TORPEDO ACOUSTIC PROCESSING U

TIMELY DETECTION, CLASSIFICATION, AND TRACKING OF TORPEDO SIGNATURES ARE OF HIGH INTEREST TO BOTH SURFACE AND SUBSURFACE SHIP DEFENSE SYSTEM PROGRAMS. THE MAJORITY OF THE WORK COMPLETED TO DATE HAS FOCUSED ON THE EXPLOITATION OF THE THREAT SUBMARINE ACOUSTIC ENERGY ENVELOPE. THE PURPOSE OF THIS PROJECT IS TO DEVELOP A LABORATORY FEASIBILITY CONCEPT DEMONSTRATION SYSTEM THAT MAKES USE OF NEURAL NETWORKS, FEATURE EXTRACTORS, AND AN EXTENDED KALMAN FILTER TRACKER, AND IS SPECIFICALLY DESIGNED TO DETECT, CLASSIFY, AND TRACK TORPEDO SIGNATURES. FIRST,ORINCON WILL SELECT TEMPLATES AND TRAIN NEURAL NETWORK RETINAS TO BE OPERATED IN SUPPORT OF DETECTION, CLASSIFICATION, AND ALERT PROCESSES. NEXT, WE WILL IMPLEMENT SIMULATED BEARING ESTIMATION AND TRACKING ROUTINES INTO A HOST COMPUTER APPLICATION, AND DEFINE THE TRIGGERING ALGORITHMS NECESSARY TO ACTIVATE TRACKER INPUTS. THIS WILL BE FOLLOWED BY A LABORATORY DEMONSTRATION OF SINGLE-CHANNEL TORPEDO DETECTIONK CLASSIFICATION, AND ALERTLING WITH SIMULATED BEARING INPUTS TO THE TRACKING SYSTEM. FINALLY, ORINCON WILL QUANTIFY THE NEURAL NETWORK DETECTION AND CLASSIFICATION PERFORMANCE.