Report 043 · Defense Tech
The counter-drone conversation is almost always about what you shoot the thing down with. The quieter question is how you find it in the first place, and most of the answers on the market share one weakness: they announce themselves. A Counter-IED officer on why the most survivable drone sensor in a jammed spectrum is the one that makes no noise of its own.
Start with a distinction that gets lost the moment a counter-drone brochure lands on the table. There are two jobs, not one. First you have to detect the drone. Then you have to defeat it. Almost all the marketing energy, and almost all the money, goes to the second job: the jammer, the net, the interceptor, the laser. The first job, sensing, is treated as solved. It is not solved. It is the part that quietly decides whether any of the rest works.
And the standard way to sense a small drone is to look for it with radio. Radar paints it with energy and reads the echo. Radio-frequency detectors listen for the control link and video downlink between the drone and its pilot. Both are effective, both are mature, and both have the same structural problem, which is the one my old job trained me to see first.
I was a Counter-IED and electronic warfare officer. In that world you learn one rule before any other: in a contested spectrum, anything that transmits is a target. A radar is a transmitter. It has to be. To see the drone it floods the sky with energy and waits for a little of that energy to come back, which means an adversary with a cheap receiver can detect the radar from much farther away than the radar can detect anything, and can often work out where it is. The jammer you deploy to protect a site becomes the brightest thing on the enemy's map. I wrote earlier in this publication about the friendly-fire cost of running your own transmitters, and about how you locate a jammer by the very signal it puts into the air. Both cut the same way here. To emit is to be found.
A passive radio-frequency detector is quieter than a radar, since it only listens for the drone's link rather than illuminating anything. But it still depends on the drone talking. Modern drones increasingly do not. A fiber-optic-tethered drone has no radio link to intercept. An autonomous or pre-programmed drone can fly its mission in radio silence. And in the environment where you most need to catch a drone, the spectrum is already being hammered by jamming from both sides, which is exactly the noise a radio sensor has to hear through. The sensor that depends on the radio band is degraded by the same electronic warfare that is going on all around it.
So the honest question is not which radio sensor is best. It is whether there is a way to detect a drone that does not touch the radio spectrum at all.
There is, and it is old physics used in a new way. A drone is loud. Its rotors and motors produce a distinctive acoustic signature, a set of tones and harmonics tied to blade count and spin rate that does not look like a bird, a car, or the wind. If you can recognize that signature, you can detect the drone by sound alone. A microphone emits nothing. It sits there and listens. There is no beam to intercept, no emission to jam, no transmitter to geolocate. In electronic-warfare terms it has essentially no signature of its own, because sound is not in the spectrum the two sides are fighting over.
That idea has moved from concept to measured result. In a peer-reviewed study published in Acta Acustica in March 2026, researcher Zakaria Ghouli built a passive system from an array of low-cost MEMS microphones, the same class of tiny microphone in a phone, paired with a machine-learning classifier trained on drone acoustic signatures. On the classification task, the system reported 91.6% detection accuracy, 94.2% precision, and 88.7% recall, and it returned a result within 150 milliseconds of hearing the sound. It did more than detect. Using the small differences in arrival time across the microphone array, it localized the source to a median error of about 1.2 meters, with a mean bearing error of 6.3 degrees. All of that from listening, with nothing radiated outward.
Here is where I have to be the person who reads the limitations section, because the limitations are the whole reason this is a layer and not a silver bullet. The same paper is refreshingly honest about them. Detection was successful out to only about 50 to 60 meters in the outdoor tests. That is close. Strong coastal wind masked the drone's low-frequency tones and cut sensitivity by an estimated 10 to 15 percent. A drone flying directly overhead, the geometry you often care about most, degraded the localization by a further 20 to 25 percent, and the false-positive rate in the affected recordings ran 5 to 8 percent.
None of that is a scandal. It is what acoustic detection is: short-ranged, weather-sensitive, and best at exactly the small, low, slow drones that radar struggles with most. Which is why the sober way to read this is not "microphones replace radar." It is "microphones cover radar's blind spot, and they keep working when the radio sensors are being jammed." A sensor with a 60-meter range and no emissions is not competing with a radar that reaches kilometers. It is the last layer that still functions when the outer layers are blinded, and it is the one that never tells the enemy where your position is.
This is not a lab curiosity looking for a customer. On January 14, 2026, the U.S. Army's C5ISR Center issued a request for information for exactly this capability: acoustic detection systems for dismounted soldiers to catch Group 1 and Group 2 drones, the small ones under roughly 55 pounds. The requirement is specific about the mechanism. In the Army's words, "the desired system shall passively detect UAS signatures through acoustics to determine location, bearing and classification," working on the move and feeding the soldier's location display through the Tactical Assault Kit and Nett Warrior. The rationale the Army gives is the same one this whole piece rests on: small drones are hard to detect because of their size and how low they fly, and passive acoustic detection offers a way to sense them without active emissions.
Read that as an institution converging on the physics. When the people who have the most radar in the world start asking for microphones, it is not because the radar stopped working. It is because they have absorbed, the expensive way, that a sensor which broadcasts its own location is a sensor the enemy can find and suppress, and that the fight is moving toward drones that give a radio sensor nothing to hear.
Every sensor answers a question by some means, and the means has a cost. A radar answers "is something there" by shouting into the dark and listening for an echo, and the cost is that everyone in the dark now knows where you are standing. A passive acoustic array answers a narrower question, "is a drone close," by saying nothing at all, and the cost is range and a sensitivity to weather. Neither is better in the abstract. They are answers to different questions with different bills attached.
So when you evaluate a counter-drone system, do not stop at how it defeats the drone. Ask how it finds the drone, and then ask what that finding costs you. If the answer is a powerful emitter, understand that you have bought a capability and a beacon in the same box. In a spectrum that is going to be jammed, contested, and mapped by both sides, the sensor that quietly listens may end up being the one still working when the loud ones have gone dark, or been silenced. The gear is new. The advantage of making no noise is as old as the ambush.
U.S. Army combat veteran (Counter-IED / Electronic Warfare), peer-reviewed researcher in microwave spectroscopy, and founder & CEO of Shroombiosis. Consults on laboratory operations, AI, and supplement formulation.