A physical statistical sonar clutter model for spatially-dispersed scatterers

The limiting influence of clutter on active sonar performance has long been recognized, with echoes from complex bathymetry, fish schools, and anthropogenic objects such as wrecks, capable of producing false target detections that can both mask real targets and overload the sonar operator with false alarms. To help reduce the impact of these false target objects, statistical distribution models have been employed to both optimize detection threshold settings and improve echo classification. While many such distribution models exist, very few have a physical basis allowing extrapolation to different environments and sonar systems. This paper describes a new 3-parameter statistical model that provides a physical context for relating the characteristics of matched-filter echo-envelope distributions to scatterer attributes. It extends our 2-parameter Poisson-Rayleigh (P-R) model [1]-[2] by adding a quantitative measure of scatterer spatial uniformity (dispersion) to the P-R model's measures of discrete scatterer density and relative strength. The resulting Conway-Maxwell-Poisson-Rayleigh (CMP-R) model's potential for isolating and characterizing clutter was quantified via shallow-water experimental data containing returns from biologic, geologic and anthropogenic objects with differing spatial separations and scattering characteristics. Analysis results demonstrate the flexibility of the CMP-R model in characterizing differing types of clutter as well as target-like objects in 2 distinct environments. Combined with a demonstrated insensitivity to echo signal-to-noise ratio (SNR), the CMP-R distribution model shows promise with regard to both its general applicability for clutter characterization and the development of robust, physics-based active classification algorithms.