ATTRIBUTE-AUGMENTED MULTI-HYPOTHESIS ACOUSTIC DATA FUSION

AN IMPORTANT PROBLEM IN TARGET TRACKING AND SURVEILLANCE IS THE TASK OF FUSING ACOUSTIC MEASUREMENT DATA FROM MULTIPLE SENSORS INTO A SINGLE TARGET REPORT. DATA FROM THE DIFFERENT SENSORS SHOULD BE ASSOCIATED SO THAT OBSERVATIONS OF THE SAME TARGET FROM DIFFERENT SENSORS ARE GROUPED TOGETHER, WHILE DISTINCT TARGETS ARE SEPARATED. THE CORRELATION PROCESS USES BOTH GEOPOSITIONAL INFORMATION (RANGE, BEARING, OR GEOLOCATION) AND ATTRIBUTE INFORMATION SUCH AS LINE FREQUENCY, BANDWIDTH, AND STABILITY IN THE CASE OF ACOUSTIC MEASUREMENTS. PREVIOUS WORK ON TACKING ALGORITHMS HAS FOCUSED ON GEOPOSITIONAL SCORING, AND LITTLE IMPORTANCE HAS BEEN PAID TO THE USE OF ATTRIBUTE INFORMATION BEYOND SIMPLE GATING TO PREVENT MIXING OF REPORTS AND OBVIOUSLY INCONSISTENT ATTRIBUTES. THE WORK PROPOSED HERE INVOLVES THE DESIGN, TEST, AND EVALUATION OF A SYSTEM TO FUSE ACOUSTIC DATA USING AN EXISTING MULTI-HYPOTHESIS DATA FUSION ALGORITHM INTERFACED WITH A RULE BASED EXPERT SYSTEM TO BE DEVELOPED IN THIS PROJECT IN ORDER TO RESOLVE AMBIGUITIES REMAINING AFTER GATING AND GEOPOSITIONAL SCORING AND TO OTHERWISE ADJUST HYPOTHESIS SCORES BY USING ATTRIBUTE INFORMATION. EMPHASIS WILL BE PLACED ON DEVELOPING EFFICIENT DATA MANAGEMENT TECHNIQUES AND KNOWLEDGE SOURCES TO ENSURE THE ABILITY TO HANDLE LARGE NUMBERS OF CONTACT REPORTS IN A TIMELY MANNER WITHIN AN ENVIRONMENT TYPIFIED, FOR EXAMPLE, BY THE ASWCS OR ACS.