MULTICHANNEL SYSTEM IDENTIFICATION AND DETECTION USING OUTPUT DATA TECHNIQUES

THE PROBLEM OF MULTICHANNEL SYSTEM IDENTIFICATION AND DETECTION O DESIRED SIHNAL INFORMATION IN THE CHANNEL WILL BE ADDRESSED BASED ON TECHNIQUES THAT UTILIZE CHANNEL OUTPUT DATA DIRECTLY, WITHOUT THE REQUIREMENT TO ESTIMATE CHANNEL OUTPUT CORRELATION MATRICES. A MODEL-BASED APPROACH IS PROPOSED FOR MULTICHANNEL SYSTEM IDENTIFICATION AND DETECTION, WHERE THE PARAMETERS OF A STATE-SPACE MODEL ARE IDENTIFIED, AND THE MODEL INNOVATIONS SEQUENCE IS USED FOR DETECTION IN A LIKELIHOOD RATION FORMULATION. RECENT STATE-SPACE SYSTEM IDENTIFICATION RESULTS WILL BE EXTENDED AND APPLIED TO THE VERY BROAD AND INCLUDES OTHER WELL KNOWN MODEL TYPES (SUCH AS AUTO-REGRESSIVE AND AUTO-REGRESSIVE MOVING AVERAGE MODELS). AS SUCH, IT OFFERS POTENTIAL FOR A BETTER MODEL FIT IN THE CASES WHERE OTHER MODELS ARE INADEQUATE OR APPROXIMATE. THE DETECTION ASPECT OF THE TECHNIQUES IN THE CONTEXT OF RADAR SURVEILLANCE.