Bayesian Multi-Array Search Aid

This work will develop a Bayesian Acoustic Multi-Array Search Aid for IUSS using the multiple-target tracker, Nodestar TMA as its basis. Nodestar TMA is currently used aboard Los Angeles class submarines. It incorporates bearing information from multiple acoustic arrays on board the submarine to produce multiple target tracks in geographic space. It associates bearing information to targets using purely kinematic information. This work will extend Nodestar to use both spectral and kinematic information in a joint Bayesian manner to perform association. The association methodology will include a confidence level that can be adjusted to require more or less confidence in an association before fusing the information. Each Nodestar target track will contain a pointer to the information used to produce the probability distribution for that track. This will allow the user to drill down to examine that information. The phase I work will demonstrate Nodestar's capability to automatically generate multiple target tracks in geographic space using joint kinematic and spectral association of information from multiple arrays. This will demonstrate the feasibility of the approach and will lay the foundation for using other feature information. In addition, we will demonstrate the ability to incorporate range estimates from single arrays