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

Vadum will implement and integrate a prototype Context-Aware Machine Learning Signal CAMLS) classification system to recognize radio frequency (RF) emissions and estimate the status of detected signals for the Army’s Next Generation Combat Vehicle (NGCV) platforms. CAMLS improves the vehicle and crew effectiveness and survivability by delivering spectrum situational awareness (SSA) including blue force signal health monitoring, and red force characteristic indication through their use of RF systems. The CAMLS system employs offline trained classifiers with optional online, on mission iterative training. Online signal and emitter characterization techniques are also employed to detect, classify and record signals and emitters of interest. CAMLS operations are based upon an intelligence and data-derived Electronic Order of Battle (EOB) framework that operates with consideration to both new and evolving conditions on the battlefield and pre-mission planning. CAMLS signal classifiers and characterization components create a Dynamic EOB (D-EOB) that evolves to handle both known and unknown signals while maintaining spectrum usage and EOB evolution history for post mission analysis. Phase II of CAMLS will focus on development of an integration level (end-to-end) prototype of at least TRL 6, that is capable of operating at or near real time in realistic multi-emitter environments.