Intelligent Wideband Signal Cueing for ESM

Today’s electromagnetic spectrum monitoring (ESM) missions require an increasingly wide radio frequency (RF) bandwidth to intercept, process, and analyze a myriad of threat signals including communications, data links, surveillance radars, and targeting radars.  These emitter waveforms often employ sophisticated modulation techniques including phase modulation, spread spectrum, linear frequency modulation, and frequency agility that further emphasizes the need for wide bandwidth coverage.  The proliferation of commercial communications systems (5G and point-to-point microwave back hauls) adds to the complexity and density of the contested RF spectrum further emphasizing the need for wide bandwidth, multi-channel ESM systems to establish and maintain dominance in the RF spectrum domain. Legacy ESM systems that routinely covered up to 6 GHz must now be expanded to include X-band (8-12 GHz) and Ku-band (12-18 GHz) to address emerging threats in that portion of the RF spectrum. A critical enabler for an enhanced wide bandwidth ESM system is a Machine Language (ML) cue generator. ML cue generators are designed to identify and extract features or "cues" from data that can be used as input to an ML signal classification model. These cues are necessary for the model to learn patterns, make predictions, or classify data effectively. A cue generator might extract characteristics from a signal that are indicative of specific patterns. The cue generator includes functions for feature extraction, preprocessing, data reduction, and pattern recognition.   Careful consideration should be given to the cue generator design to accommodate operations over wide bandwidth, scalability, real-time processing, and external interfaces for pattern of life data.BANC3 specializes in the design, development, integration, and test of extremely wide instantaneous bandwidth software defined RF spectrum monitoring systems for military/aerospace missions.  We have successfully deployed wide bandwidth systems covering 10 MHz to 18 GHz in challenging real world ESM conditions.  Building on this experience, we have implemented and tested an ML cue generator for use in wide bandwidth RF threat signal recognition.  This technology has been successfully tested using radar, radio control, and low probability of intercept communication signals. This cue generator will be the basis for the proposed research to develop an N-channel cue generator for use in future ESM systems.  The cue generator architecture will provide multiple parallel cue generator pipelines and a cue generator process that prioritizes and schedules the signal types for action by the ESM system.  The proposed configuration can be hosted on commercially available field programmable gate array (FPGA), embedded processor, and server. The proposed architecture will be tested in simulation and laboratory configurations.