AIBRIEF
DroneShield's new RfAI 3, its third generation radio frequency AI engine for counter drone systems, scans a wider radio band and assigns a confidence score to drone emissions the system has never catalogued.
DroneShield, an Australian counter-drone vendor, says its third-generation radio-frequency engine can flag drone emissions no existing database has catalogued, and assign a confidence score to signals the system has never seen.
The product, RfAI-3, extends a line the company has been building since 2018, when it released the original RfAI, and 2023, when RfAI-2 turned the company's RF signature library into an asset shipped to deployed customers. Counter-drone, or CUAS, systems are the sensors and software used to spot hostile or unauthorized drones near airports, prisons, military bases, stadiums, and other sensitive sites. Most rely on a catalog: they look for emitters that match signatures already on file, and ignore anything that does not.
DroneShield says RfAI-3 works the other way around. The engine scans a wide radio band first, fingerprints any drone-like emissions it finds, and generates a signature plus a confidence score for signals that do not match anything in the prior library. The system keeps each new signature, so the next encounter has something to compare against.
CEO Angus Bean called the release "a fundamental shift" and said wideband detection "changes the calculus for our operators and for the adversaries trying to evade them." That framing is the vendor's own.
The press release does not include independent benchmarks, field-test data, named customers, or third-party validation of the wideband claim in contested environments. (sUAS News republishing the DroneShield announcement)