Sensor Fusion: Detecting RF-Silent Fiber Optic Drones on the Battlefield
By Benjamin Cook
Expect every truck, tank, and vehicle to look like the mast of a naval ship in the near future.
NOTE: Below is a practical analysis about how to defend against fiber optic drones. I won’t bury the lede. Defense against any and all small-cheap-drone is based on first seeing/sensing it. It’s all about the sensors. There is no new tech here. We have everything we need to sense a drone in enough time to shoot it down. As with everything, is it scalable? Yes. I think so. Is there something on the market now? No, but many systems are being tested. There are many systems for shooting down cruise missiles and Shaheds. But nothing yet that is modular for good point and convoy defense vs small-cheap-drones. Why don't we already have this on the battlefield? EW has, until now, worked very well. Fiber drones changed this.
On today’s battlefields, fiber-optic guided drones present one of the most difficult detection challenges faced by defenders. These drones don’t emit radio-frequency signals, making them invisible to standard electronic warfare systems that rely on RF detection and jamming. Controlled via an ultra-thin tethered cable, they fly low, use terrain for cover, and exploit gaps in conventional sensor systems. Both Ukraine and Russia have deployed these “unjammable” drones for reconnaissance and precision strikes.
Detecting fiber-optic drones requires a layered approach. No single sensor can cover all the weaknesses these drones exploit. The solution lies in combining acoustic, radar, and optical sensors into a single fusion system. Each sensor type has its strengths and weaknesses, but together they can compensate for each other and create a viable detection umbrella.
Hate Subscriptions? Me too! You can support me with a one time contribution to my research at Buy Me a Coffee. https://buymeacoffee.com/researchukraine
Acoustic sensors pick up the noise of drone motors and propellers. Fiber-optic drones often require slightly more thrust to carry the weight of their tether cable, making them somewhat noisier than standard drones. Microphone arrays, sometimes with dozens of microphones, can listen for these sounds even when the drone is hidden by terrain or flying behind obstacles. These arrays use beamforming to calculate the drone’s bearing and sometimes elevation. However, their range is limited, typically only a few hundred meters. Background noise from battlefields, wind, vehicles, or even distant helicopters can overwhelm the drone’s acoustic signature.
Radar systems offer the longest detection ranges, sometimes exceeding 1 to 3 kilometers depending on the radar’s power and the size of the drone. Unlike acoustic systems, radar works in all weather and all lighting conditions. Advanced radars use micro-Doppler processing to identify drones by the unique spinning signatures of their propellers, distinguishing them from birds or debris. But radar requires direct line-of-sight, and drones flying close to terrain features can mask themselves in ground clutter, causing radars to miss them until very late in their approach.
Optical and infrared cameras provide the clearest confirmation and identification. A camera can visually confirm whether an object is a drone and even identify payloads. Thermal imaging can detect heat signatures from motors and batteries, especially at night. One novel technique under development involves using infrared laser illuminators to sweep the sky. Fiber-optic cables can reflect this IR light, making the otherwise invisible tether visible to sensors. However, optical systems suffer in fog, rain, and smoke, and often have a limited field of view. They are typically cued by radar or acoustic sensors to focus on a suspected location rather than scanning blindly.
The key to effective fiber-optic drone detection is fusion—linking these three sensor types into a single integrated system. When radar detects a possible drone, it can automatically cue cameras to zoom in for identification. If acoustic arrays pick up a sound from a hidden drone, they can cue radar to scan that sector more intensely. A fusion engine correlates data from all sensors in real time, filtering out false alarms and producing a unified air picture. If radar detects something and acoustic sensors also register drone-like noise, confidence rises that it's a valid target. When combined with machine learning algorithms, this multi-modal approach becomes increasingly reliable.
Two practical systems have emerged from recent battlefield experience and development work.
For approximately $500,000, a truck-mounted system can be deployed to protect forward operating bases, convoys, or border sectors. These mobile units typically include a 3D radar mounted on an elevating mast for maximum coverage, such as the Spotter Global AX250 or Blighter A422. The radar feeds continuous track data to a ruggedized command and control (C2) system in the vehicle. High-end EO/IR gimbals are mounted on stabilized masts, automatically slewing to visually confirm targets identified by radar. These cameras often include cooled mid-wave infrared sensors with detection ranges up to several kilometers. Acoustic arrays, such as the Squarehead Discovair G2, provide additional close-in coverage, filling blind spots where terrain may block radar signals. All sensor feeds merge inside the truck’s C2 system, which displays radar tracks, video feeds, and acoustic bearings in a unified interface. This allows operators to make rapid decisions and cue possible countermeasures. Similar systems already exist in active service, such as the U.S. Army’s X-MADIS or Teledyne FLIR’s LVSS ADA platforms.
For about $100,000, a portable ground-based system can be built to defend trenches, forward outposts, or sensitive fixed sites. In this configuration, radar may be scaled down to micro-radar units capable of detecting drones out to 500 meters. These smaller systems might borrow from automotive radar designs or perimeter security radars. Acoustic coverage can be provided by several compact microphone arrays placed around the defensive position, offering 200–300 meters of detection range. Optical confirmation might rely on fixed thermal cameras or unfocused IR laser illuminators, reflecting off fiber-optic cables to expose nearby drones. Processing power is provided by a ruggedized laptop or embedded mini-PC running open-source fusion software. The NATO Innovation Challenge has called for exactly these types of modular, under-100kg systems for frontline use, as they allow small teams to rapidly deploy effective drone detection without the size, cost, or power demands of full mobile platforms.
Both systems share modular architecture. Many of the software algorithms, sensors, and processing frameworks can be used across both truck-based and portable setups, allowing flexible adaptation depending on mission needs and available budgets.
Real battlefield experience has validated the value of sensor fusion. In Ukraine, fiber-optic drones have repeatedly exploited gaps in RF detection. Early attempts at single-sensor solutions produced high false alarm rates. Birds triggered radar-only systems. Acoustic-only setups struggled in noisy conditions. But combining multiple sensors allowed defenders to filter out false targets, verify real threats, and provide soldiers with actionable early warnings. Even when radar momentarily loses a low-flying drone masked by terrain, acoustic arrays can provide seconds of advanced warning—critical time for defenders to react.
As fiber-optic drones proliferate and evolve, defenders must continue to develop adaptable, modular sensor fusion platforms that can be deployed across both high-end military and more austere frontline environments. Which is a good way of saying the systems must be scalable/affordable and possibly even attritable.
Benjamin Cook continues to travel to, often lives in, and works in Ukraine, a connection spanning more than 14 years. He holds an MA in International Security and Conflict Studies from Dublin City University and has consulted with journalists and intelligence professionals on AI in drones, U.S. military technology, and open-source intelligence (OSINT) related to the war in Ukraine. He is co-founder of the nonprofit UAO, working in southern Ukraine. You can find Mr. Cook between Odesa, Ukraine; Charleston, South Carolina; and Tucson, Arizona.
Hate Subscriptions? Me too! You can support me with a one time contribution to my research at Buy Me a Coffee. https://buymeacoffee.com/researchukraine
Mr. Cook’s Substack:
Sources
● Robin Radar
● Spotter Global
● Business Insider
● Photonics Online
● Hensoldt C-UAS
● FLIR Government & Defense
● DroneShield
● NATO C-UAS Innovation Challenge 2025
● SUAS News
● C-UAS Hub
Thanks Ben. I would like to know how the parties identify the affiliation of the many drones in the air in the frontline zone, how the fighters/operators determine which drones are enemy and which are "friendly", which ones need to be shot down and which ones can be ignored.
I acknowledge that the following information is a bit off-topic. Nevertheless, it's worth considering this topic as well. Who knows, it might not be long before it becomes part of our everyday lives.
EMP – How dangerous are electromagnetic weapons?
Electromagnetic weapons are considered the most dangerous weapons technology that could cripple countries, cities, power grids or entire regions to gain a strategic advantage during an attack. To date, there exist no such weapons officially in existence that work with an EMP (electromagnetic pulse). In secret, the military worldwide has been researching on electromagnetic weapons for decades that could eventually be used.
This technology would render all flying, unprotected drones completely unusable within the radius of explosion, whether or not they are connected to fiber optic cables.