Insights · Passive Sensing & Detection

Detecting the Drone That Doesn't Transmit

September 2026 · 4 min read

AMBy Amardeep Mond, Founder & Chief Technology Officer, Aerodyne AI

A growing share of hostile unmanned aircraft never transmit. Fiber-optic FPV drones carry their control link in a spool of glass; long-range one-way attack drones fly pre-programmed routes with no operator link at all. Detection architectures built around radio-frequency scanning are structurally blind to both. The physics of what such aircraft cannot help emitting — sound, radar reflection, heat — points to a different architecture.

The RF-scanning assumption, and how it broke

The first generation of counter-UAS sensing was built on a sound observation: consumer and commercial drones chatter. Command links, telemetry, and video downlinks occupy predictable bands, and a passive receiver network can detect, classify, and often localize a drone — and sometimes its operator — from those emissions alone. The approach is inexpensive, licensable almost anywhere, and genuinely effective against the aircraft it was designed for.

The battlefield then optimized against it. Fiber-optic FPV variants, now produced at scale, replace the radio link with an unspooling optical fiber: full video and control with zero radiated signature, immune to both detection and jamming. At the other end of the size spectrum, one-way attack aircraft of the Shahed type navigate by inertial and satellite guidance along pre-programmed routes; there is no control link to intercept for most of the flight. Publicly reported employment of both classes has grown continuously since 2022. An emissions-only sensor network registers neither until the terminal moment — if at all.

What a drone cannot stop emitting

A powered aircraft is a machine that must move air, reflect electromagnetic energy, and dissipate heat. Each is a detection channel with a mature open literature.

Acoustics

Propulsion is loud in structured ways. A piston engine of the type used by long-range one-way attack aircraft produces strong tonal content at its cylinder firing rate and harmonics; electric multirotors produce blade-passage tones set by rotor RPM and blade count, with a characteristic harmonic comb that survives in the spectrum even when broadband noise is high. The academic literature on acoustic detection and localization of small UAS — using microphone arrays, time-difference-of-arrival processing, and harmonic-line tracking — goes back well over a decade, with published work from university groups demonstrating detection and bearing estimation of small aircraft at operationally interesting ranges under favorable conditions [1]. The physics is honest about its limits: detection range varies strongly with wind, terrain, ambient noise, and the aircraft's propulsion class, which is exactly why acoustics belongs in a layered mix rather than alone.

Passive radar

Radar detection does not require the defender to transmit. A bistatic passive radar uses existing illuminators of opportunity — FM broadcast, digital television, cellular downlinks — as its transmitter, and correlates the direct signal against reflections from moving targets. The foundational engineering literature is well established [2], and published experimental systems have demonstrated detection of aircraft using FM broadcast illumination since the mid-2000s [3]. For counter-UAS work the attraction is twofold: the sensor emits nothing, so it cannot be targeted by anti-radiation weapons or located by electronic reconnaissance; and the widespread, low-frequency illumination is poorly suited to the radar cross-section reduction techniques optimized against conventional microwave radar. The engineering challenges — direct-path interference cancellation, low signal-to-noise ratios, bistatic geometry effects — are equally well documented, and are the reason passive radar performance is highly site- and geometry-dependent.

Thermal and electro-optical

Combustion engines and motors dissipate heat; airframes present visual signatures. EO/IR sensors offer high-confidence classification and precise angular measurement, but over narrow fields of view — which makes them best employed as confirmation sensors, cued by wider-area acoustic or radar detections rather than searching open sky unaided.

The architectural conclusion: no single channel, and no single site

Each channel fails differently. Acoustics degrades in wind; passive radar depends on illuminator geometry; EO/IR is weather-limited and narrow. A defensible architecture therefore fuses multiple channels — and, just as importantly, multiple sites. A single sensor position, however multi-modal, measures one bearing set and can be sited around, saturated, or destroyed. A network of distributed nodes converts single-site bearings into multilateration, extends coverage from a point to an area, and degrades gracefully under node loss. The fusion problem this creates — combining detections from heterogeneous sensors across a bandwidth-constrained network — is a substantial engineering discipline in its own right, and the subject of a separate article in this series.

The strategic point stands independent of any particular implementation: as the threat has evolved toward aircraft that emit nothing, credible detection has necessarily become passive-first, multi-modal, and distributed. That is not a product claim; it is where the physics leads.

References

1. See, e.g., published work on acoustic UAV detection and localization with microphone arrays from academic groups including the Stevens Institute of Technology and others, in IEEE conference proceedings on homeland security technologies and acoustics.
2. H. D. Griffiths and C. J. Baker, An Introduction to Passive Radar, Artech House, 2017.
3. P. E. Howland, D. Maksimiuk, and G. Reitsma, "FM radio based bistatic radar," IEE Proceedings — Radar, Sonar and Navigation, vol. 152, no. 3, 2005.

This article synthesizes publicly available academic literature and public reporting. It describes concepts and results from the open published record and contains no Aerodyne technical data, system specifications, or performance information.