Autonomous Classification of Acoustic Signals

U. S. Navy tactical and strategic forces need a real-time, autonomous classification capability to realize the full potential of fixed acoustic surveillance sensors. In this work, we demonstrate the feasibility of integrating a novel, physics-based automatic classification algorithm into a new, large aperture Planar Array Prototype (PAP). The approach differs from traditional"intel-based"classification methods by using fundamental knowledge of normal mode propagation constraints to exploit natural differences in the way surfaced and submerged sources excite the shallow water waveguide. OASIS has successfully employed this approach in the past on both towed and fixed horizontal line arrays. The Phase I effort will determine the feasibility and identify any technical issues associated with applying the concept to the PAP arrays under consideration by PMS-485. Under the Phase I Option, the sizing and timing requirements associated with the implementation of the PAP classification algorithm on a commercial off-the-shelf (COTS) hardware platform will be analyzed. Finally, a plan will be written that details the steps for integration of the algorithm into the Shallow Water Surveillance System (SWSS) autonomous detection, classification, and tracking (DCT) baseline, which leverages the OASIS-led real-time DCT development and integration effort under the ONR Persistent Littoral Undersea Surveillance (PLUS) program.