RF Signal Processing for Drone Detection

𝗙𝗶𝗯𝗲𝗿-𝗼𝗽𝘁𝗶𝗰 𝗙𝗣𝗩 𝗱𝗿𝗼𝗻𝗲𝘀 𝗮𝗿𝗲 𝗻𝗼𝘁 𝗶𝗻𝘃𝗶𝘀𝗶𝗯𝗹𝗲 There is a persistent misconception in the drone debate. Because fiber-optic FPV drones emit 𝗻𝗼 𝗿𝗮𝗱𝗶𝗼 signal, many assume they are impossible to detect. That is not true. 📡 Fiber-optic drones may be invisible to classical electronic intelligence (#SIGINT / #ESM), but they remain 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 objects moving through space. And physics still betrays them. One of the most promising detection methods relies on 𝗽𝗮𝘀𝘀𝗶𝘃𝗲 𝗿𝗮𝗱𝗮𝗿. Instead of transmitting its own signal, the system uses existing emissions in the environment — DVB-T towers, GSM networks, or LTE infrastructure — as an illumination source. An SDR-based receiver then analyzes how these signals are reflected by objects in the air. 🔎 The real discriminator comes from 𝗺𝗶𝗰𝗿𝗼-𝗗𝗼𝗽𝗽𝗹𝗲𝗿 signatures. Traditional radar often struggles to distinguish small UAVs from birds because their radar cross-sections can look similar. Micro-Doppler changes that. By analyzing the tiny motion patterns inside the reflected signal, the system can detect the unique spectral fingerprints created by rapidly rotating rotor blades. ⚙️ What this enables: • 𝗣𝗿𝗲𝗰𝗶𝘀𝗲 𝘁𝗮𝗿𝗴𝗲𝘁 𝗱𝗶𝘀𝗰𝗿𝗶𝗺𝗶𝗻𝗮𝘁𝗶𝗼𝗻   Rotor rotation generates a distinctive high-frequency Doppler pattern that differs fundamentally from bird wing flaps or vegetation movement. • 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗼𝗳 𝗵𝗼𝘃𝗲𝗿𝗶𝗻𝗴 𝗱𝗿𝗼𝗻𝗲𝘀   Even when the aircraft itself has near-zero radial velocity, spinning rotors continue to produce a measurable signal. • 𝗧𝗵𝗿𝗲𝗮𝘁 𝗰𝗵𝗮𝗿𝗮𝗰𝘁𝗲𝗿𝗶𝘇𝗮𝘁𝗶𝗼𝗻   Rotor frequency and geometry can reveal drone class, swarm formations, and enable accurate cueing for counter-UAS systems. 🧠 The real breakthrough comes when this data is fed into AI models. Using Short-Time Fourier Transform (STFT), radio reflections are converted into time-frequency spectrograms. These images can then be analyzed by convolutional neural networks in the same way computer vision systems analyze pictures. The result: • automated drone classification   • drastically reduced false alarms   • reduced cognitive load for air-defense operators ⚠️ The bottom line is simple. A fiber-optic drone may defeat jamming. But it cannot defeat physics. Its rotors will always write a signature in the radio spectrum. 💬 𝘐𝘯 𝘮𝘰𝘥𝘦𝘳𝘯 𝘥𝘳𝘰𝘯𝘦 𝘸𝘢𝘳𝘧𝘢𝘳𝘦, 𝘵𝘩𝘦 𝘤𝘢𝘣𝘭𝘦 𝘮𝘢𝘺 𝘣𝘦 𝘴𝘪𝘭𝘦𝘯𝘵, 𝘣𝘶𝘵 𝘱𝘩𝘺𝘴𝘪𝘤𝘴 𝘯𝘦𝘷𝘦𝘳 𝘪𝘴.

🚁 Distributed Autonomy + Radar Intelligence in Drone Swarms In this simulation, I demonstrate how a swarm of autonomous drones can cooperatively search, detect, track, and neutralize a dynamic target — without any central controller. Each drone operates with its own directional radar, limited field-of-view, and noisy measurements. Individually, their perception is imperfect. Collectively, it becomes powerful. Here’s what’s happening under the hood: ✅ Distributed radar-based area coverage ✅ Probabilistic target detection under SNR and beam-pattern constraints ✅ Multi-sensor fusion for precise localization ✅ Confidence-driven mode switching (Search → Focus → Hunt & Destroy) ✅ Cooperative containment geometry for safe engagement ✅ Fully decentralized decision-making When a single drone detects a target, it shares its estimate. As more radars observe the same object from different angles, localization uncertainty collapses through geometric diversity — just like in real multi-static radar networks. Once collective confidence crosses a threshold, the swarm automatically transitions from exploration to coordinated pursuit and encirclement. No “master” drone. No centralized planner. Just local intelligence + communication + control. This kind of architecture is highly relevant for: • Defense and surveillance • Airspace security • Search-and-rescue • Law Enforcement • Large-scale robotic systems And it’s a great example of how signal processing, estimation theory, control, and AI come together in real systems. Still plenty to optimize — but a strong foundation for truly autonomous cooperative sensing. Happy to discuss the math, radar models, or system design in the comments. 👉 About me: I’m Dr. Nir Regev — a professor and radar engineer with 28 years of industry experience. I work at the intersection of sensors, statistical signal processing, AI, and autonomous systems. I also teach engineers and innovators how to turn theory into real-world systems at Regev’s Radar & AI Academy: academy.drnirregev.com #AutonomousSystems #Radar #MultiSensorFusion #SwarmIntelligence #AIEngineering #Robotics #SignalProcessing #DistributedSystems #DefenseTech

Stealthy detection of UAS via passive RF sensing. What if you could detect drones using RF passive sensing without relying on radar, ISAC or similar methods? No RF illuminator needed. Using ELF signatures emitted from spinning rare-earth magnets in the UAS motors we can not only detect, but also identify the type of drone via neurosymbolic AI methods. See our latest paper in IEEE Sensors. This work is in collaboration with the University of Ruhuna, and the Brookhaven Nation Labs NY. Coauthors include Chatura Wickrema Seneviratne Soumyajit Mandal Nimasha Hiruni Silva Supun Ganegoda Sudeepa Ranasinghe Dilshara Herath Early accepted paper is attached.

UAV/Drones can be detected and tracked using a variety of technologies. Two key approaches are 'Active' and 'Passive' detection. Active detection, where a signal is Transmitted to an object and reflection analyzed has a disadvantage in military use because the emitted signals can be detected by enemy. Passive detection focuses on Receiving, and analyzing, signals emitted by objects and are not detected by adversaries. With massive rise of counter-UAVs and drones, can expect huge technology growth in both forms of detection? Thoughts? "Active drone detection: involves emitting signals to detect the presence of drones. Systems transmit radio waves or light pulses, which reflect off objects and return to the sensor, providing information about the object’s location and movement. Active sensors send out a pulse of energy and detect the changes in the return signal. Radars emit radio waves that bounce off objects and return to the sensor, providing information about the objects’ location and movement. LiDARs emit light pulses that reflect off objects and return to the sensor, providing information about the objects’ location and movement. Active drone detection methods carry risks in military operations, as they emit signals that can be detected by adversaries." "Passive drone detection involves using methods to detect the drone and/or its controller without emitting any signals. Passive sensors detect energy emitted or reflected from an object. RF drone detection systems receive and analyze the communication signals between the drone and its controller. Acoustic sensors receive and analyze the sounds emitted by drone motors and propellers. Cameras detect drones by capturing and, in some cases, analyzing images or videos of the surrounding area. These can be standard optical cameras and/or more advanced systems, such as infrared cameras for night vision. Passive radar systems detect and analyze reflected signals from other electromagnetic sources. They provide information on the location and movement of objects without emitting their own signals." #uavdronedetection #uavdronetracking https://lnkd.in/egR5iMwe

YOU DON’T NEED ACCESS TO DETECT A DRONE. 📡 This concept highlights an often overlooked reality in counter UAS: A drone does not need to successfully connect to a network to reveal itself. 📡 Detection through signaling, not connection Even when a network rejects access, the drone continues to send attach requests. These signaling attempts move through the core network and can be analyzed. This creates valuable indicators: Cell location of the device Repeated attach attempts across cells Mobility patterns that differ from handheld devices Device identifiers such as IMEI where available. 🧠 What this means operationall Every failed connection attempt becomes a data point. Over time, these points form a movement pattern that can indicate the presence of a drone. This turns existing telecom infrastructure into a passive detection layer. No additional sensor deployment required. ⚙️ From communication network to sensing layer Traditionally, networks were built to connect devices. Now they can also help identify anomalies in the airspace. The key is not allowing access. The key is understanding the behavior of devices trying to connect. 🛡️ Relevance for security and critical infrastructure For perimeter protection, this adds a new dimension: Wide area detection without line of sight Integration into existing infrastructure Early indication of suspicious movement patterns. This is especially relevant in urban environments where classical sensors face limitations. 💡 What to consider Data access and privacy constraints Integration into command and control systems Correlation with other sensor layers such as RF or EO Clear definition of normal versus abnormal mobility patterns. It requires the right operational concept. ❓ Question for discussion Do you see telecom networks becoming a standard layer in counter UAS architectures or will they remain a niche capability for specific environments?