ELINT and COMINT Fusion: Combining Electronic and Communications Intelligence

Signals intelligence has two primary collection disciplines, each capturing a different layer of the electromagnetic environment. Electronic Intelligence (ELINT) focuses on non-communications emissions — radar signals, weapon system guidance waveforms, electronic warfare system signatures. Communications Intelligence (COMINT) intercepts intentional communications — voice transmissions, encrypted data links, messaging systems. In isolation, each discipline produces partial visibility. A radar emission tells you that a system is radiating; it does not tell you what orders the operator received. A command radio transmission tells you what was ordered; it does not tell you which weapon system executed the order. Fusing ELINT and COMINT produces the coherent picture that commanders require: what systems are operating, what those systems have been ordered to do, and when capability and intent align into an imminent threat.

ELINT Collection and Parametric Analysis

ELINT collection measures the physical characteristics of radar and other non-communications emissions. The primary parameters captured for each emission are: carrier frequency (the center frequency of the transmission), pulse repetition interval (the time between successive radar pulses), pulse width (the duration of each pulse), scan rate and pattern (how the antenna rotates or phases), and modulation characteristics (AM, FM, pulse Doppler waveforms). These parameters, taken together, constitute the emitter's electronic fingerprint — a set of characteristics that, compared against a parametric emitter library, identifies the system type with high confidence.

Emitter Identification (EID) is the process of matching observed signal parameters against library entries. The library contains known parameter ranges for each radar type — the AN/APG-77's expected pulse repetition frequency range, the S-300's surveillance radar scan rate, the Pantsir fire control radar's frequency agility pattern. When ELINT collection produces a new emitter observation, EID compares the observed parameters against library entries and returns a ranked list of candidate identifications with confidence scores. EID works even when the emitter uses frequency agility (pseudo-random frequency hopping) if the agility pattern itself is characteristic of the system type.

The operational output of ELINT collection is an emitter record: a geographic position (derived by geolocation techniques — discussed separately), a system type identification, an activity timeline (when the emitter was first detected, when it changed mode, when it ceased radiating), and a confidence score for the identification. Multiple emitter records from the same geographic area constitute an electronic order of battle — an inventory of the electromagnetic capabilities present in a defined area of operations.

COMINT Collection and Traffic Analysis

COMINT collection operates on intentional communications. For unencrypted voice communications, this includes both content analysis (what was said) and traffic analysis (who communicated with whom, when, and how long). For encrypted communications, content is unavailable but traffic analysis remains fully applicable: communication patterns, network topology, message frequency, and transmission timing all carry intelligence value independent of the content.

Traffic analysis techniques applicable to military COMINT include: network mapping (identifying which stations communicate with which others, revealing the command network structure — a high-traffic hub station is likely a headquarters or relay), activity analysis (changes in communication volume or frequency that correlate with operational activity — a fourfold increase in encrypted traffic may indicate an imminent operation), direction finding (locating communication transmitters using TDOA or AOA techniques), and link analysis (identifying communication patterns that associate specific call signs or frequencies with specific units or functions).

Call sign exploitation is a classic COMINT technique. Military radio operators use call signs that rotate on a schedule, but the rotation schedule itself may be predictable, or operators under stress may break discipline and use a previous call sign. Building a database of observed call signs, their associated frequencies, observed times, and geolocated positions enables pattern-of-life analysis even when content is encrypted or absent.

Fusion Architecture: Correlating ELINT and COMINT

The fusion process correlates ELINT emitter records with COMINT observation records to produce combined intelligence products. The correlation operates on several dimensions simultaneously:

Geographic correlation: An ELINT emitter geolocated to grid reference X, and a COMINT transmitter geolocated to the same grid reference, likely belong to the same physical platform or unit. The geographic correlation window must account for geolocation error ellipses from both collection systems — if the ELINT position has a 500-meter CEP and the COMINT position has a 300-meter CEP, the correlation threshold should be approximately 1200 meters rather than expecting exact coincidence.

Temporal correlation: An ELINT emitter that activates at T1, and COMINT traffic observed at the same frequency from the same approximate location at T1, likely represent the same activity. Temporal correlation is particularly powerful for detecting coordination events — if a radar activates and encrypted voice traffic increases at the same time from the same location, the system is likely transitioning to an active engagement posture.

Entity correlation: Emitter library entries for specific system types can be cross-referenced against COMINT databases to associate a weapon system type with a unit. If ELINT identifies an SA-21 acquisition radar at location X, and COMINT has previously associated a specific radio network with the 7th Air Defense Brigade at location X, the fusion product is: the 7th Air Defense Brigade has an operational SA-21 at this location.

Electronic Order of Battle Maintenance

The output of sustained ELINT/COMINT fusion is an Electronic Order of Battle (EOB) — a structured database of adversary electronic capabilities, their locations, their unit associations, their activity patterns, and their current status. An EOB entry for a fire control radar includes: radar type, emitter ID confidence, geographic location with error ellipse, associated unit, observed activity timeline, last observed status (radiating or silent), threat envelope parameters (from the emitter library), and any COMINT associations that have been confirmed.

EOB maintenance is a continuous process, not a one-time product. Emitters go silent (power-off, reposition, or emissions control), new emitters appear, units relocate. The EOB system must age entries when a previously active emitter has not been observed for a configurable threshold period, flag geographic inconsistencies when the same emitter ID is reported from locations too far apart to be the same platform, and merge duplicate entries when multiple collection assets have independently observed the same emitter.

Software Architecture for ELINT/COMINT Fusion Systems

A production ELINT/COMINT fusion system requires several distinct software components. The collection front-end processes raw signal measurements into standardized emitter observation records — normalizing parameter formats, applying geolocation calculations, and packaging records for downstream processing. The correlation engine applies the multi-dimensional correlation logic to match ELINT and COMINT records against each other and against the existing EOB. The EOB database stores the fused entity records with full history, supporting both current-state queries and historical replay for analytical tasks. The analyst workstation provides the human interface for reviewing correlation candidates, confirming or rejecting proposed fusions, and manually adding context that automated systems cannot infer.

The correlation engine is the most computationally intensive component. For a system processing hundreds of ELINT emitter observations and thousands of COMINT records per hour, the correlation engine must evaluate each new observation against all existing records within plausible geographic and temporal windows. Efficient spatial indexing (R-tree or PostGIS GiST) reduces the candidate set for each correlation from the entire database to the small number of records within the relevant geographic window.