Drones of all sizes, from your basic “hobbyist” version (although many are quite sophisticated) to UAVs the size of some human-piloted aircraft, are getting lots of attention as their capabilities increase dramatically in terms of range, navigation, sensors, payloads, and autonomous behavior. They have quickly and effectively improved form being “merely” observers to being weapons platforms which can carry anything from grenade-sized loads to small missiles.
This danger, of course, brings a need to detect the drone’s approach and presence, and that’s not an easy task. Due to their relatively small radar cross section, ability to fly low, low acoustic noise, and high maneuverability, they are hard to detect until it may be too late to do “something.”
What does it take to detect drones before that too-late period? In brief, it takes a multidisciplinary approach which combines conventional radar and optical sighting, of course, but also intense RF-signal capture and signal analysis. That RF aspect is severely challenging, as a drone’s uplink/downlink signals are in a sea of RF sharing the same spectrum, and is likely encoded as well.
A recent issue of Microwave Journal included an Aerospace and Defense Electronics Supplement which addressed the drone-detection problem under the title “Drone Detection and Location Systems.” Rather than have a single article or author, the issue linked three articles from leading providers of such systems or the functions and subsystems needed to build them: Keysight Technologies, Rohde & Schwarz, and Aaronia AG. While each sub-article had some self-serving aspects – as you’d expect – each also had a wealth of insight into the many problems associated with drone detection, the solutions, and the difficulties that each solution encounters.
The problem of capturing a fairly unknown RF signal is formidable. In his landmark 1968 textbook, “Detection Estimation and Modulation Theory, Part I: Detection, Estimation, and Filtering Theory,” (See Figure below) Harry L. Van Trees classified signal-recovery difficulty into classes of increasing difficulty. (Note that Parts II, III, and IV are even-more intense and insightful, if you have the need.) The easiest case is simple detection (absence/presence) of a “known signal in known environment” where the signal format, modulation, and frequency are known in advance, with a modest SNR and co-signal environment. The most difficult is the estimation (analog demodulation) of a signal with characteristics which are largely unknown in advance, in an environment which is largely unknown and hostile as well. The drone RF-signal detection and recovery problem is at the more-difficult end of that range.
Despite being first published nearly 50 years ago, this classic book still has considerable relevant insight into the fundamental problems of signal capture and recovery in a wide variety of environments and scenarios.
For drone detection, the problem naturally begins with signal capture in an unknown and challenging RF environment. That’s where the integration of analog functions from front-end amplifiers through the RF chain is critical.
It’s interesting to view the complementary nature of the situation. Drone technology has advanced largely due to convergence of dramatic improvements in electric motors, batteries, MEMS-based accelerometers and gyros, video capture, GPS, microprocessors in many forms, and more. Therefore, the defensive drone-detection side has had to “up its game” by improvements in, and convergence among, its own technologies including software-defined radio, signal and signature analysis, and more. This is the 21st-century chapter in an offense/defense story which has occurred many times in the past.
Were you aware of the multifaceted challenges of the drone-detection problem, or what is being done to address it?