SIGNAL PROCESSING TECHNIQUES FOR CONSTANT FALSE ALARM RATE DETECTORS IN A DENSE TARGET ENVIRONMENT

IT IS PROPOSED TO DEVELOP AND EVALUATE A NEW CLASS OF CONSTANT FALSE ALARM RATE (CFAR) ALGORITHMS FOR USE IN DENSE TARGET ENVIRONMENTS, BASED ON ADVANCED MULTIDIMENSIONAL DIGITAL SIGNAL PROCESSING METHODS. EXISTING CFAR TECHNIQUES EMPLOY A VARIETY OF PRIMITIVE ONEAND TWODIMENSIONAL DIGITAL FILTERS, BOTH LINEAR AND NON-LINEAR. MORE POWERFUL FILTERING AND PROCESSING METHODS HAVE EMERGED FROM RECENT RESEARCH EFFORTS IN MULTIDIMENSIONAL SIGNAL PROCESSING. THE PROPOSED WORK IS DIRECTED AT ADAPTING AND APPLYING THESE NEW METHODS TO THE DENSE-TARGET CFAR PROBLEM. THE BASIC APPROACH WILL BE TO TREAT RANGE-DOPPLER RADAR MAPS AND ANGLE-ANGLE IR MAPS AS IMAGES THAT CAN BE PROCESSED BY A VARIETY OF MULTIDIMENSIONAL SIGNAL ENHANCEMENT ALGORITHMS IN ORDER TO ELIMINATE STATISTICAL ANOMALIES FROM THE DATA. THRESHOLDS FORMED FROM THE PROCESSED IMAGE WILL BE FREE OF TARGET CONTAMINATION AND, WHEN APPLIED TO THE ORIGINAL DATA, SHOULD YIELD GOOD CFAR PERFORMANCE REGARDLESS OF TARGET DENSITY. THE COMPUTATIONINTENSIVE NATURE OF MULTIDIMENSIONAL SIGNAL PROCESSING ALGORITHMS IMPLIES THAT SMALL, LIGHTWEIGHT HARDWARE IMPLEMENTATIONS WILL REQUIRE THE USE OF VHSIC DEVICES AND/OR CUSTOM VLSI CHIPS. INCLUDED IN THE PROPOSAL IS A HIGH LEVEL ARCHITECTURAL DESIGN TASK AIMED AT ASSESSING THE COMPATIBILITY OF THESE DEVICES WITH ADVANCED CFAR REQUIREMENTS.