How Military Satellites Actually Work: From Launch to Orbital Intelligence

Military satellites are among the most expensive, most classified, and most consequential machines ever built. A single reconnaissance satellite programme can consume billions of euros and take a decade from concept to first image. The engineering constraints are severe: the hardware must survive launch vibration, operate in hard vacuum with temperature swings of 300 degrees Celsius, tolerate constant radiation bombardment, and function autonomously for years without physical maintenance. And all of this must produce intelligence products that are militarily useful within minutes of collection.

Despite the classification barriers, the underlying physics and engineering principles are well understood. Orbital mechanics is textbook material. Radiation effects on silicon are studied in open literature. Infrared detection physics has been published for decades. What makes military satellite engineering interesting is not secret physics but the extreme optimisation of known principles under constraints that commercial systems rarely face.

This post covers the full stack: orbits, bus architecture, radiation hardening, imaging sensors, infrared missile warning, signals intelligence, military communications, constellation design, launch, and the growing threat environment in orbit.

1. Orbits and Why They Matter for Military Missions

Every satellite mission begins with orbital selection, and the choice of orbit determines almost everything about the satellite's design, capability, and cost. The governing equation is Kepler's third law, which relates orbital period to altitude:

T = 2 * pi * sqrt(a^3 / mu)
 
where:
  T  = orbital period (seconds)
  a  = semi-major axis (metres), equal to Earth's radius + altitude for circular orbits
  mu = Earth's gravitational parameter = 3.986 * 10^14 m^3/s^2

For a circular orbit at altitude h above Earth's surface, the semi-major axis is a = R_E + h, where R_E = 6,371 km.

Low Earth Orbit (LEO): 200 to 2,000 km

LEO is the workhorse altitude for reconnaissance. At 400 km altitude, the orbital period is approximately 92.5 minutes, giving roughly 15.5 orbits per day. The satellite moves at about 7.67 km/s relative to the ground. The key advantage is proximity: a telescope at 400 km altitude resolves far finer detail than the same optic at 36,000 km.

Most imaging satellites operate in sun-synchronous orbits (SSO), a special class of LEO where the orbital plane precesses at exactly one degree per day to match Earth's motion around the Sun. This guarantees consistent solar illumination angle on every pass, which is critical for change detection. An SSO at 500 km altitude requires an inclination of approximately 97.4 degrees.

The European CSO (Composante Spatiale Optique) constellation, operated by France's DGA, uses SSO. CSO-1 launched in December 2018 into an orbit near 800 km for wide-area surveillance. CSO-2 operates at a lower altitude near 480 km for higher resolution imaging. CSO-3, launched in 2023, completes the constellation.

The drawback of LEO is limited access time. A satellite at 500 km altitude has a ground footprint roughly 2,400 km in diameter, and it traverses that footprint in about five minutes. Revisit time for a single satellite is typically 12 to 24 hours depending on latitude and swath width.

Medium Earth Orbit (MEO): 2,000 to 35,786 km

MEO is primarily used for navigation constellations. GPS operates at approximately 20,200 km altitude with an orbital period of about 11 hours 58 minutes (half a sidereal day), arranged in six orbital planes with four or more satellites each. Europe's Galileo constellation orbits at approximately 23,222 km with an orbital period of about 14 hours 5 minutes, using three orbital planes inclined at 56 degrees.

The choice of altitude for navigation satellites is driven by geometry: you need at least four satellites visible from any point on Earth at any time, and MEO altitudes provide the right balance between constellation size and visibility geometry. Higher orbits mean fewer satellites needed, but longer signal propagation delay and larger positional dilution of precision.

Military GPS receivers use the encrypted P(Y) code on L1 (1575.42 MHz) and L2 (1227.60 MHz) frequencies, plus the newer M-code on L1 and L2 designed specifically for military operations with improved anti-jam capability and a more directional broadcast pattern. Galileo's Public Regulated Service (PRS) provides a similar encrypted, access-controlled positioning capability for European governments.

Geostationary Earth Orbit (GEO): 35,786 km

GEO is where the satellite's orbital period matches Earth's rotation, so the satellite appears stationary relative to the ground. The orbital altitude is precisely 35,786 km for a circular equatorial orbit. Three GEO satellites spaced 120 degrees apart can see most of Earth's surface between about 70 degrees north and south latitude.

GEO is used for two major military applications: early warning (infrared missile detection) and communications. The SBIRS (Space Based Infrared System) GEO satellites stare continuously at entire hemispheres, watching for the infrared signature of ballistic missile launches. Military communication satellites like the UK's Skynet 5 constellation and France's Syracuse IV operate from GEO because the fixed geometry simplifies ground terminal pointing and provides continuous coverage of a theatre of operations.

The trade-off is distance. Signal propagation to GEO takes approximately 120 milliseconds one way, creating a minimum round-trip latency of about 240 milliseconds. For imaging, the resolution penalty is severe: at 35,786 km, the same telescope that achieves 0.3-metre ground sample distance from 500 km would produce roughly 21-metre GSD, which is inadequate for tactical reconnaissance.

Highly Elliptical Orbits (HEO): Molniya and Tundra

Russia has a persistent problem that GEO satellites cannot solve: coverage above 70 degrees north latitude, which includes most of its territory. The Molniya orbit addresses this with high eccentricity (approximately 0.74), a 12-hour period, an inclination near 63.4 degrees (chosen to eliminate apsidal precession), and an apogee over the Northern Hemisphere around 39,000 to 40,000 km.

A satellite in Molniya orbit spends roughly eight hours near apogee, moving slowly and providing extended dwell time over high-latitude regions. Two satellites in Molniya orbits phased 180 degrees apart can provide near-continuous coverage of the Arctic and northern latitudes.

The United States operates the SBIRS HEO payloads, which are infrared sensors hosted on satellites in highly elliptical orbits to provide missile warning coverage of northern launch areas that GEO sensors see at extreme look angles.

Tundra orbits are a 24-hour variant with similar inclination but a period of one sidereal day, providing a ground track that dwells over one hemisphere.

2. Satellite Bus Architecture

The satellite bus is the platform that keeps the payload alive and pointed. Military satellite buses are built to tighter reliability and survivability standards than most commercial platforms, but the fundamental subsystems are the same.

Power

Solar arrays are the primary power source for every military satellite outside of a few nuclear-powered deep-space missions. Modern triple-junction gallium arsenide solar cells (InGaP/GaAs/Ge) achieve efficiencies around 30 to 32 percent. Airbus Defence and Space and Leonardo (formerly Selex) manufacture cells and array assemblies for European military programmes.

A typical GEO communications satellite deploys solar arrays generating 15 to 20 kilowatts at beginning of life. LEO reconnaissance satellites generally require less power, typically 2 to 8 kilowatts, because their payloads operate intermittently during passes over target areas.

Batteries store energy for eclipse periods. GEO satellites experience eclipses near equinoxes lasting up to 72 minutes. LEO satellites at 500 km altitude experience roughly 35-minute eclipses every 92-minute orbit. Lithium-ion batteries have largely replaced nickel-hydrogen in modern designs, with specific energy around 150 to 200 Wh/kg at the cell level.

The power bus typically runs at 28 V or 100 V, with a power conditioning and distribution unit (PCDU) managing the interface between arrays, batteries, and payload loads. Thales Alenia Space builds PCDUs for multiple European military satellite programmes.

Attitude Determination and Control (ADCS)

Pointing accuracy is critical. An imaging satellite at 500 km with a 0.3-metre GSD needs to know where it is pointing to within a few microradians. The attitude determination chain typically includes:

• Star trackers: optical cameras that identify star patterns and compute attitude to approximately 1 to 5 arcseconds (5 to 25 microradians). Sodern (a Safran subsidiary) manufactures the SED36 star tracker used on numerous European military and scientific missions.
• Inertial measurement units (IMUs): fibre-optic or ring-laser gyroscopes that measure angular rates. These provide high-bandwidth attitude information between star tracker updates.
• Sun sensors and Earth sensors: coarser sensors used for safe mode and initial acquisition.

Attitude control uses a combination of:

• Reaction wheels: spinning flywheels that exchange angular momentum with the spacecraft body. Three or four wheels in a redundant configuration allow three-axis pointing. Typical wheels store 20 to 75 Nms of angular momentum.
• Control moment gyroscopes (CMGs): gimballed spinning rotors that provide higher torque than reaction wheels for the same mass. CMGs are used on larger, more agile platforms that need rapid slew capability, such as reconnaissance satellites that must slew between targets during a single pass.
• Magnetic torquers: electromagnets that interact with Earth's magnetic field to desaturate reaction wheels. Only effective in LEO where the field is strong enough.
• Thrusters: used for coarse attitude control, orbit maintenance, and momentum management at GEO where magnetic torquers are ineffective.

For an imaging satellite that needs to slew 30 degrees between targets in under 25 seconds, the required torque is substantial. The moment of inertia of a 3,000 kg satellite with deployed solar arrays can easily exceed 5,000 kg*m^2. Achieving a 30-degree slew in 25 seconds (approximated as a bang-bang profile) requires peak angular acceleration of roughly:

theta = 0.5 * alpha * (t/2)^2   (half the slew in half the time, then decelerate)
alpha = 4 * theta / t^2
 
For theta = 0.524 rad (30 deg), t = 25 s:
alpha = 4 * 0.524 / 625 = 0.00335 rad/s^2
 
Torque = I * alpha = 5000 * 0.00335 = 16.8 Nm

This is well within CMG capability but beyond typical reaction wheel output, which explains why agile reconnaissance satellites favour CMGs.

Thermal Control

In LEO, the Sun-facing side of a satellite can reach +150 degrees Celsius while the shadow side drops to -150 degrees Celsius. Thermal management uses passive and active techniques:

• Multi-layer insulation (MLI): the characteristic gold or silver foil blankets that reduce radiative heat exchange.
• Radiators: panels with high emissivity coatings that reject waste heat to space.
• Heat pipes: two-phase systems that transport heat from hot components to radiators.
• Heaters: resistive heaters keep components above their minimum operating temperatures during eclipse or cold phases.
• Louvers: mechanically actuated covers over radiators that adjust the effective radiating area. These are used on some military GEO buses.

Optical payloads have especially tight thermal requirements. The primary mirror of an imaging telescope must maintain its figure to fractions of a wavelength (tens of nanometres) across temperature variations. This typically requires active thermal control with stability better than 0.1 degrees Celsius over orbital cycles.

Propulsion

Military satellites use several propulsion types:

• Monopropellant hydrazine: the traditional workhorse, with specific impulse around 220 to 230 seconds. Simple, reliable, well-characterized. Used for orbit maintenance and attitude control on many European military platforms.
• Bipropellant (MMH/NTO): higher specific impulse around 310 to 320 seconds. Used for larger orbit manoeuvres, including the apogee motor that circularises GEO transfer orbits.
• Electric propulsion: Hall-effect or ion thrusters with specific impulse of 1,500 to 3,000 seconds. Increasingly used for station-keeping on GEO military communications satellites. The Safran PPS-5000 Hall thruster is used on the Syracuse IV satellites. Electric propulsion saves substantial propellant mass but produces very low thrust (typically 50 to 300 millinewtons), making it unsuitable for rapid manoeuvres.

Propellant mass is a primary driver of satellite mass. A GEO communications satellite carrying 15 years of station-keeping propellant (chemical) might allocate 25 to 30 percent of its launch mass to fuel. Electric propulsion can reduce this significantly.

3. Radiation-Hardened Electronics

Space is a radiation environment that would destroy commercial electronics within weeks or months. Understanding this environment and the engineering countermeasures is central to military satellite design.

The Radiation Environment

Three primary radiation sources affect satellites:

Van Allen radiation belts: two toroidal regions of trapped charged particles (primarily protons and electrons) held by Earth's magnetic field. The inner belt, centred near 1.5 Earth radii (about 3,000 km altitude), is dominated by energetic protons with energies up to several hundred MeV. The outer belt, centred around 4 to 5 Earth radii, contains primarily electrons. LEO satellites below 1,000 km pass beneath the inner belt but are exposed in the South Atlantic Anomaly (SAA), where the belt dips to as low as 200 km altitude. GEO satellites sit outside the outer belt but are exposed to solar energetic particles.

Solar particle events (SPEs): the Sun periodically releases massive bursts of protons and heavier ions during coronal mass ejections and solar flares. A severe SPE can deliver a proton fluence of 10^10 protons/cm^2 (for energies above 10 MeV) over a few days. The October 1989 event was among the worst recorded.

Galactic cosmic rays (GCRs): very high-energy ions originating outside the solar system. Individual GCR ions can have energies exceeding 10^18 eV. While the flux is low (a few particles per cm^2 per second), the high linear energy transfer (LET) of heavy ions like iron nuclei can deposit enormous charge in a single transistor gate.

Radiation Effects on Electronics

Total Ionising Dose (TID): cumulative damage from ionising radiation, measured in rads or grays (1 gray = 100 rads). TID shifts transistor threshold voltages, increases leakage currents, and eventually causes functional failure. Commercial CMOS processes fail at TID levels of 5 to 20 krad(Si). A GEO satellite accumulates roughly 10 krad per year behind 2 mm of aluminium shielding. Military satellites require electronics rated to 100 krad to 1 Mrad depending on mission duration and orbit.

Single-Event Upsets (SEUs): a single energetic particle (proton or heavy ion) deposits enough charge in a memory cell to flip its state. SEU rates vary with orbit, shielding, and device technology, but a typical rate for an unprotected SRAM in LEO might be 10^-7 to 10^-6 upsets per bit per day. For a system with millions of bits, this means multiple bit flips per day.

Single-Event Latchup (SEL): a parasitic thyristor in CMOS is triggered by particle impact, creating a low-impedance path between power supply and ground. If not detected and cleared quickly (by power cycling), the resulting current can destroy the device.

Displacement Damage: protons and neutrons displace atoms from crystal lattice positions, degrading minority carrier lifetime in semiconductor materials. This is particularly damaging for solar cells and optocouplers.

Radiation Hardening Techniques

Hardened-by-design (HBD): circuit-level techniques that make transistors inherently tolerant. These include guard rings around NMOS transistors to prevent leakage paths, enclosed layout transistors (ELTs) where the gate completely surrounds the drain, and special oxidation processes that resist charge trapping. BAE Systems manufactures several lines of radiation-hardened processors and ASICs, including the RAD750 processor used on numerous military and deep-space missions.

Hardened-by-process (HBP): manufacturing processes specifically developed for radiation tolerance. Silicon-on-insulator (SOI) technology reduces latchup susceptibility by placing transistors on an insulating layer that interrupts the parasitic thyristor path. The BAE Systems RH1020 and more recent RHFL devices use SOI processes.

Triple Modular Redundancy (TMR): three identical logic chains process the same inputs, and a majority voter selects the output. If one chain suffers an SEU, the other two outvote it. The cost is roughly three times the silicon area and power. TMR is used extensively in FPGAs deployed in space; Xilinx (now AMD) and Microchip (formerly Microsemi) both offer space-grade FPGAs with built-in TMR or TMR-friendly architectures.

Error Detection and Correction (EDAC): memories are protected with codes that detect and correct bit flips. Single-error-correct, double-error-detect (SECDED) Hamming codes are standard. More advanced Reed-Solomon or BCH codes protect larger memory arrays. Every radiation-hardened memory system in a military satellite uses EDAC.

The performance penalty is significant. The RAD750, based on a PowerPC 750 core, runs at 132 to 200 MHz and delivers roughly 266 MIPS. Compare that to a modern commercial processor at several GHz. Radiation-hardened parts are typically two to three technology generations behind commercial silicon.

European manufacturers include Atmel (now Microchip) in Nantes, France, which produces the AT697 and AT7913 SPARC-based processors for space applications, and Cobham Advanced Electronic Solutions (now part of Renesas) which has produced radiation-hardened memory devices.

4. Electro-Optical Imaging Satellites

Reconnaissance imaging is the most visible (if that word can apply to classified systems) military satellite mission. The physics of space-based imaging is constrained by fundamental optics.

Resolution and the Diffraction Limit

The theoretical angular resolution of a circular aperture is:

theta = 1.22 * lambda / D
 
where:
  theta  = angular resolution (radians)
  lambda = wavelength (metres)
  D      = aperture diameter (metres)

The Ground Sample Distance (GSD) is the angular resolution projected onto the ground:

GSD = theta * h = 1.22 * lambda * h / D
 
where:
  h = orbital altitude (metres)

For visible light at 550 nm wavelength, a 2.4-metre aperture (the reported size of the KH-11 primary mirror, matching the Hubble Space Telescope) at 250 km altitude:

GSD = 1.22 * 550e-9 * 250e3 / 2.4 = 0.070 metres = 7 cm

At 500 km altitude, the same system gives 14 cm GSD. Atmospheric turbulence further limits practical resolution, though military systems employ sophisticated image processing to partially compensate.

Pushbroom Sensors

Modern imaging satellites use pushbroom (linear array) sensors rather than the older whiskbroom (scanning mirror) design. A pushbroom sensor has a linear CCD or CMOS array aligned perpendicular to the satellite's velocity vector. As the satellite moves, each line of pixels sweeps across the ground, building up a two-dimensional image.

The advantages are significant: longer integration time per pixel (improving signal-to-noise ratio), no moving scan mirror, and simpler mechanics. The integration time per pixel for a pushbroom sensor is:

t_int = GSD / v_ground
 
For GSD = 0.3 m and ground velocity approximately 7 km/s (accounting for Earth's rotation):
t_int = 0.3 / 7000 = 43 microseconds

This is far longer than a whiskbroom scanner would allow for the same GSD, which translates directly to better signal-to-noise ratio.

Real Systems

KH-11 / CRYSTAL (USA): the long-running US optical reconnaissance satellite programme, operated by the National Reconnaissance Office (NRO). The KH-11 series has been in operation since the late 1970s, with the current block sometimes referred to as "Enhanced CRYSTAL" or EIS (Evolved Imagery Satellites). The primary mirror is widely reported as 2.4 metres in diameter based on the 2011 NRO donation of two surplus telescope assemblies to NASA, which were confirmed to have 2.4-metre primary mirrors. Lockheed Martin is the primary contractor. Typical orbits are between 250 and 1,000 km, with the satellite manoeuvring to lower perigee for high-resolution passes.

Helios (France/Europe): the Helios 1 and Helios 2 series were France's first dedicated military imaging satellites, built by Matra Marconi Space (now Airbus Defence and Space). Helios 2A (launched 2004) and 2B (launched 2009) operated in SSO at approximately 680 km altitude with reported resolution around 35 cm in optical and about 1 metre in infrared. These have been succeeded by CSO.

CSO (France): the Composante Spatiale Optique is the current French military reconnaissance constellation, built by Airbus Defence and Space and Thales Alenia Space. The system uses three satellites:

• CSO-1 at approximately 800 km for wide-area surveillance
• CSO-2 at approximately 480 km for very high resolution (reportedly sub-20 cm GSD)
• CSO-3 at approximately 800 km with infrared and optical capability

The CSO system cost is estimated at approximately 3.6 billion euros for development and the first operational capability.

Pleiades Neo (Airbus): while primarily a commercial/dual-use system, Pleiades Neo is instructive. These are 30 cm GSD optical satellites in 620 km SSO, weighing approximately 920 kg. The telescope has a 65 cm aperture. Airbus Defence and Space manufactures and operates these from Toulouse. The four-satellite constellation provides revisit times under one day. Several European military customers use Pleiades Neo imagery.

OPTSAT-3000 (Israel/Italy): built by Israel Aerospace Industries (IAI) for Italy's defence ministry, operating since 2017 in SSO at about 450 km. The system reportedly achieves sub-50 cm GSD using the IAI Jupiter space telescope. This illustrates the international trade in military reconnaissance capability.

Ofek series (Israel): IAI builds Israel's Ofek reconnaissance satellites, launched retrograde (westward) from Palmachim Air Base to avoid overflying neighbouring countries. Ofek-16, launched in 2020, is the latest publicly acknowledged electro-optical variant. The retrograde orbit costs significant launch performance but is a hard operational constraint.

Spectral Bands

Military imaging satellites do not just capture visible light. Most carry multispectral imagers operating across several bands:

The shortwave infrared (SWIR) band is particularly valuable militarily because many camouflage materials that appear natural in visible light are readily distinguishable in SWIR.

5. Infrared Missile Warning

Detecting ballistic missile launches from space is one of the most demanding sensor problems in military engineering. The requirement is to detect, characterise, and report a missile launch within seconds of ignition, from a sensor platform 36,000 km away.

The Physics of Infrared Detection from GEO

A ballistic missile in boost phase produces an exhaust plume with temperatures between 1,500 and 2,500 Kelvin, radiating strongly in the short-wave infrared (SWIR, 2 to 3 micrometres) and mid-wave infrared (MWIR, 3 to 5 micrometres) bands. The peak emission wavelength from Wien's displacement law:

lambda_max = 2898 / T  (micrometres, with T in Kelvin)
 
For T = 2000 K: lambda_max = 1.45 micrometres
For T = 1000 K (later in boost): lambda_max = 2.9 micrometres

The plume is brightest against the cold space background when viewed from above. From GEO altitude, the signal must be distinguished from Earth's own infrared emission, sunglint, cloud patterns, and various false alarm sources (gas flaring, forest fires, volcanic eruptions).

DSP Legacy

The Defence Support Program (DSP) satellites, first launched in 1970 and operating through the 2000s, used a scanning infrared sensor. The satellite spun at 6 RPM, and a linear infrared detector array swept across the Earth disk with each rotation. This "spinning sensor" approach was simple and reliable but had limited sensitivity and spatial resolution. DSP could detect large ICBM launches reliably but struggled with smaller, shorter-burn tactical ballistic missiles.

DSP satellites were built by Northrop Grumman (originally TRW) and operated in GEO. A total of 23 DSP satellites were launched over the programme's lifetime.

SBIRS Architecture

The Space Based Infrared System replaced DSP with far more capable sensors. SBIRS has two components:

GEO satellites: Lockheed Martin builds the SBIRS GEO payload, which carries two distinct infrared sensors:

• Scanning sensor: inherits the DSP concept but with far greater sensitivity. It uses a large-format infrared focal plane array and a scanning mirror to survey the full Earth disk.
• Staring sensor: a step-stare or continuous-stare system that can dwell on specific regions of interest. The staring mode provides much higher sensitivity and temporal resolution, enabling detection of dimmer, shorter-duration events like tactical ballistic missile launches and even some types of aircraft.

The combination matters. The scanning sensor provides broad area surveillance, while the staring sensor provides focused, high-sensitivity coverage. An initial detection by the scanner can trigger the starer to focus on the region for characterisation.

SBIRS GEO-1 through GEO-6 have been launched as of early 2026. The follow-on programme, Next-Generation Overhead Persistent Infrared (Next-Gen OPIR), is being developed by Lockheed Martin (GEO component) and Northrop Grumman (polar orbit component), with first launches scheduled for the late 2020s.

HEO payloads: SBIRS HEO sensors are hosted payloads on classified satellites in highly elliptical orbits, providing coverage of northern high-latitude launch areas. The specific host satellites are not publicly identified, but the orbits are understood to be similar to Molniya type.

Detection Chain

The timeline from launch to warning is measured in seconds:

1. Missile motor ignites, producing infrared plume
2. Infrared photons travel to GEO sensor (approximately 120 ms propagation)
3. Sensor detects above-threshold signal in one or more pixels
4. On-board processing characterises the event (distinguishing from false alarms)
5. Alert message is transmitted to ground station
6. Ground processing correlates data from multiple sensors
7. Warning is disseminated to command authorities

The total time from launch to initial detection can be under 60 seconds. Ground processing and dissemination add additional time, but the goal is to provide warning within 90 to 120 seconds of launch, giving decision-makers minutes of warning time for ICBMs with 25 to 30 minute flight times.

European Capabilities

France has developed limited space-based infrared capability through the SPIRALE (Systeme Preparatoire Infrarouge pour l'ALErte) technology demonstrator, launched in 2009. SPIRALE consisted of two microsatellites in HEO that collected infrared data to characterise the background and develop detection algorithms. The follow-on operational system, part of the broader French defence space strategy, is still in development.

6. SIGINT and ELINT Satellites

Signals intelligence from space involves intercepting electromagnetic emissions, whether communications (COMINT) or radar and other electronic systems (ELINT), from orbital vantage points. This is among the most classified military satellite missions, but the physics and general architecture are understood.

The Challenge: Signal Strength from Orbit

A terrestrial radio transmitter radiates a signal that follows the inverse-square law. By the time that signal reaches LEO (say 500 km), it has spread over an enormous area. The received power at the satellite is:

P_r = P_t * G_t * G_r * (lambda / (4 * pi * R))^2
 
where:
  P_t = transmitter power
  G_t = transmitter antenna gain toward the satellite (usually low, since the signal is aimed at the ground)
  G_r = satellite receive antenna gain
  R   = slant range to satellite
  lambda = wavelength

For a mobile phone transmitting 0.25 watts at 900 MHz (lambda = 0.33 m) toward a satellite at 500 km range with a 20 dBi satellite antenna, the received power is roughly:

P_r = 0.25 * 1 * 100 * (0.33 / (4 * pi * 500000))^2
    = 0.25 * 100 * (8.35e-8)^2 / ... 

The numbers work out to something on the order of 10^-16 to 10^-15 watts, which is extremely weak but detectable with sensitive receivers and large antennas. This is why SIGINT satellites carry some of the largest antenna structures ever deployed in space.

Large Deployable Antennas

The US Mentor/Orion programme (also known as Advanced Orion) operates from GEO and reportedly deploys mesh antennas with diameters estimated at 100 metres, making them among the largest structures in orbit. These satellites are built by Northrop Grumman and launched on heavy-lift vehicles to GEO. The large antenna provides the gain needed to intercept weak signals from 36,000 km altitude.

The antenna diameter directly determines the system's angular resolution for geolocation and its sensitivity. For a 100-metre antenna operating at 1 GHz (lambda = 0.3 m):

Beamwidth = 1.22 * lambda / D = 1.22 * 0.3 / 100 = 0.00366 radians = 0.21 degrees

From GEO, 0.21 degrees subtends roughly 130 km on Earth's surface, which is adequate for regional interception but insufficient for precise geolocation from a single satellite. More precise geolocation requires either multiple satellites, time-difference-of-arrival techniques, or frequency-difference-of-arrival analysis.

Geolocation Techniques

Time Difference of Arrival (TDOA): two or more satellites receive the same signal at slightly different times due to different path lengths. The time difference defines a hyperbola on the Earth's surface. Two pairs of satellites provide two hyperbolas whose intersection locates the emitter. Accuracy depends on timing precision and satellite geometry. With nanosecond-level timing from atomic clocks and satellites separated by hundreds of kilometres, geolocation accuracy of a few kilometres is achievable.

Frequency Difference of Arrival (FDOA): relative motion between satellites and emitter causes different Doppler shifts at each satellite. Combined with TDOA, FDOA provides another constraint that improves geolocation accuracy.

European SIGINT Programmes

France operates the CERES (Capacite de Renseignement Electromagnétique Spatiale) constellation, launched in November 2021. CERES consists of three satellites flying in formation in LEO, built by Airbus Defence and Space and Thales. The three-satellite formation enables TDOA and FDOA geolocation of electromagnetic emitters. CERES is designed to detect and characterise radar systems, communication networks, and other military electronic emissions.

Germany operates the SAR-Lupe constellation for radar reconnaissance (synthetic aperture radar rather than SIGINT) and has been developing electronic intelligence capabilities. Italy operates the COSMO-SkyMed SAR constellation, built by Thales Alenia Space, which provides radar imaging useful for all-weather, day/night surveillance.

Israel's ISA (Israel Space Agency) has operated SIGINT capabilities through the TecSAR and other classified programmes, with Israel Aerospace Industries as the primary contractor.

7. Military Satellite Communications

Communication satellites form the backbone of modern military command and control. Without them, deployed forces cannot receive orders, transmit intelligence, or coordinate across theatres. The engineering challenge is providing reliable, high-bandwidth, secure, jam-resistant communications to users ranging from strategic command centres to individual soldiers with small terminals.

Frequency Bands and Their Trade-offs

Higher frequencies allow smaller antennas (for a given beamwidth) and wider bandwidths, but suffer more from rain attenuation and require more precise pointing. EHF is favoured for the most critical military links because the narrow beamwidths make jamming geometrically harder, and the wide bandwidth supports spread-spectrum anti-jam waveforms.

Anti-Jam Techniques

Military SATCOM systems employ several layers of protection:

• Spread-spectrum modulation: the signal is spread across a wide bandwidth, making it difficult to jam without enormous power. Direct-sequence spread spectrum (DSSS) and frequency hopping are both used.
• Narrow spot beams: instead of illuminating a continent with a single beam, the satellite uses a phased-array antenna to create small spot beams that concentrate receive sensitivity and transmit power over small areas. A jammer outside the spot beam is geometrically excluded.
• Nulling antennas: adaptive beamforming that places antenna pattern nulls in the direction of detected jammers. This is computationally intensive but highly effective.
• Low probability of intercept (LPI): waveforms designed to be indistinguishable from noise to an unintended receiver.

Major Systems

MILSTAR / AEHF (USA): the Milstar constellation (five satellites, GEO) provided protected EHF communications from the mid-1990s. The follow-on Advanced Extremely High Frequency (AEHF) system, built by Lockheed Martin, provides order-of-magnitude improvements in data rate while maintaining anti-jam capability. AEHF operates at 44 GHz (uplink) and 20 GHz (downlink) with crosslinks at 60 GHz. The system supports data rates from 75 bps (for small terminals in severe jamming) to approximately 8 Mbps (for large ground stations). Six AEHF satellites were launched between 2010 and 2020.

Skynet 5 (UK): the UK's military communications constellation, operated by Airbus Defence and Space under a Private Finance Initiative (PFI). Skynet 5A, 5B, 5C, and 5D operate in GEO, providing UHF, SHF, and EHF communications to British and allied forces. The Skynet 5 system supports frequency hopping and adaptive antenna nulling. The follow-on Skynet 6A programme, awarded to Airbus Defence and Space, is under development with launch expected in the late 2020s.

Syracuse IV (France): France's Syracuse IV constellation provides strategic and tactical SATCOM for French forces. Syracuse IV-A launched in October 2021 and IV-B in 2023, both built by Thales Alenia Space and Airbus Defence and Space. The satellites operate in GEO and carry military X-band and Ka-band payloads with anti-jam features. The platform uses the Spacebus Neo bus with electric propulsion (Safran PPS-5000 Hall thrusters) for orbit raising and station-keeping.

SATCOM BW (Germany): Germany's SATCOMBw programme uses two dedicated military satellites (COMSATBw-1 and COMSATBw-2) in GEO, launched in 2009 and 2010, built by Thales Alenia Space and Astrium (now Airbus Defence and Space). These provide SHF and UHF military communications. The follow-on programme, Heinrich Hertz, includes both military and experimental communications payloads.

SICRAL (Italy): Italy's SICRAL-1 and SICRAL-2 satellites provide military UHF, SHF, and EHF communications. SICRAL-2, launched in 2015, was jointly funded by Italy and France.

Data Rates and Link Budgets

A simplified link budget for a military GEO SATCOM link illustrates the engineering trade-offs:

Transmitter: ground terminal with 1.2 m antenna, 20 W at X-band (8 GHz)
Slant range: 36,000 km (GEO minimum, directly overhead)
Free space path loss: 20 * log10(4 * pi * R / lambda)
  lambda = 0.0375 m
  FSPL = 20 * log10(4 * pi * 3.6e7 / 0.0375) = 201.6 dB
Ground antenna gain: 37.5 dBi (1.2 m dish at 8 GHz)
Satellite receive antenna gain: 35 dBi (large reflector with feed array)
EIRP: 10 * log10(20) + 37.5 = 50.5 dBW
Received power: 50.5 - 201.6 + 35 = -116.1 dBW
 
For a noise temperature of 500 K:
Noise power density: k * T = 1.38e-23 * 500 = 6.9e-21 W/Hz = -201.6 dBW/Hz
C/N0 = -116.1 - (-201.6) = 85.5 dB-Hz
 
For QPSK with FEC at Eb/N0 = 5 dB:
Max data rate = 10^(85.5/10) / 10^(5/10) = 3.55e8 / 3.16 = ~112 Mbps (theoretical)

In practice, atmospheric losses, rain fade margins, modulation inefficiency, and required margins reduce achievable rates substantially. But this shows why larger antennas and higher power translate directly to capability.

8. Constellation Design and Revisit Time

A single satellite in LEO provides a brief, periodic look at any given location. Military operations often require more frequent revisit, approaching continuous surveillance of certain areas. This drives constellation design.

Revisit Time Fundamentals

For a single satellite in a circular orbit at altitude h, the swath width (W) of its sensor determines how much ground it covers per pass. The revisit time at the equator for a single satellite is approximately:

T_revisit = Earth's circumference at equator / (W * orbits per day)
 
For W = 15 km (narrow swath, high-resolution mode), orbits per day = 15:
T_revisit = 40,075 / (15 * 15) = 178 hours = 7.4 days

This is far too long for operational needs. Increasing the number of satellites is the direct solution.

Walker Constellations

Robert Walker formalised the notation for symmetric circular constellations: Walker i:T/P/F, where:

• i = inclination
• T = total number of satellites
• P = number of equally spaced orbital planes
• F = phasing factor (determines the relative spacing of satellites in adjacent planes)

GPS uses a Walker 55:24/6/1 constellation (approximately). Galileo uses Walker 56:24/3/1 (when complete).

For a reconnaissance constellation wanting daily revisit at the equator with a 15 km swath, you would need roughly:

N = T_revisit_single / T_desired = 178 / 24 = ~7.4 satellites minimum

In practice, you need more because coverage is not uniform, and you want some redundancy. A constellation of 12 to 16 LEO imaging satellites can provide sub-daily revisit for most of Earth's surface.

Sun-Synchronous Considerations

Military imaging constellations overwhelmingly use sun-synchronous orbits. The specific local time of the ascending node (LTAN) is chosen based on mission requirements:

• Morning pass (10:00 to 10:30 LTAN): low Sun angle provides long shadows that reveal vertical structure (building heights, vehicle types). Preferred for many reconnaissance tasks.
• Noon pass (12:00 to 13:00 LTAN): minimum shadows, maximum illumination. Good for multispectral analysis but poor for interpretation of vertical features.
• Consistent pass time: Sun-synchronous orbits ensure the same location is always imaged at the same local time, making change detection more reliable because lighting conditions are constant.

Real Constellation Sizes

Most nations cannot afford large dedicated reconnaissance constellations. The table below shows approximate sizes of known military or dual-use imaging constellations:

The small number of dedicated military satellites is why nations increasingly purchase commercial imagery from operators like Airbus (Pleiades, SPOT), Maxar (WorldView), and Planet Labs to supplement national systems.

SAR Satellites: All-Weather Reconnaissance

Synthetic aperture radar (SAR) satellites deserve mention because they provide imagery regardless of cloud cover or daylight. Italy's COSMO-SkyMed constellation (Thales Alenia Space) and Germany's SAR-Lupe system (OHB) both use X-band SAR. The second-generation COSMO-SkyMed (CSG) satellites achieve sub-metre resolution in spotlight mode.

SAR works by transmitting radar pulses and recording the reflected signal. The "synthetic aperture" is created by using the satellite's motion to simulate a much larger antenna: as the satellite moves along its orbit, it records returns from multiple positions, and coherent processing combines these returns to achieve azimuth resolution equivalent to a physical antenna hundreds of metres long.

The azimuth resolution of a SAR system is:

delta_az = D / 2
 
where D = physical antenna length in the along-track direction

This counterintuitive result (smaller antenna gives better resolution) arises because a shorter antenna has a wider beam, illuminating a ground target for a longer time and providing a longer synthetic aperture. For COSMO-SkyMed, with a physical antenna approximately 5.7 metres long, the finest azimuth resolution is approximately 2.85 metres, further improved in spotlight mode by electronically steering the beam to dwell on a target.

9. Launch and Deployment

Getting military payloads to orbit presents unique challenges beyond the normal difficulty of space launch.

Dedicated Versus Rideshare

Military satellites almost never ride-share with commercial payloads. The reasons are operational security (the payload's mass, size, and orbital parameters reveal information about its capabilities), schedule assurance (the military cannot wait for a ride-share manifest to fill), and orbital specificity (military orbits are chosen precisely for the mission, not for commercial convenience).

The exception is hosted payloads, where a military sensor rides on another satellite. SBIRS HEO payloads are the primary example, riding on classified host satellites.

Launch Vehicles

Ariane 5 / Ariane 6 (Europe): Ariane 5 has launched numerous European military satellites, including Syracuse III, Skynet 5, and SICRAL. Ariane 6, which entered service in 2024, continues this role with both A62 (two solid boosters) and A64 (four solid boosters) configurations. Ariane 6 can deliver approximately 5,000 kg to GTO in the A62 configuration and 11,500 kg in A64. ArianeGroup and ESA have worked to ensure European sovereign access to space, meaning Europe can launch military satellites without depending on non-European launch services.

Vega / Vega-C (Europe): the smaller European launcher, built by Avio, is suitable for LEO military payloads. Vega-C can deliver approximately 2,200 kg to a 700 km SSO. CSO and CERES satellites were launched on Soyuz from Kourou, but Vega-C is the intended European launcher for similar future payloads.

Atlas V / Vulcan (USA): United Launch Alliance's Atlas V has been the primary US military launch vehicle for decades, launching SBIRS, AEHF, and numerous NRO missions. Vulcan Centaur, its successor, achieved first flight in early 2024. ULA is certified for National Security Space Launch (NSSL) missions.

Falcon 9 / Falcon Heavy (USA): SpaceX's Falcon 9 was certified for NSSL in 2018 and has launched multiple GPS III satellites and other military payloads. Falcon Heavy has launched the USSF-44 and USSF-67 missions carrying classified payloads to GEO. The reusability of Falcon 9's first stage has significantly reduced launch costs, though the military pays a premium over commercial missions for mission assurance requirements.

Soyuz (Russia): historically used for French military launches from the Guiana Space Centre (Kourou), but this ended after Russia's 2022 invasion of Ukraine. This created an urgent need for European sovereign launch alternatives, accelerating Ariane 6 and Vega-C development.

Shavit (Israel): Israel's Shavit launcher, built by IAI, is a small solid-fuelled rocket that launches the Ofek reconnaissance satellites westward from Palmachim. Shavit can deliver approximately 350 to 800 kg to its retrograde low orbit. The launch azimuth restriction (westward over the Mediterranean to avoid overflying hostile neighbours) costs roughly 30 percent of payload capacity compared to an eastward launch.

NRO and USSF Launch Programmes

The US National Reconnaissance Office (NRO) is the agency responsible for developing and operating reconnaissance satellites. The NRO does not publicly acknowledge most of its launches until after the payload reaches orbit, and sometimes not even then. NRO missions are designated NROL (NRO Launch) followed by a number. The mission number sequence is not chronological with launch dates, adding a layer of obfuscation.

The US Space Force's Space Systems Command manages the National Security Space Launch programme, which certifies and procures launch services. The current NSSL Phase 2 contracts are split between ULA (Vulcan) and SpaceX (Falcon 9/Heavy), with Phase 3 (Lane 1 for smaller payloads) also competed.

On-Orbit Deployment

Many military satellites have complex deployment sequences:

1. Separation from launch vehicle
2. Solar array deployment
3. Antenna deployment (for SIGINT or communications satellites, this may involve very large deployable structures)
4. Orbit raising (for GEO satellites, typically from geostationary transfer orbit)
5. Attitude acquisition and sensor calibration
6. Commissioning and operational handover

For a large GEO communications satellite using chemical propulsion, orbit raising from GTO to GEO takes a single apogee burn. With electric propulsion (as on Syracuse IV), orbit raising can take several months of continuous low-thrust spiralling. This extended deployment period is a trade-off for the significant mass savings that electric propulsion provides.

10. Vulnerabilities and Space Domain Awareness

The growing dependence on military satellites has made space a contested domain. The ability to deny, degrade, or destroy an adversary's space capabilities is now an active area of military development by several nations.

Anti-Satellite (ASAT) Threats

Direct-ascent kinetic kill: a ground-launched missile that physically collides with a satellite. China demonstrated this in January 2007 by destroying its own Fengyun-1C weather satellite at 865 km altitude, creating over 3,500 pieces of trackable debris that remain a collision hazard. India conducted a similar test (Mission Shakti) in March 2019, though at a lower altitude (approximately 300 km) to reduce long-lived debris. Russia tested its Nudol system against Kosmos-1408 in November 2021, creating over 1,500 trackable fragments.

The debris problem is the critical issue. Each kinetic ASAT test creates a cloud of fragments that threatens all satellites at similar altitudes for decades. The 2007 Chinese test alone increased the LEO debris population by roughly 25 percent.

Co-orbital ASAT: a satellite manoeuvres close to its target and destroys or disables it through collision, explosive fragmentation, or directed energy. Russia's Kosmos-2542 and Kosmos-2543 demonstrated inspection and close approach capabilities in 2020, with Kosmos-2543 reportedly releasing a sub-satellite that manoeuvred near a US reconnaissance satellite.

Electronic warfare: ground-based or space-based jammers can deny satellite communications or GPS navigation without physical destruction. This is far more reversible and less escalatory than kinetic attack. GPS jamming and spoofing have been documented extensively in the eastern Mediterranean, the Baltic region, and around conflict zones. Satellite communications can be jammed by directing sufficient power at the satellite's receive frequency from within its antenna footprint.

Laser dazzling/blinding: ground-based lasers can temporarily saturate or permanently damage electro-optical sensors on reconnaissance satellites. China has reportedly tested ground-based lasers capable of dazzling imaging satellites at LEO altitudes. The energy required to permanently damage an optical sensor from the ground depends on the aperture, wavelength, atmospheric conditions, and range, but multi-kilowatt lasers can pose a genuine threat to LEO assets.

Cyber attacks: satellite ground segments, command-and-control links, and data downlinks are all potential targets for cyber intrusion. In February 2022, the Viasat KA-SAT network was disrupted by a cyber attack attributed to Russia, affecting both commercial and military users across Europe. The attack targeted ground modems rather than the satellite itself, demonstrating that the ground segment is often the weakest link.

Space Situational Awareness (SSA)

Knowing what is in orbit is a prerequisite for defending space assets. Space situational awareness involves tracking every object in orbit, characterising its behaviour, and detecting anomalous manoeuvres.

Space Fence (USA): an S-band radar system on Kwajalein Atoll in the Marshall Islands, operated by the US Space Force. Space Fence can track objects as small as 10 cm in LEO and detects approximately 200,000 observations per day. The system uses a ground-based phased-array radar that creates an electronic "fence" across the sky. Any object passing through the radar's field is detected and tracked.

Space Surveillance Network (USA): the broader US network of radars, optical telescopes, and space-based sensors that collectively maintain a catalogue of over 47,000 objects. This includes mechanical and phased-array radars at multiple sites, optical telescopes at sites like the Ground-Based Electro-Optical Deep Space Surveillance (GEODSS) system in White Sands (New Mexico), Diego Garcia, and Maui.

EU Space Surveillance and Tracking (EU SST): Europe's SST programme, coordinated by the EU Agency for the Space Programme (EUSPA), pools sensors from France (GRAVES radar, TAROT telescopes), Germany (TIRA radar at Fraunhofer FHR), Spain (ESA's Optical Ground Station, S3T radar), and other member states. GRAVES (Grand Reseau Adapte a la Veille Spatiale) is a bistatic radar near Dijon, France, that detects and tracks objects in LEO. TIRA (Tracking and Imaging Radar) at Wachtberg, Germany, operates at L-band and Ku-band and can image space objects with sub-metre resolution.

Space-based sensors: the US operates the SBSS (Space Based Space Surveillance) satellite and the GSSAP (Geosynchronous Space Situational Awareness Program) satellites, which manoeuvre in near-GEO orbits to inspect and characterise objects near the geostationary belt. GSSAP satellites, built by Orbital Sciences (now Northrop Grumman), have been observed approaching other nations' GEO satellites for close inspection.

Orbital Debris: The Long-Term Threat

The Kessler syndrome, first described by NASA's Donald Kessler in 1978, posits that above a critical density of objects in orbit, collisions generate more debris than natural decay removes, creating a cascading chain reaction. Current debris models suggest that some altitude bands (particularly around 800 to 1,000 km) may already be at or near this threshold.

Military implications are direct: LEO reconnaissance satellites operate in exactly the altitude bands most congested with debris. A debris cascade could render these orbits unusable for decades. This creates a paradox where ASAT weapons, by generating debris, could deny the attacker's own use of space.

Debris mitigation guidelines require satellites to deorbit within 25 years of end of life. Active debris removal (ADR) technologies are being developed by ClearSpace (a Swiss company contracted by ESA for the ClearSpace-1 mission) and Astroscale (a Japanese company that demonstrated proximity operations with its ELSA-d mission).

Resilience Through Architecture

The recognition that large, exquisite satellites are vulnerable is driving architectural changes:

Proliferated LEO constellations: instead of a few large, expensive satellites, deploy many smaller, cheaper ones. The loss of any individual satellite has minimal impact on capability. The US Space Development Agency (SDA) is building the Proliferated Warfighter Space Architecture (PWSA), a mesh network of hundreds of small satellites in LEO providing communications, missile tracking, and data transport. Tranche 0 satellites were launched in 2023 and 2024.

Disaggregation: separating functions that were previously combined on one satellite. Instead of one satellite that does imaging, communications, and signals intelligence, build separate specialised platforms. This increases survivability because an adversary must target multiple systems to degrade a capability.

Manoeuvre capability: providing satellites with sufficient propulsion to evade threats. This requires substantial delta-v reserves beyond what is needed for station-keeping. The trade-off is mass and cost.

Rapid reconstitution: maintaining the ability to rapidly launch replacement satellites if orbital assets are lost. This drives interest in responsive launch capabilities and pre-built spare satellites. The US Space Force's Tactically Responsive Space programme aims to launch replacement satellites within days of a loss, though this remains aspirational for large systems.

Closing Notes

Military satellite engineering sits at the intersection of orbital mechanics, materials science, semiconductor physics, signal processing, and systems engineering at the most demanding scale. The constraints are extraordinary: hardware must work for 15 years without repair, survive radiation that would fry consumer electronics in days, and produce actionable intelligence within minutes.

The trend lines are clear. Constellations are getting larger and more distributed. Commercial and military capabilities are converging, with governments purchasing commercial imagery and launching military payloads on commercial rockets. Electronic warfare in space is becoming normalised. And the debris environment is getting worse, creating a collective action problem that no single nation can solve.

For European nations, the strategic lesson of the past decade is that sovereign access to space matters. The loss of Soyuz launches from Kourou after 2022 forced a rapid (and expensive) pivot to European vehicles. The development of indigenous SIGINT (CERES), optical reconnaissance (CSO), and communications (Syracuse IV, Skynet 6A) capabilities reflects a recognition that dependence on non-European suppliers for military space is an unacceptable strategic vulnerability.

The engineering is hard. The physics is unforgiving. The orbital environment is getting more hostile. But for any nation that wants to see what is happening on Earth's surface, listen to what is being transmitted, and communicate securely with its forces, military satellites remain indispensable. Nothing else provides the global, persistent, sovereign access to information that space systems deliver.