LONG-TERM GOALS

To significantly reduce the probability of false alarm in Navy active sonar systems using predictive physics-based models. OBJECTIVES To develop and apply probability density functions (PDF's) to describe non-Rayleigh echoes from single realistic objects for eventual use in classifiers. APPROACH The echo statistics associated with a randomly oriented prolate spheroid randomly located in the sonar beam are modeled. The cases of a smooth and rough spheroid, both in the water column (i.e., case in which boundaries are not involved), and on a boundary are investigated. The modeling begins with the analytical physics-based expression for the scattering amplitude of a smooth prolate spheroid. The following properties are then incorporated which leads to a random echo: randomly rough outer boundary of spheroid, random orientation, and random location in the sonar beam. The contributions of each property to the echo statistics are studied both separately and collectively. Graduate studen