DroneShield today unveiled RfAI-3, the third generation of its proprietary radio frequency intelligence engine. RfAI-3 marks a significant evolution in counter-drone doctrine, prioritizing wideband detection to sense drone threats beyond known and pre-classified emitters, delivering decisive situational awareness in the complex, rapidly evolving threat environments that define modern operations. Industry standard RF detection systems operate on a classified emitter model: a drone must be cataloged in an RF dataset before it can be reliably identified. As drone proliferation accelerates and adversaries deploy increasingly novel, commercially sourced, and purpose-built systems, this limitation creates a critical and growing gap. RfAI-3 inverts the conventional approach. Rather than searching the spectrum for signals it already knows, the model detects the emissions present and then fingerprints each one against previously observed drone signatures. Where a match exists, the emitter is classified with full detail. Where it does not, RfAI-3 generates a signature for the unknown emission and presents it to the operator with a confidence assessment, ranked against the wider spectrum picture. Nothing is discarded for being unfamiliar, and every new signature the model generates becomes a reference for the next encounter.
The RfAI Lineage
RfAI, initially developed in 2018, established the foundational RF detection architecture, delivering real-time classification of known drone emitters with precision that rule-based systems could not match.
RfAI-2, released in 2023, turned the RF intelligence dataset into a continuously expanding asset, with new drone models, protocols, and threat vectors identified through DroneShield’s ongoing acquisition and characterization program and distributed to the deployed fleet.
RfAI-3, unveiled today, builds on this robust foundation while breaking the paradigm. Wideband detection extends capability beyond the RF dataset entirely, surfacing drone emissions outside of any catalog. Combined with the continuously expanding RF intelligence dataset inherited from RfAI-2, the result is a detection capability that compounds over time while simultaneously operating beyond the boundaries of what it has previously encountered.
A defining characteristic of the RfAI architecture, retained and accelerated in RfAI-3, is the continuously expanding RF intelligence dataset. DroneShield’s engineering teams maintain a rigorous program of emitter identification, classification, and integration, ensuring that known threat signatures are cataloged and distributed to the full deployed fleet. This means the system’s performance against classified emitters compounds over time. An RfAI-3-powered deployment today will be more capable six months after commissioning than it is on day one, not because of a hardware upgrade, but because the intelligence it draws on continues to grow. RfAI-3 is specifically designed for DroneShield’s next-generation hardware platforms, with initial releases expected in the second half of calendar year 2026 and further releases through 2027. DroneShield’s existing product lineup continues to be supported under the company’s software subscription model, including ongoing delivery of expanded emitter coverage as the RF intelligence dataset grows.
DroneShield (ASX:DRO) is a publicly listed company with headquarters in Australia, the United States, and the Netherlands. DroneShield specializes in counter-UAS technology leveraging RF sensing, AI, sensor fusion, and electronic warfare. Our solutions protect military, government, law enforcement, and critical infrastructure globally, offering both custom and off-the-shelf products for land, sea, and air platforms. ISO 27001 Certified Australian Government DISP Member.
