Uplink interference geolocation using TDOA and FDOA from GEO pairs

When a terrestrial emitter interferes with a GEO satellite uplink, time and frequency differences across two or more satellites can triangulate its position. This page explains the geometry, the error sources, and what commercial services can and cannot resolve.

Sensors

Intelsat Geolocation Service (GEO fleet): Uses pairs or triplets of Intelsat GEO satellites as passive receivers of the same uplink signal. Published accuracy is typically 10–150 km depending on satellite geometry, signal bandwidth and ephemeris quality. The service is documented in ITU coordination filings and Intelsat technical bulletins.
Kratos Monics platform: A commercial carrier-monitoring and geolocation system that ingests telemetry from GEO transponders. Monics performs TDOA/FDOA cross-correlation across satellite pairs and is widely used by satellite operators for interference management. Kratos publishes platform capability sheets but does not disclose per-event accuracy figures publicly.
SES Interference Management Service (GEO fleet): SES has published case studies through the Satellite Interference Reduction Group (IRG) describing successful geolocation of interferers in Africa and the Middle East using TDOA across SES GEO assets. Reported location errors in those published cases ranged from tens to over a hundred kilometres.
Inmarsat GEO fleet (L-band and Ka-band): Inmarsat's four-satellite GEO constellation provides L-band coverage with sufficient orbital separation in some arc segments to support TDOA geolocation. L-band signals have narrower bandwidth than Ku/Ka carriers, which degrades TDOA cross-correlation sharpness and typically widens position uncertainty.

Two satellites, one emitter, and the geometry that connects them

When a terrestrial transmitter illuminates a GEO satellite, it also illuminates every other GEO satellite whose footprint covers the same ground. If two satellites receive the same uplink signal, the difference in propagation path length produces a measurable time difference of arrival (TDOA). That TDOA constrains the emitter to lie on a hyperboloid of revolution whose foci are the two satellite positions. On the ground, this projects to a hyperbolic arc.

A second independent measurement, frequency difference of arrival (FDOA), exploits the Doppler shift caused by the relative velocity between the emitter and each satellite. GEO satellites are not perfectly stationary: station-keeping manoeuvres and residual orbital eccentricity produce velocities on the order of a few metres per second. That tiny motion creates a measurable frequency offset. The FDOA measurement generates a second family of curves. Where the TDOA hyperbola and the FDOA iso-Doppler curve intersect, the emitter is located. In principle, two satellites and two measurements are sufficient. In practice, the intersection is often poorly conditioned or ambiguous.

Why two satellites are rarely enough

The TDOA/FDOA pair from a single satellite pair produces two candidate intersection points, symmetric about the satellite baseline extended to the ground. Resolving which candidate is real requires either prior knowledge of the emitter's approximate region, a third satellite measurement, or an additional physical constraint such as the emitter being known to lie on land or within a licensed service area.

Geometry matters enormously. Two GEO satellites separated by only a few degrees of arc in the geostationary belt subtend a very small baseline as seen from the ground. Small TDOA values are harder to measure precisely, and the resulting hyperbola is nearly parallel to the satellite arc, producing a long, narrow uncertainty region rather than a compact position fix. Operators deliberately choose satellite pairs with larger angular separations when the coverage geometry allows it. Published Intelsat geolocation guidance recommends a minimum separation of around 10 degrees of arc to achieve operationally useful accuracy, though even then the position ellipse can span tens of kilometres.

A third satellite collapses the ambiguity cleanly. Three pairwise TDOA measurements produce three hyperbolas; their common intersection is unique (subject to measurement noise). This is the standard approach when a third satellite with adequate footprint overlap is available.

Ephemeris error: the limit that signal processing cannot fix

TDOA geolocation accuracy is bounded by how precisely the satellite positions are known at the moment of measurement. A GEO satellite's position is not a fixed point. Station-keeping cycles, solar radiation pressure and gravitational perturbations from the Moon and Sun all move the spacecraft. Ephemeris errors of even a few hundred metres translate directly into TDOA errors and therefore into ground position errors. At GEO altitude (approximately 35,786 km), a 100-metre ephemeris uncertainty in the along-arc direction produces a ground position error that depends on the geometry but can easily reach 10–30 km.

Commercial operators using their own fleets have the advantage of knowing their satellites' positions from ranging data far better than the published two-line element sets available publicly. That internal ephemeris quality is one reason why operator-run services such as Intelsat's and SES's consistently outperform third-party attempts to replicate the calculation from public orbital data alone.

Signal bandwidth and the cross-correlation floor

TDOA is estimated by cross-correlating the signal received at two satellites after compensating for the nominal path delay. The sharpness of the correlation peak, and therefore the precision of the delay estimate, scales with the signal's bandwidth. A wideband Ku-band carrier occupying 36 MHz of transponder bandwidth gives a correlation peak with a theoretical width of roughly 28 nanoseconds. At the speed of light, 28 nanoseconds corresponds to about 8 metres of path length difference. Translated through the geometry to a ground position, the contribution from correlation noise alone might be a few kilometres or less under good conditions.

Narrowband signals tell a different story. A 200 kHz SCPC carrier produces a correlation peak roughly 180 times wider. The TDOA estimate becomes noisy, and the ground position uncertainty grows accordingly, sometimes to hundreds of kilometres. Continuous wave interference, such as an unmodulated carrier, has essentially zero bandwidth and makes TDOA estimation unreliable. FDOA becomes the primary observable in that case, but GEO relative velocities are so small that FDOA precision is also limited. Continuous wave interferers on GEO uplinks are among the hardest targets for this technique.

What the published record shows about operational accuracy

The Satellite Interference Reduction Group has documented numerous geolocation cases in its published interference reports. Across cases involving Ku-band carriers with moderate bandwidth and favourable satellite geometry, reported position accuracies cluster in the 10–50 km range. In less favourable conditions, particularly in regions of Africa and Southeast Asia where satellite arc geometry is poor or only two satellites with narrow separation are available, errors of 100–200 km have been reported.

Those figures are sufficient to identify the country of origin and often the region within a country. They are rarely sufficient to identify a specific building or vehicle without follow-up ground-based direction-finding. The ITU Radio Regulations (Article 15 and associated procedures) require administrations to investigate and suppress harmful interference once an emitter is located to national territory. In practice, the GEO geolocation result triggers a diplomatic or regulatory process; ground-based DF teams complete the final few kilometres.

Satellize incorporates published TDOA/FDOA geolocation outputs into interference attribution workflows for clients who need to correlate RF event timing with other orbital observations, such as SAR or optical passes over candidate sites.

Honest limits and the cases where this technique fails

Mobile interferers, such as a misconfigured VSAT terminal on a ship or truck, can move between the time of geolocation and the time any ground response is mounted. The technique produces a position at the moment of measurement, not a track. Repeated measurements at short intervals can establish motion, but GEO-based TDOA has no inherent revisit advantage: the measurement is continuous while the interference persists, which is actually useful for mobile targets if the processing pipeline is fast enough.

Multipath in urban or mountainous terrain corrupts the signal wavefront arriving at the uplink dish, adding apparent delay that the geolocation algorithm cannot distinguish from true path length difference. Interference events in cities with dense building stock are systematically harder to locate than those from open rural sites. Finally, deliberate spoofing of the uplink signal, such as a retransmission of a legitimate carrier from a different location, can in principle deceive the TDOA measurement entirely. This is rare in practice but not unknown in contested spectrum environments.

Typical figures

Typical ground position accuracy (Ku-band, good geometry)	10–50 km (published operational cases)
Typical ground position accuracy (Ku-band, poor geometry or narrow satellite separation)	100–200 km
Minimum useful signal bandwidth for TDOA	Approximately 1 MHz; accuracy degrades significantly below this
Frequency bands supported	C-band (3.7–4.2 GHz uplink), Ku-band (13.75–14.5 GHz), Ka-band (27–30 GHz), L-band (1.6–1.7 GHz)
Minimum satellite arc separation for useful geometry	~10 degrees (Intelsat published guidance); wider is better
Ambiguity resolution	Requires third satellite or independent geographic constraint; two-satellite solution yields two candidate positions
Ephemeris error contribution to ground position	Tens of kilometres per 100 m of satellite position uncertainty, geometry-dependent
Measurement latency (signal to position estimate)	Minutes to hours depending on operator processing pipeline; near-real-time in automated commercial systems
Continuous wave interferer performance	TDOA unreliable; FDOA-only solution possible but with substantially larger uncertainty

Analytics Satellize can run

Interference event geolocation report	TDOA/FDOA cross-correlation across GEO satellite pair or triplet, hyperbolic intersection	PDF or structured JSON report giving position estimate, uncertainty ellipse, satellite pair used and signal characteristics at time of event
Emitter country-of-origin attribution	Geolocation position overlaid against national boundaries and licensed earth-station registry (ITU MIFR data)	Attribution memo identifying administrations to be notified under ITU Article 15 procedures
Interference timeline and persistence analysis	Time-series of carrier power and TDOA estimates across an interference episode	Annotated event log showing onset, duration, power variation and any apparent emitter movement
Candidate site shortlist for ground-based DF	Geolocation ellipse intersected with road network, earth-station licence database and land-use classification	GIS layer of candidate sites ranked by prior probability, formatted for field team navigation
Multi-event pattern analysis	Clustering of repeated interference geolocations over time to identify habitual or systematic emitters	Summary report with cluster map and recurrence statistics for regulatory or legal proceedings
Geometry feasibility assessment for a new satellite pair	Analytical modelling of TDOA/FDOA dilution of precision across a coverage region for a proposed satellite arc configuration	Accuracy heat map (GIS raster) showing expected position uncertainty across the service area

Sources

Who does the work

We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.