CAPTURE - Context Aware Preprocessing Technology for Ultra -Wide Band RF Exploitation

The radio frequency (RF) spectrum encompasses many technological infrastructures which are critical to the Navy. Such are the global navigation satellite system (GNSS), early warning radars, command-and-control for unmanned aerial systems (UAS), among others.  Naval operations require real-time monitoring of large segments of the RF spectrum to determine the impact of RF transmissions on missions and operations. Etegent proposes CAPTURE, a Context Aware Preprocessing Technology for Ultra-Wide Band RF Exploitation. CAPTURE utilizes proven deep learning (DL) techniques, advanced scheduling combined with proven distributed edge computing to provide the Navy with real-time monitoring of ultra-wide spectrum bands.  CAPTURE divides multi-gigahertz instantaneous RF bands into megahertz sub-bands where data can be processed by inherently parallel DL algorithms which perform detection and estimation on 2D transformations of the raw I&Q data. Novel DL networks provide detection, classification, and anomaly detection to discriminate which spectral bands should be stored/post-processed or discarded effectively reducing compute and storage requirements of down-stream Navy systems.   During Phase I, Etegent will leverage an ultra-wide band RF collection system to stream 4GHz of instantaneous bandwidth to be pre-processed by CAPTURE. This setup enables Etegent to assess the feasibility of key components of the proposed solution and demonstrate its performance, scalability and through-put. The base Phase I effort assesses key CAPTURE technologies, while the Phase I option investigates performance and optimization of those technologies.