NONLINEAR SIGNAL PROCESSING ALGORITHMS FOR ENHANCED C3I APPLICATIONS

THERE IS A GROWING INTEREST IN THE ROLE OF NONLINEAR SIGNAL PROCCESSING (NSP) ALGORITHMS UTILIZING HIGHER-ORDER MOMENTS AND HIGHER-ORDER SPECTRAL MEASUREMENTS FOR FERRETING OUT INFORMATION ABOUT UNKNOWN SIGNALS THAT ARE BEYOND THE REACH OF LINEAR TECHNIQUES. NSP TECHNIQUES DISSECT HIGHER ORDER STATISTICAL SIGNAL STRUCTURE WHICH PERMIT ENHANCED DIFFERENTIATION FROM BACKGROUND. UNLIKE HEURISTICS, NSP OFFER RIGOROUS METHODS TO IDENTIFY AND EXTRACT INFORMATION NOT AVAILABLE FROM LINEAR TECHNIQUES, E.G., THE FFT IS BLIND TO CERTAIN INFORMATION THAT CAN ONLY BE EXTRACTED BY MULTISPECTRA, I.E., "HIGHER-ORDER' SPECTRA. PHASE I DEMONSTRATES THE UTILITY OF FUNDAMENTAL NSP TECHNIQUES UNDERLYING ENHANCED DETECTION/CLASSIFICATION OF BROADBAND SIGNALS EMANATING FROM NONLINEAR SOURCES. THE PROBLEMS ARE: NONLINEAR SYSTEMS WHICH ARE ACTIVELY PROBED, OR PASSIVELY MONITORED, THE LATTER BEING MORE DIFFICULT AND REPRESENTING TWO FUNDAMENTAL PROBLEMS IN DOD SURVEILLANCE REQUIREMENTS: DETECTION AND CLASSIFICATION. A CASE STUDY USES SPREAD-SPECTRUM EMITTERS BECAUSE OF THEIR IMPORTANCE AND BECAUSE THEY GENERATE BROADBAND SIGNALS FROM AN INHERENT NONLINEAR SIGNAL GENERATION MECHANISM. PHASE I DEMONSTRATES THE UTILITY OF THE NSP ALGORITHMS, USING A SET OF PREVIOUSLY SBIR FUNDED NSP DISCRIMINANTS; AND WILL DEVELOP HIGHSPEED COMPUTING ARCHITECTURES FOR NSP APPLICATIONS. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - RECOGNITION OF THE IMPORTANCE OF NSP ALGORITHMS TO PROBLEMS TO DOD. A CHIP ARCHITECTURE FOR HIGH SPEED IMPLEMENTATION OF A FUNDAMENTAL NSP ALGORITHMS FOR PIGGYBACK ENHANEMENT OF DOD SENSOR PLATFORMS WITHOUT MODIFICATION OF FRONT INSTRUMENTATION, AND SPECTROSCOPY. KEY WORDS - BISPECTRUM, NONLINEAR, MULTISPECTRA, SPREAD SPECTRUM, VOLTERRA.