APPLICATION OF ADAPTIVE NEURAL NETWORKS FOR SPECIFIC EMITTER RECOGNITION/IDENTIFICATION

THIS IS THE FIRST PHASE OF A MULTI-PHASE ADVANCED DEVELOPMENT EFFORT TO PROVIDE ADAPTIVE SIGNAL RECOGNITION ALGORITHMS THAT CAN UNIQUELY IDENTIFY AND FOLLOW THE CHANGING NATURE OF COMMUNICATION EMITTERS AND TO IMPLEMENT THESE ALGORITHMS IN A NEURAL NETWORK ARCHITECTURE. IT WILL DEMONSTRATE THE UTILITY OF EXPLOITING ART-E OR SIMILAR NEURAL NETWORKS TO ACHIEVE A GREATLY ENHANCED CAPABILITY TO PERFORM SPECIFIC EMITTER IDENTIFICATION. THIS WILL BE ACCOMPLISHED BY CONDUCTING A PERFORMANCE ANALYSIS USING AN EXISTING SIGNAL RECOGNITION SIMULATOR. THIS INITIAL PHASE WILL ALSO ENHANCE EXISTING SEI METHODS BY DEFINING A NEW NEURAL NETWORK PROCESSING CONCEPT WHEREBY THE RECOGNIZER WILL NOT ONLY FOLLOW THE SIGNALS IN TIME (AS THEY CHANGE THROUGH NORMAL AGING OF COMPONENTS OR MODE CHANGES), BUT ALSO BE ABLE TO "LEARN" OVER TIME TO PERFORM MORE ACCURATE RECOGNITION.