But with fielded platforms advancing rapidly, relying solely on a living catalog of continually updated signal information creates an issue. Updates keep the catalog current, but not ahead of the pace of drone innovation.
To match pace with the threat, DroneShield created the RfAI-3 detection engine to sense drones that have never been encountered before, with RfRecon placing that capability into the hands of operators in the field.
Remote ID Is Limited
Remote ID is the first and most basic level of drone detection. A large majority of retail drone models are equipped with the technology, which acts as a digital license plate, broadcasting the platform’s identity and location information. While valuable for identifying retail drones operating as intended, it’s limited to aircraft that are actively transmitting Remote ID data.
The Catalog Has a Ceiling
Beyond basic Remote ID, traditional RF detection systems operate on the same essential logic:
1. A drone emits a signal.
2. That signal gets matched against a catalog of known RF signatures.
3. If there’s a match, the drone is detected.
4. If there isn’t, the drone is invisible.
Despite its enduring utility, sole reliance on the catalog has limitations. A drone’s signals must first be recorded, analyzed and a signature created before the platform can be reliably identified. This must occur frequently to maintain currency.
When a new emitter is first encountered, there may be no confirmed alert to a drone threat because the detection engine cannot match the platform’s emissions to a known signature.
The issue lies in the fact that commercial manufacturers continue to release new models, while novel DIY and purpose-built platforms are being developed across a growing range of frequencies, protocols and operating behaviors. The signal database must continue to grow.
Signature Matching Can Be Exploited
Commercial drones may frequency hop to maintain a reliable connection, making their signals harder to match as they move rapidly across an already crowded RF spectrum.
Adversarial systems can take this further, using unconventional frequencies and protocols specifically to avoid signature matching.
As adversaries create DIY and purpose-built UAS operating on frequencies and protocols outside known parameters, catalog-dependent detection creates a vulnerability that can be actively exploited.
RfAI-3 Can Sense Drones Never Encountered Before
RfAI-3 resets the approach. While the knowledge gleaned from the catalog remains significant, signature matching alone cannot keep pace with the threat.
Built on three generations of RF artificial intelligence development, years of signal data collection and operational field experience, RfAI-3 uses machine learning to analyze RF emissions and recognize activity characteristic of UAS operation.
Rather than requiring a known signature, RfAI-3 can detect entirely new drone models that have never been encountered before.
Detection is no longer limited to platforms whose signals have already been captured, characterized and added to the catalog.
Fielding RfAI-3 Capability with RfRecon
To move RfAI-3 beyond the lab and into the field, DroneShield launched RfRecon.
Powered by the RfAI-3 detection engine, RfRecon is a rugged, rapidly deployable RF detector that allows operators to identify known and previously unseen drone activity. It combines ultra-wideband RF sensing and AI-powered analysis in a portable system designed for rapidly changing operational environments.
For operators, the advantage is clear. Even an entirely new drone platform can be detected and a line of bearing can be established during its first encounter.
Military and government operators face the challenge of detecting new and modified drone platforms in contested environments, while public safety and critical infrastructure teams face expanding commercial and DIY drone activity around sensitive locations.
In each case, operators require reliable detection of known and novel drone activity.
