Australian company DroneShield announced the rollout of its first fully Machine Learning/AI based detection and classification software to all of its existing customer systems.
DroneShield utilises its proprietary techniques in signal processing and Machine Learning/AI to do near-real time detection and identification of unmanned robotic systems and, more broadly, other potential threats in the Electronic Warfare fields. The result is a dramatic increase in detection responsiveness, lower false positives and a significant increase in the speed at which new threats are detected, classified and tracked by the DroneShield systems.
One of the key achievements that sets DroneShield’s technology substantially ahead of the existing technologies globally, is a very lightweight machine learning architecture designed to run on low power FPGA (Field-Programable Gate Array) hardware. This enables the system to be deployed for long term periods in power scarce, air-gapped environments.
The software is designed to run on all DroneShield platforms including RfPatrol, DroneSentry and DroneSentry-X.
Future device software updates will build on this system architecture and increase performance and the number of detectable threats.