C-UAS detection: Why RF and radar sensors fail against autonomous drones

Most fielded Counter-UAS (C-UAS) systems are built around a single assumption: the drone is controlled by a radio link. Jam that link, detect its emissions, or triangulate its operator, and the threat collapses. That assumption held for years, but it is no longer universal. A growing class of aerial threats is deliberately built to operate without one — and against them, RF and radar-first architectures leave a dangerous gap.

The RF layer is listening for a signal that isn't there

RF-based C-UAS sensors detect, classify, and geolocate drones by intercepting the command-and-control link between the aircraft and its ground station. They are effective against consumer quadcopters, modified FPV rigs, and many military systems that still rely on 2.4 GHz, 5.8 GHz, or proprietary datalinks.

The weakness is in the dependency. If the aircraft is not emitting, the RF layer has nothing to fingerprint. Two operational threat classes exploit this directly.

Fibre-optic-controlled drones

These aircraft trail a thin optical fibre from the airframe to the operator. The command channel is physical, not electromagnetic. There is no control signal to intercept, no uplink to jam, and no spectrum signature to classify. Widely used in current combat operations, they convert the drone into a line-of-sight weapon that RF sensors cannot detect by emissions alone.

Fully autonomous drones

Autonomous systems carry preloaded missions, onboard navigation, and terminal guidance. After launch, there is no continuous link. No operator to geolocate. No jamming surface. The aircraft navigates by GNSS, inertial reference, or vision alone. Against these targets, an RF-only kill chain never begins.

Radar is not always the answer

Radar solves part of the problem. It does not require the target to emit; it paints the aircraft directly. But radar-first C-UAS has its own limitations. Small drones have low radar cross-sections. Flying low and slow, they hide in ground clutter. In urban environments, multipath and shadowing make consistent tracking difficult. Classification — telling a drone from a bird, a balloon, or debris — is hard without additional sensor context.

Radar tells you that something is there. It does not necessarily tell you what it is, whether it is hostile, or where to point an effector. In layered defence, that ambiguity matters.

EO/IR: the necessary additional layer

Electro-optical (EO) and thermal (IR) sensors see the physical aircraft. They do not wait for emissions, they do not rely on radar reflectivity, and they are not confused by RF silence. A camera or thermal imager resolves the airframe against the sky or terrain background, producing the visual evidence that RF and radar cannot.

Combined with edge-AI classification and tracking, EO/IR becomes a detection modality that is indifferent to the drone's command method. Radio, fibre, or fully autonomous — the aircraft is still a physical object, and it can still be seen. This is why mature C-UAS architectures are moving toward layered sensor fusion: RF for emissions, radar for all-weather presence, and EO/IR for positive identification and tracking.

What this means for layered defence

A layered C-UAS system should not treat any single sensor as sufficient. RF is fast and scalable but blind to RF-silent threats. Radar extends coverage but struggles with classification and clutter. EO/IR closes the identification gap and provides the visual confirmation needed for engagement decisions.

For organisations assessing C-UAS procurement, the question is no longer "which sensor is best?" but "which combination covers the full threat set?" Any architecture that omits an EO/IR layer leaves an open door to the drones that are designed specifically to exploit that omission.