Feature Aided Association Module (FAAM)

In this Phase II SBIR project, Daniel H. Wagner Associates will develop a full-scale prototype Feature Aided Association Module (FAAM) that will improve the accuracy of Track Management or Data Fusion systems when processing measurement data containing non-kinematic or feature information. FAAMs processing of the feature information will significantly improve the ability of the Track Management or Data Fusion system to create hypotheses that are more likely to contain the correct correlation decisions and to more accurately estimate the probabilities of correctness for each hypothesis. The underlying technical mechanism by which FAAM treats the feature information is a Bayesian Network (BN) based on the taxonomy of air targets and the types of measurements available for estimating the various attributes that characterize the targets. This BN estimates the identification/classification of each track, which permits a more accurate calculation of the association likelihood between a given track and a sensor measurement containing feature data. In Phase II, we will enhance FAAM to reflect more realistic sensors and taxonomy, to accommodate the processing of dependent data, and to use ID/classification evidence derived from kinematic data.