recognizing a shape. The seeker can reject a flare because a flare does not look like an aircraft, it can select a specific aimpoint on a vehicle — engine deck, turret ring — and it can perform automatic target recognition against stored templates. Resolution of 256×256 to 640×512 in the 3–5 µm or 8–12 µm band is typical.

Countermeasure resistance and aimpoint selection are the advantages, and they are large: an imaging seeker rejects the flares and much of the directed-infrared jamming that defeat reticle seekers, and hitting a specific part of a target changes lethality substantially. It also enables true fire-and-forget attack against ground targets. The weaknesses are cost and weather. A cooled imaging seeker costs several times a reticle seeker, and every infrared band still loses range in cloud, rain, and smoke. Automatic target recognition performance against unexpected target configurations remains the soft spot, and it is hard to test honestly.

Pick an imaging infrared seeker where countermeasure resistance or precise aimpoint selection justifies the cost: top-attack anti-armor missiles, modern air-to-air missiles, and precision strike weapons against distinct targets. If the target is a large emitter in clear weather and cost matters, a simpler seeker will do. If weather is the dominant risk, pair imaging infrared with millimeter-wave radar in a dual-mode seeker — the radar finds the target through obscurants and the imager confirms the aimpoint, which is now the standard architecture for high-end anti-armor weapons.

Focal plane array 256×256 to 640×512 · 3–5 µm or 8–12 µm band · detector cooled to roughly 77 K by a Stirling cooler · dual-mode versions pair it with a 94 GHz millimeter-wave radar · several times the cost of a reticle seeker, dominated by the array and its cooler.

Multispectral smoke and obscurants block it. Thermal signature management reduces contrast, and decoys designed to present a plausible thermal shape rather than just a hot spot are effective if well made. High-power directed infrared countermeasures can still dazzle or damage the array. Automatic target recognition can be defeated by presenting targets in configurations outside its training set, which is the practical version of an adversarial attack and is much easier than the academic literature makes it sound.

Javelin's imaging seeker, the AIM-9X and IRIS-T, the Spike family, Brimstone's dual-mode seeker, the Python-5, and the terminal seekers on several loitering munitions.

Imaging seeker cost is dominated by the cooled focal plane array and its cooler, which are the same components as high-end thermal imaging and share that concentrated supply base. Prices have fallen as detector production has matured, which has moved imaging seekers from exquisite weapons into much broader use, including relatively inexpensive loitering munitions. Export control remains strict, and the seeker is often the item that gates a weapon transfer.

15 minThe Insane Engineering of the JavelinReal Engineering · 5m+ views

6 minJavelin Missile | How a Javelin Missile worksAiTelly · 1m+ views

Fundamentals of Infrared Detector Materials (SPIE) · Infrared System Test and Evaluation at APL (Johns Hopkins APL Technical Digest)

An anti-radiation seeker homes on the radar it is attacking. A wideband receiver and interferometer antenna set in the missile nose measures the direction of arrival of the target emitter and steers the missile down the bearing. Because the target is broadcasting a strong signal, acquisition range is long — well beyond the radar's own detection range against the launching aircraft in many cases. Modern seekers store a threat library so they can be told which emitter type to attack, and better ones include an inertial memory mode that continues to the last known position if the radar shuts down.

Long standoff and complete passivity are the advantages, and the weapon imposes a dilemma on the defender: keep the radar on and be attacked, or turn it off and stop defending. That dilemma is worth as much as the kills. The weaknesses are what happens when the radar shuts down. Without a memory mode the missile goes ballistic; with one, it hits the last known location, which may be empty if the radar moved. Decoy emitters are cheap and effective. Discriminating the intended radar from other emitters in a dense environment is a real processing challenge.

Use anti-radiation weapons in suppression and destruction of enemy air defenses, and understand that the operational effect is usually suppression rather than destruction — forcing radars off the air is the point. If the target radar is expected to shut down quickly, combine the anti-radiation weapon with GPS or millimeter-wave terminal guidance so the weapon can still hit the site. Loitering anti-radiation drones invert the timing problem entirely by waiting in the area until the radar transmits, which is a much better match to the actual behavior of a disciplined air-defense operator.

Wideband interferometer seeker giving bearing, not range · acquisition range often longer than the target radar's own detection range against the launching aircraft · standoff commonly quoted above 100 km, more on extended-range types · several hundred thousand to over a million dollars per round · decoy emitters that draw a shot cost a small fraction of that · counter-tactic timescale in minutes: transmit briefly, shut down, displace.

Emission control defeats it: the standard counter is to transmit briefly, shut down, and displace within minutes. Decoy emitters mimicking the radar are cheap and draw missiles onto empty ground. Placing radars near protected sites exploits collateral-damage rules. Multistatic and passive radar architectures give nothing to home on. Blinking several radars in sequence confuses seekers that expect continuous emission.

The AGM-88 HARM and its AARGM and AARGM-ER developments, the ALARM with its loiter parachute mode, the Kh-31P and Kh-58, and loitering anti-radiation systems such as the Harpy and Harop.

Anti-radiation missiles are expensive — several hundred thousand to over a million dollars each — and inventories are small relative to the number of emitters a campaign would face. That cost asymmetry against cheap decoy emitters is the central economic problem of the mission, and it is a large part of why loitering munitions and cheap one-way attack drones have become attractive for the same role. The seeker itself is a wideband receiver, and improvements track the same digital receiver technology as electronic support systems.

18 minSEAD - Suppression of Enemy Air DefenseCovert Cabal · 500k+ views

23 minInvisible Battlefield: Jamming & Anti-Radiation MissilesMilitary History not Visualized · 10k+ views

Defense Primer: Electronic Warfare (Congressional Research Service) · Anti-Radiation Missiles vs. Radars (International Journal of Electronics and Telecommunications)

A semi-active laser seeker homes on laser energy reflected off a target that someone else is designating. Four quadrant detectors behind a lens compare received energy to derive an error signal, and the seeker steers to null it. The designator pulses in a coded pattern that the seeker is set to accept, so multiple engagements can run simultaneously in the same area without cross-talk. The seeker itself is simple and cheap — a lens, four detectors, and modest electronics — which is why laser guidance kits are the least expensive way to make a bomb or rocket precise.

Accuracy is the headline: circular error of one to two meters, better than GPS guidance, and it works against moving targets because the designator simply keeps the spot on them. Cost is very low, with laser guidance kits for rockets running a few thousand to $30k. The weaknesses are the dependency chain. Someone has to hold the designation through the terminal phase, which means an exposed observer or an orbiting aircraft, and cloud, smoke, or dust between designator and target breaks the engagement. Designation also alerts the target through laser warning receivers.

Use semi-active laser guidance for precision against moving targets, in tight collateral-damage situations, and wherever cost per round matters and someone can designate. It is the standard for armed helicopters, close air support, and guided rockets. If the weather is poor or nobody can maintain line of sight, GPS or inertial guidance works through cloud at lower accuracy. Dual-mode weapons carrying both are increasingly standard because the two failure modes are almost completely disjoint.

Circular error 1–2 m, better than GPS guidance · four quadrant detectors behind one lens · designator at 1,064 nm, pulse-coded at roughly 10–20 Hz so several engagements can run at once · guidance kits from a few thousand dollars to about $30k, with the APKWS section around $25k · line of sight from designator to target required through the terminal phase.

Smoke and obscurants break the beam and are cheap and fast to deploy. Laser warning receivers give the target seconds to react, and some vehicles slew countermeasures onto the designator. Cloud below the launch aircraft blocks it. Moving out of line of sight after designation begins defeats a weapon already in flight. Decoy laser spots, though rarely used, would work against seekers that accept any correctly coded energy.

The Paveway family, Hellfire and its many derivatives, APKWS and other laser guidance kits for 70 mm rockets, the Russian Krasnopol and Kitolov guided projectiles, and the laser seekers on numerous loitering munitions.

This is the cheapest precision guidance available, and the APKWS case is the clearest illustration: bolting a $25k guidance section between an existing rocket motor and warhead converts a cheap unguided rocket into a precision weapon, and it has since been adapted as a low-cost air-defense interceptor. That cost-per-effect argument dominates the segment. Production is broad, the technology is mature, and the main innovation is in packaging the seeker into ever cheaper and smaller munitions.

2 minTwo-minute tech: Paveway IV bomb, the RAF's go-to weaponBFBS Forces News · 50k+ views

5 minBAE's Advanced Precision Kill Weapon SystemAviationWeek · 100k+ views

7 minThe AGM-114 Hellfire & How it Operates | Munitions of Battle Scenario Fulfillment · 10k+ views

Defense Primer: U.S. Precision-Guided Munitions (Congressional Research Service) · TRADOC Pam 34-3, Joint Laser Designation Procedures (US Army Training and Doctrine Command)

An active radar homing seeker carries its own radar transmitter and receiver in the missile nose. Typically the missile flies most of the way on inertial navigation with datalink updates from the launching platform, then turns its seeker on in the last 10–20 km and completes the engagement autonomously. That mid-course-plus-terminal architecture is what makes beyond-visual-range air combat work: the launching aircraft can turn away once the missile goes active, instead of holding an illumination lock all the way to impact.

Fire-and-forget at long range is the decisive advantage, and it works in weather that stops infrared. The seeker gives range and closing velocity directly, which supports fuzing and endgame maneuvering. The weaknesses are cost, aperture, and countermeasures. A missile nose is a small antenna, so seeker detection range is limited and the mid-course phase must deliver the missile into a small acquisition basket. DRFM jamming that generates coherent false targets is effective against many seekers, and low-observable targets are hard for a small X-band aperture. The seeker also announces the terminal phase, giving the target warning to maneuver.

Use active radar homing for beyond-visual-range air-to-air missiles, anti-ship missiles, and long-range surface-to-air systems where all-weather fire-and-forget performance is needed. If the engagement is within visual range and clear, an imaging infrared seeker is cheaper and gives no terminal warning. Against heavily jammed targets, consider a dual-mode seeker with a passive home-on-jam mode, which turns the jammer into a beacon. Getting the mid-course guidance right matters at least as much as seeker performance, since a good seeker pointed at empty sky finds nothing.

Seeker goes active in the last 10–20 km · X-band, with an antenna typically 150–200 mm across on an air-to-air missile · mid-course on inertial navigation plus datalink updates from the launching platform · engagement ranges from tens of kilometers to well past 100 km on current types · seeker among the most expensive components in the missile, dominated by the RF front end and antenna.

DRFM repeater jamming creates false targets and range-gate pull-off. Towed decoys present a more attractive return than the aircraft. Chaff still works against some seekers. Notching — flying perpendicular to the missile so Doppler shift falls into the clutter rejection filter — is a classic and still effective maneuver. Low observability reduces seeker acquisition range enough to break the endgame. Breaking the mid-course datalink leaves the missile searching in the wrong place.

The AIM-120 AMRAAM, MBDA Meteor, the R-77, PL-15, the seekers on Harpoon and most modern anti-ship missiles, and the active seekers on Patriot PAC-3 and Aster interceptors.

Active seekers are among the most expensive components in a missile, and the cost is dominated by the RF front end and the antenna. The important trend is AESA seekers, made viable by falling transmit-receive module prices, which give better jam resistance and multiple simultaneous functions in the same nose volume. Several countries now field or are fielding AESA seekers, and the module cost curve is what determines how far down the missile inventory that capability spreads.

21 minActive Radar Homing - The Guidance of the AMRAAM, MICA, R77 etc.Millennium 7 * HistoryTech · 50k+ views

16 minTRACKING: how missile sensors follow their targets (AMRAAM, Sidewinder and all the others)Millennium 7 * HistoryTech · 10k+ views

8 minThe AIM-120 AMRAAM and How it Works.| munitions of battle Scenario Fulfillment · 10k+ views

Defense Primer: U.S. Precision-Guided Munitions (Congressional Research Service) · Radar Principle (Radar Tutorial)

Semi-active radar homing puts the transmitter on the launching platform and only a receiver in the missile. The ground radar or aircraft illuminates the target continuously, and the missile homes on the reflected energy, using a rear-facing reference antenna to compare against the direct signal for Doppler processing. Because the illuminator can be far more powerful than anything that fits in a missile nose, effective range is long. The cost is that the launching platform must keep the target illuminated until impact, which ties it to the engagement and exposes it.

Illuminator power gives long range at a much lower missile cost than an active seeker, and the missile itself is simpler and cheaper. For ground-based air defense with a large fixed radar, this is an efficient arrangement. The weakness is the illumination requirement: the launching platform cannot maneuver freely or engage other targets during the missile's flight, and it is emitting continuously, which makes it an obvious target for anti-radiation weapons and electronic support. The number of simultaneous engagements is limited by the number of illumination channels.

Semi-active homing still makes sense for ground-based air defense where a powerful illuminator already exists and missile cost matters, and for legacy fleets. For new air-to-air designs it has been superseded almost everywhere by active homing, because the tactical penalty of holding a lock is severe. If you are assessing a system that uses it, the questions worth asking are how many illumination channels it has and how quickly the illuminator can displace after firing, because both determine how it performs against a real attack rather than a single target.

Receiver only in the missile, transmitter on the launching platform · illumination held continuously from launch to impact · rear reference antenna for the Doppler comparison against the direct signal · engagement ranges typically 20–70 km on fielded types · simultaneous engagements capped by the number of illumination channels, usually a handful · missile cost well below an equivalent active-seeker round.

Jamming the reflected signal or the illuminator is effective, and noise jamming against a continuous-wave illuminator is comparatively easy. Notching to fall inside the Doppler clutter filter works. Chaff decoys the seeker. Most importantly, forcing the illuminator off the air with anti-radiation weapons or the threat of them breaks every missile in flight simultaneously, which is a systemic vulnerability that active seekers do not share.

The AIM-7 Sparrow and its Sea Sparrow derivatives, the R-27R, the Hawk and early Patriot engagement modes, most S-300 variants in their older configurations, and many naval point-defense systems.

This is a legacy architecture in slow decline, sustained by large fielded inventories and by the low missile cost that makes it attractive for high-volume ground-based air defense. Modernization programs generally replace it with active seekers as module costs fall. Its continued presence in export inventories means that suppression tactics developed against it remain operationally relevant for a long time yet.

17 minSparrow, Skyflash, Aspide: the Age of Semi Active Radar HomingMillennium 7 * HistoryTech · 10k+ views

14 minFrom failure to redemption | The AIM-7 SparrowAustralian Military Aviation History · 100k+ views

Radar Principle (Radar Tutorial) · Defense Primer: U.S. Precision-Guided Munitions (Congressional Research Service)

In beam-riding guidance the launcher points a beam — historically radar, now almost always laser — at the target and the missile flies up the center of that beam. A rearward-facing detector on the missile senses its position within the beam's cross-section and steers to stay centered. The missile carries no seeker at all, only a detector looking backward at the launcher, which makes it very cheap and essentially immune to countermeasures aimed at forward-looking seekers. The beam is usually coded and shaped so the missile can compute its offset precisely.

Cost and jam resistance are the advantages. Because the missile looks backward at a known beam rather than forward at the target, decoys, flares, and target-side jamming have nothing to work against, and the missile itself can cost a fraction of a homing weapon. Accuracy is good at short range. The weaknesses are range and geometry. Beam divergence means guidance accuracy degrades with distance, the launcher must track the target continuously through impact, and the operator is exposed. Smoke and obscurants between launcher and target break guidance, and the technique does not scale to long range.

Pick beam-riding for short-range anti-tank and air-defense missiles where cost per round is the dominant consideration and the launcher can maintain track: infantry anti-armor systems, vehicle-mounted short-range air defense, and light naval point defense. If the engagement is beyond a few kilometers or the launcher must not remain exposed, use a homing seeker. Beam-riding is also attractive where countermeasure resistance matters more than fire-and-forget convenience, since active protection systems and soft-kill countermeasures have far less to attack.

Engagement ranges typically 4–8 km · rearward-facing detector only, with no forward-looking seeker · flight times from a few seconds to roughly 15 s, with the launcher tracking throughout · cost roughly a third to a half of a comparable homing missile · accuracy falls off with range as the beam diverges.

Smoke and obscurants block the beam. The launcher must remain exposed and tracking, so suppressive fire against the operator defeats the engagement. Laser warning receivers detect the beam and cue countermeasures or evasive movement. At longer ranges beam divergence degrades accuracy. Against fast crossing targets, the requirement to keep the launcher's beam on the target through flight time limits engagement geometry.

The Russian 9M119 Refleks and Kornet anti-tank missiles, the Swedish RBS 70 short-range air defense system, the Starstreak with its three darts riding a laser beam, and the older radar beam-riding systems such as Sea Slug and the original Nike Ajax.

Beam-riding weapons are cheap relative to homing missiles — often a third to a half the cost — and remain in production for exactly that reason, particularly for infantry and short-range air defense. The Starstreak case is notable because its beam-riding architecture makes it very difficult to counter with directed infrared countermeasures, which has kept a 1990s design operationally relevant. Expect this niche to persist wherever cost per engagement and countermeasure resistance beat fire-and-forget convenience.

5 minStarStreak Missiles: Ukraine's Weapon That’s Three Times the Speed of Sound | WSJ EquippedThe Wall Street Journal · 500k+ views

13 minThe Kornet-E: How This Russian Missile Actually Destroys TanksSYG DESIGNWORKS · 50k+ views

4 minPortable Air Defence Demo: How To Use Saab's RBS 70 NGBaltic Defence Review · 5k+ views

Radar Principle (Radar Tutorial) · The Talos Guidance System (Johns Hopkins APL Technical Digest)

Command guidance keeps the intelligence off the missile. A ground or airborne system tracks both the target and the missile, computes the intercept, and sends steering commands over a link — radio, fiber-optic wire, or a datalink from the launching aircraft. The missile carries only a receiver and actuators. Fiber-optic guided missiles trail a spool of fiber and give the operator a video feed from the missile's nose plus full control, which turns the weapon into a remotely piloted vehicle that happens to explode.

The missile is cheap because the expensive sensing and processing stay on the launcher and get reused. Man-in-the-loop control allows aimpoint changes, target reassignment, and abort in flight, which matters enormously under restrictive rules of engagement. Fiber-optic links are essentially unjammable. The weaknesses are link dependence and operator workload. Radio command links can be jammed. Fiber spools limit range and can snag on terrain. The operator must stay engaged through the whole flight, so one operator generally handles one missile, which limits the rate of engagement.

Use command guidance where cost per round matters, where a human decision is needed late in the engagement, or where jam resistance through a physical link is worth the range limit. Fiber-optic guided drones and missiles have become prominent in current conflicts precisely because they defeat the electronic warfare that stops radio-controlled drones. If you need fire-and-forget or high engagement rates, a homing seeker is necessary. For air defense, command guidance from a capable ground radar remains an economical way to field large numbers of cheap interceptors.

Fiber-optic spool range typically 10–25 km · wire-guided TOW range about 3.75–4.5 km · spools cost tens to hundreds of dollars · one operator per missile in flight, which caps the engagement rate · missile carries only a receiver and actuators, so unit cost sits at the low end of guided weapons · fiber links cannot be jammed at all.

Radio command links are jammable, and this is the standard counter. Fiber-optic links cannot be jammed but the fiber can break, snag, or be cut, and the range is set by the spool length, typically 10–25 km. Killing or suppressing the operator ends the engagement. Command systems that must track both missile and target need line of sight to both, so terrain masking works. Engagement rate limits mean saturation attacks overwhelm the system.

The TOW's wire-guided command link, the fiber-optic Spike NLOS and the Russian and Ukrainian fiber-optic FPV drones now widely used, command-guided air-defense systems such as the Rapier and parts of the S-300 engagement sequence, and the mid-course datalinks on modern air-to-air missiles that are technically a command-guidance phase.

Command guidance is the cheap end of guided weapons and it has become newly important because fiber-optic control is the most effective available answer to drone jamming. Spools cost tens to hundreds of dollars, and the resulting drone is immune to the electronic warfare that has otherwise dominated the small-drone fight. That has produced rapid adoption on both sides in Ukraine and a corresponding scramble to develop counters, which currently amount to physical interception rather than electronic attack.

15 minBGM-71 TOW Anti-tank missile | WIRE GUIDED WONDERMatsimus · 100k+ views

18 minRapier Air Defense Missile SystemMatsimus · 100k+ views

14 minWhat are Fiber Optic FPV Drones, and Why are They so Deadly? - Feat. @DefenseTechforUkraineNorth Wind Aerial · 5k+ views

Defense Primer: U.S. Precision-Guided Munitions (Congressional Research Service) · Guidance of Missiles, Chapter 8: The Line of Sight Guidance Law (NPTEL)