Signals intelligence systems use the established Shannon-Nyquist approach of signal acquisition, setting a sampling requirement based on data rate. This sampling approach is generating operationally impractical amounts of data, most of which goes unused or arrives late. The conventional rate of data-to-information (DTI) conversion is low, and there exists high demand for a solution. A new approach to signal sampling achieves efficient DTI conversion. Our approach exploits mathematical sparsity to produce high accuracy information sensing (ISENSE) with reduced sampling rates. This technique employs the relatively new science of compressive sensing (CS) and compressive processing (CP), which extract information with fewer samples compared than when sampling at the Nyquist rate. As an extension to known CS techniques, ISENSE continuously validates sparsity patterns, taking CS from theory to a deployable technology. We apply ISENSE to the intercept, detection, characterization, classification and identification of emitters. We will illustrate, by qualitative and quantitative measures, how the ISENSE capability maintains DTI accuracy and detection probabilities despite reduced sampling rates, thus improving Naval airborne SIGINT data collection, storage, transport, and processing efficiency.
Keywords: Specific Emitter Identification, Specific Emitter Identification, Sigint/Electronic Support, Electronic Warfare, Nyquist, Compressive Classifier, Compressive Sampling, Relativ
