NASHUA, N.H. — A drone does not have to be visible for a military unit to know it is nearby. If the aircraft or its operator is transmitting radio-frequency signals, equipment on the ground may be able to detect those signals and provide an early warning.

Figuring out exactly what produced the signal is harder. Counter-UAS systems can compare detected RF activity with libraries containing information about signals associated with known drones and controllers. That works particularly well when a system encounters something it already recognizes. Modified aircraft and changing frequencies can make the comparison less straightforward.

As drones and their communications continue to change, engineers are looking beyond simply matching signals with known entries. The larger challenge is understanding what is happening across the electromagnetic spectrum, even when a signal has not been encountered before.

Related: What is an uncrewed aircraft and how does it actually fly without a pilot onboard?

RF libraries give systems something to compare

An RF signature library gives a sensor a reference for identifying signals it encounters in the field. A match can help identify a known drone or another emitter. The limitation is that somebody generally has to encounter and characterize a new signal before it can become part of the library.

“A library-only system is structurally reactive: it can only recognise what someone has already captured, analysed and pushed back out to the field,” Bill Ostrowski of MyDefence told Military + Aerospace Electronics.

Ostrowski explained that libraries built through close cooperation with front-line units can shorten that process by incorporating information collected during real operations. But even a frequently updated library remains dependent on what has already been observed.

A newly modified drone, for example, may communicate differently from the version already represented in the library. Home-built aircraft and proprietary communications protocols can introduce signals that a system has not previously catalogued.

“The library gives precision against the known threat, the AI covers what has not been catalogued yet, and because the AI runs on the device itself the decision is made where the threat is rather than at the end of an update cycle,” Ostrowski said.

“Libraries do not disappear,” he continued. “They stop being the whole system and become the foundation that AI builds on.”

Detection and intelligence answer different questions

At its simplest, RF detection provides an alert when a sensor observes activity. “Detection tells you that something is there,” Ostrowski said. “It is close to binary, and what it buys you is time.”

RF intelligence goes further by characterizing that activity. Depending on the signal and equipment involved, a system may be able to provide information about what it is doing or where the emitter may be located.

“RF intelligence tells you what it is, what it is doing, info on position, and whether it is the same you saw yesterday,” he said. “That is what turns an alert into a decision, because you cannot choose a proportionate response to a signal you cannot characterise.”

Knowing that a sensor has detected RF activity is different from knowing that the activity appears to be associated with an uncrewed aircraft and understanding where it is coming from. Spectrum awareness widens the view again.

“Spectrum awareness is the wider electromagnetic picture, including your own emissions, jamming and interference,” Ostrowski said. “It is the context that explains why your sensors and your radios are behaving the way they are, and it increasingly shapes how a unit manoeuvres.”

That broader view can help explain what a sensor is detecting and what may be affecting the unit's communications and other electronic systems.

Passive RF listens without transmitting

RF sensing also differs from some other methods of detecting aircraft because it can operate passively. A passive RF sensor listens for energy already being transmitted and does not need to send out a signal simply to search for a drone. An active radar, by comparison, transmits RF energy and listens for the return.

“Because it does not transmit, it adds nothing to your own signature, which matters more as adversaries hunt emissions,” Ostrowski said.

There is a tradeoff, however. The target has to produce a useful signal for the sensor to hear.

Some drones make that increasingly difficult. A fiber-optic-controlled aircraft can receive commands through a physical fiber rather than a conventional RF control link. An aircraft capable of carrying out more of its mission autonomously may also have less need for continuous communication with an operator. Passive RF therefore cannot provide every piece of the counter-UAS picture.

“It is not always sufficient on its own,” Ostrowski said. “Fibre-optic controlled and fully autonomous platforms present little or no control link, which is exactly why passive RF belongs inside a layered architecture.”

Different sensors fill different gaps

Other sensors can help fill those gaps. Radar can detect an object by transmitting energy and measuring what returns. Electro-optical and infrared equipment can provide imagery, while acoustic sensors can listen for sounds associated with an aircraft.

RF adds another source of information. When a drone is transmitting, a passive sensor may provide an early indication of activity and potentially help characterize what is producing it. No individual sensing method works equally well against every aircraft or in every environment. Combining them allows one technology to contribute information when another has less to work with.

The spectrum contains much more than drones

The challenge grows as friendly and adversary drones become more common above military units.

“At that point, an alert saying that something is in the air is close to useless,” Ostrowski said. “The system has to answer whose it is, fast enough to act on. That is a much harder problem than detection, and it is where the real separation will happen.”

RF sensors also encounter far more than drone control links. Jammers, communications equipment and other electronic systems all contribute to the spectrum around a unit.

“The bigger shift is that these systems stop being just drone detectors,” Ostrowski said. “The drone is one emitter in a spectrum full of emitters.”

That gives the same sensing hardware potential value beyond counter-UAS. Instead of looking only for a particular drone signature, the system can contribute information about the electromagnetic environment surrounding the force.

The engineering problem is no longer limited to recognizing a drone that somebody has seen before. It is increasingly about making sense of a spectrum that keeps changing.