← All articles

The Drone With No Signal

Fibre optic FPV drones emit zero RF. That breaks the fundamental assumption most counter-drone systems are built on.


The assumption everyone was comfortable with

For the last couple of years, the counter-drone world has operated on a pretty tidy premise: drones talk to their operators by radio, radio waves are detectable, therefore drones are detectable. RF sensing, passive direction-finding, spectrum analysis, protocol fingerprinting, all of it flows from that one assumption. Some systems treat RF as their primary detection method. Some treat it as their only one.

You can probably see where this is going.

Fibre optic FPV drones, now in active use in Ukraine, don't use radio at all. They communicate entirely through light pulses in glass fibre. They emit no RF. None. For any system built around intercepting radio signals, these drones simply do not exist.

How it works (and why it's annoyingly simple)

Take a standard first-person-view drone. The kind that both sides in Ukraine have been using by the thousand for reconnaissance and strikes. Now bolt on a lightweight spool of single-mode optical fibre, wound on a bobbin.

As the drone flies, the fibre pays out behind it. It's absurdly thin, about 250 microns with coating, barely thicker than a hair. A few kilometres of the stuff weighs almost nothing.

The camera feed travels back through the fibre as modulated light. Control inputs go forward the same way. The entire data link is photonic. No RF transmitter on the drone, no antenna, no electromagnetic emissions of any kind. As far as the spectrum is concerned, the drone might as well not be there.

And the fibre? Disposable. Fractions of a penny per metre. You fly the mission and leave it behind like a spider's thread. Nobody's going back to collect it.

What happened when these turned up

Fibre optic FPV drones started appearing in Ukrainian combat footage in late 2023. By mid-2024, they were being produced in quantity by multiple units. Not a novelty any more. Just another tool.

The operational reports make for uncomfortable reading if you sell RF detection equipment. Units with functioning RF-based drone detection described engagements where the system was on, working correctly, and detected nothing at all. The first indication of an incoming drone was either a soldier physically spotting it or, in the worst cases, the detonation.

This broke something quite fundamental about how defensive positions had been set up. If your RF system typically gives you 30 seconds of warning, you build your entire response timeline around those 30 seconds. When the warning drops to zero, the response chain never even starts.

Jamming, which had been the primary active countermeasure against FPV drones, became irrelevant too. You can't jam a signal that isn't propagating through the air. Direction-finding equipment, which locates drones by triangulating their RF emissions, produced no bearing. On the electronic warfare picture, the drone just wasn't there.

This isn't a sensitivity problem

Here's the bit that really matters. RF-based systems facing fibre optic drones don't need better antennas, more sensitive receivers, or updated threat libraries. The problem isn't that the signal is faint. The problem is that the signal doesn't exist.

This makes fibre optic drones categorically different from, say, a drone using clever frequency hopping to avoid detection. That drone is still emitting. Better kit might find it. A fibre optic drone has moved the entire communications link out of the electromagnetic domain. Within the RF modality, detection isn't difficult. It's impossible.

For any counter-drone system that leans heavily on RF, this creates a rather binary outcome. Either you detect the drone through other means, or you don't detect it.

What still works

The fibre optic link removes one detection surface: the RF communication channel. It does not make the drone physically invisible. Several modalities carry on working perfectly well.

Acoustic detection

Multirotor drones are noisy. Motors whine, propellers chop air. These sounds are completely independent of how the drone communicates. A microphone array that can detect a standard FPV drone will perform identically against a fibre optic one, because it sounds exactly the same. In quiet environments, which is precisely where forward positions tend to be, acoustic detection can pick up drones at useful ranges.

Radar

Radar bounces off the physical airframe. The drone's body, motors, and spinning propellers all reflect radar energy. Radar genuinely could not care less how the drone is talking to its operator. Micro-Doppler signatures from spinning blades are a particularly neat discriminator. Works just as well against fibre optic drones as any other.

Thermal and visual

Cameras, both visible-spectrum and infrared, detect the drone's physical and thermal presence. Motors get hot. Batteries get hot. The airframe is visible against the sky. Thermal contrast is especially useful at night, which is exactly when a lot of fibre optic drone operations happen. Again, completely indifferent to the communications method.

Multi-sensor fusion

No single alternative modality is the answer on its own. The answer is combining them. Acoustic might give you the initial cue. Radar confirms the track and gives you range and velocity. Electro-optical classifies the target. Each covers the others' blind spots. A fusion system that combines three or more independent sensing methods maintains detection capability regardless of whether the target emits RF.

The bit nobody wants to think about yet

Fibre optic drones aren't the last thing that will break the RF assumption. Autonomous drones that need no communication link at all, pre-programmed with GPS waypoints or using onboard visual navigation, present exactly the same RF-invisible problem. As autonomy improves, the proportion of drone threats that emit any detectable RF signature is only going to shrink.

The systems that will still be relevant in two years are the ones that were designed from the start around multi-modal sensing. RF as one input among several, not the backbone.

Any architecture that can't detect a drone it can't hear on the spectrum has a shelf life. And for the threats already flying in Ukraine, that shelf life has already expired.