Machine Learning for Radio Frequency (RF) Signatures Detection and Classification System

Gathering the spectrum blue print of the operational environment is pivotal in gaining actionable intelligence in a tactical electronic warfare scenario. The electronic countermeasure systems of combat vehicles possess SIGINT systems that allow operators to observe the RF operational environment and identify the friendly/foe systems in their vicinity. The tactical operational scenarios often involve multiple RF transmissions which motivates the need for a cognitive SIGINT system that can operate in congested RF scenarios. To address these challenges, ANDRO proposes INtelligent SIgnal Detector and classifiER (INSIDER), an automated machine learning based unified software approach which will be engineered to operate in remote congested hostile tactical scenarios. INSIDER’s modularity and open framework allows for seamless integration to existing APS and Modular Active Protection System (MAPS) solutions. To be in pace with the evolving RF systems, INSIDER’s unique design is envisioned to accommodate future waveforms without requiring any additional human involvement in developing algorithms. Additionally, INSIDER will be trained and tested with a comprehensive dataset that comprises a rich collection of modern LPI, conventional radar and communication signals developed during this effort.