What the payload actually measures
An RF signal-mapping payload is, at its simplest, a wideband software-defined radio receiver with a precise onboard clock. It records the complex baseband of whatever signals fall within its tuned band during a pass. The geolocation comes not from a single satellite but from comparing those recordings across two or more spacecraft flying in coordinated formation.
Two measurements do the work. Time-difference-of-arrival (TDOA) exploits the fact that a ground emitter's signal reaches each satellite at a slightly different instant, because the propagation paths differ in length. Frequency-difference-of-arrival (FDOA) exploits the differential Doppler shift that arises because each satellite moves at a slightly different velocity relative to the emitter. Together, TDOA and FDOA intersect two hyperbolic surfaces in three-dimensional space; their crossing line, projected onto the Earth ellipsoid, gives a position fix. HawkEye 360's published cluster architecture uses three spacecraft separated by tens to low hundreds of kilometres in the same orbital plane to execute exactly this geometry.
Frequency coverage and the software-defined advantage
The defining characteristic of an SDR payload is that the frequency plan lives in firmware, not in fixed filter banks. HawkEye 360's Clusters 1 through 7 are publicly documented as covering VHF through Ku-band, roughly 144 MHz to 18 GHz, in selectable sub-bands. That range spans maritime VHF voice and AIS (156–162 MHz), aeronautical DME and ACAS transponders, L-band EPIRB distress beacons, S-band radars, C-band satellite uplinks and X/Ku-band communications terminals. A single hardware design, reprogrammed on orbit, can pivot between those targets as mission priorities shift.
Bandwidth per sub-band is a genuine constraint. Capturing 100 MHz of instantaneous bandwidth requires onboard ADC rates and storage that drive mass and power upward fast. Most operational smallsat SDR payloads process 20–40 MHz per channel in practice, which means a wideband survey requires multiple dwells or multiple receiver chains in parallel. Missions that need simultaneous multi-band coverage typically fly larger payloads or accept that some signals will be missed in any given pass.
Formation flying: the geometry you cannot skip
TDOA accuracy scales directly with baseline length and clock precision. Shorten the baseline between satellites and the TDOA differential shrinks toward the noise floor of the timing system. Lengthen it too far and the assumption that both satellites are illuminated by the same emitter at the same time breaks down, because the emitter may have stopped transmitting. Published analyses of HawkEye 360's architecture suggest baselines of roughly 100–300 km offer a practical balance for maritime and terrestrial emitter classes.
Maintaining that geometry requires active station-keeping or very careful differential drag management. Smallsats in low Earth orbit experience differential atmospheric drag that, left uncorrected, will pull a three-spacecraft cluster apart within weeks at 500 km altitude. Propulsion adds mass, cost and complexity; propellant-free drag compensation using differential attitude is possible but limits flexibility. Formation acquisition after launch is also non-trivial: the cluster must be deployed into a configuration that converges on the target geometry, not just released simultaneously from a single dispenser.
Clock synchronisation is equally critical. TDOA resolution of 100 nanoseconds corresponds to roughly 30 metres of range difference. Achieving sub-100 ns inter-satellite timing alignment either requires tight onboard oscillator discipline, inter-satellite ranging links, or post-processing cross-correlation against a common reference signal. Most operational systems use a combination of GPS-disciplined clocks and ground-based cross-correlation of known beacon signals to calibrate residual timing errors.
Geolocation accuracy: honest numbers
HawkEye 360 publishes geolocation accuracy figures in the range of 1–5 km CEP (circular error probable) for their operational service, depending on signal type, dwell time, geometry and emitter behaviour. That is sufficient to attribute an emission to a vessel, a specific cell of an airport, or a border region, but it will not resolve two ships in the same anchorage or two terminals on the same factory roof.
Accuracy degrades with weak signals, short dwell times, and poor satellite geometry. An emitter that transmits for only a few seconds may not produce enough signal energy for reliable cross-correlation. Emitters that hop frequency or use spread-spectrum waveforms require more sophisticated signal processing and longer integration. The FDOA component depends on knowing each satellite's velocity to better than a few centimetres per second, which in turn demands precise orbit determination, typically from onboard GPS receivers cross-checked against ground tracking.
Where the method falls short
RF signal mapping cannot geolocate what is not transmitting. A vessel that has switched off its VHF radio and its AIS transponder, and is not actively using a satellite phone or radar, is invisible to this payload class. It remains visible to synthetic aperture radar and, in daylight, to optical sensors. The methods are complementary, not interchangeable.
Dense signal environments create ambiguity. In a busy port or a congested spectrum band, multiple emitters transmit simultaneously on overlapping frequencies. Separating their individual TDOA signatures requires signal separation algorithms that work well for narrowband, isolated emitters but become unreliable as spectral density rises. Urban RF environments are largely intractable for this technique without very high signal-to-interference ratios.
Regulatory access to spectrum is also a real constraint. Receiving any frequency from orbit is generally permissible; retransmitting, processing for intelligence purposes, or sharing intercept data may trigger national and international legal obligations depending on the signal type and the jurisdiction of both the operator and the target. Governments procuring this capability need legal review alongside the engineering specification. ITU Radio Regulations and national spectrum management frameworks govern what can be collected and how the data can be used.
Sizing a sovereign RF-mapping mission
A minimum viable TDOA/FDOA capability requires at least two satellites; three is the practical floor for reliable two-dimensional geolocation without additional geometric assumptions. Each spacecraft in the HawkEye 360 class sits in the 15–30 kg range with payload power draws of 20–50 W per receiver chain. That puts the full cluster within the capacity of a single rideshare launch to a sun-synchronous orbit between 500 and 575 km, which is the sweet spot balancing revisit rate, atmospheric drag lifetime and signal collection geometry.
Revisit rate for a three-satellite cluster in LEO is typically two to six passes per day over a mid-latitude target, depending on orbital inclination and the cluster's phasing. That is adequate for maritime domain awareness over open ocean but insufficient for near-real-time monitoring of a specific port or land border. Constellations of six or more clusters, as HawkEye 360 has built, compress revisit to sub-hourly for most latitudes. A sovereign programme with more modest ambitions might instead rely on a small national cluster for strategic awareness and access commercial tasking for surge requirements.
Engineering parameters
| Spacecraft mass per node (HawkEye 360 class) | 15–30 kg |
| Payload power per receiver chain | 20–50 W |
| Typical frequency coverage | VHF to Ku-band (approx. 144 MHz – 18 GHz), selectable sub-bands |
| Instantaneous bandwidth per channel | 20–40 MHz (typical operational SDR smallsat) |
| Geolocation accuracy (published, HawkEye 360) | 1–5 km CEP depending on signal type and geometry |
| Inter-satellite baseline for TDOA/FDOA | 100–300 km (practical range for maritime/terrestrial emitters) |
| Minimum cluster size for 2-D geolocation | 3 satellites |
| Typical operational orbit | 500–575 km SSO |
| Revisit rate (3-satellite cluster, mid-latitude) | 2–6 passes per day |
| Clock synchronisation requirement | Sub-100 ns inter-satellite timing alignment for 30 m range resolution |
One contract, one accountable engineer
Commissioned as one programme, not a stack of contracts: spacecraft, launch, ground segment, mission control, training and handover are priced together. Source-access terms and audit rights are agreed in writing before signature. Discuss cluster architecture for your orbit.