Neural Network Probability Data Association for Radar Tracking

DANIEL H. WAGNER ASSOCIATES, INC. PROPOSES TO DEVELOP AND DEMONSTRATE A NEURAL NETWORK BASED ALGORITHM FOR TRACKING RADAR TARGETS IN THE PRESENCE OF CLUTTER. THE ALGORITHM WILL CLOSELY APPROXIMATE THE TRACKING PERFORMANCE OF A JOINT PROBABILISTIC DATA ASSOCIATION (JPDA) FILTER, BUT THROUGH THE USE OF NEURAL NET HARDWARE, DO DO AT A MUCH HIGHER DATA RATE THAN OTHERWISE POSSIBLE. IN PHASE I, WAGNER ASSOCIATES WILL DEFINE AND IMPLEMENT A PROTOTYPE VERSION AND COMPARE ITS PERFORMANCE AGAINST A MULTIPLE HYPOTHESIS DATA ASSOCIATION ALGORITHM, SEQUENTIAL NEAREST NEIGHBOR (SNNA), AND THE TRADITIONAL JPDA.