HIGH-SPEED TRACKER ALGORITHMS STUDY

ADVANCES IN SIGNAL COLLECTION AND PROCESSING CAPABILITIES HAVE PLACED EVER INCREASING DEMANDS ON INFORMATION PROCESSING CAPABILITIES. INCREASED DATA RATES AND SENSOR CAPABILITIES HAVE ALSO INCREASED THE IMPORTANCE OF CORRELATION TECHNIQUES. THIS HAS LED TO A PROLIFERATION OF TRACKING AND CORRELATION ALGORITHMS, BUT THE PROCESS OF SELECTING AMONG VARIOUS MODELS AND TECHNIQUES IS STILL MORE ART THAN SCIENCE. IN FACT, THE SEEMINGLY BEWILDERING ARRAY OF TRACKING ALGORITHMS CURRENTLY AVAILABLE ARE NOT AS DISPARATE AS THEY APPEAR. MOST ARE DERIVED FROM A VERY LIMITED SET OF UNDERLYING COMPONENTS, WITH THEIR DIFFERENCES TRACEABLE TO THE MODELING ASSUMPTIONS BEHIND THESE COMPONENTS AND THE MANNER IN WHICH COMPONENTS ARE COMBINED. WE PROPOSE, IN PHASE I, TO DEVELOP A FRAMEWORK WITHIN WHICH TO STUDY, MEASURE AND COMPARE THE CHARACTERISTICS AND PERFORMANCE OF MODERN, HIGH-SPEED, TRACKING ALGORITHMS, PAYING PARTICULAR ATTENTION TO THE EFFECTS OF THEIR VARIOUS COMPONENTS. THIS EFFORT WILL INCLUDE AN ANALYSIS OF THE IMPLICATIONS OF THE UNDERLYING MODELS AS WELL AS COMPUTER SIMULATIONS DESIGNED TO HIGHLIGHT THE DIFFERENCES BETWEEN ALGORITHMS. THE RESULTS OF THIS EFFORT WILL BE A FUNCTIONING SYSTEM CAPABLE OF ANALYZING THE CHARACTERISTICS OF THE COMPONENTS OF A MODERN TRACKER AND OF THE TRACKER AS A COMPLETE PACKAGE.