Joseph M. Fialkowski

Methods for Identifying and Controlling Sonar Clutter

Three methods for identifying, characterizing, and controlling active sonar clutter are described and demonstrated on broadband, low-frequency (<;1800 Hz) normalized data collected in two range-dependent, shallow-water environments: the Malta Plateau off Sicily and the Stanton Banks off Scotland. One method examines short time-scale clutter persistence to stabilize/minimize coherent propagation effects. A second method correlates strong clutter echoes with the depth-weighted topographic slope (relative to the sonar) to help predict the significant scatterers. A third method involves moment-based statistical measures [Poisson-Rayleigh-inspired method (PRIM), K-distribution shape parameter, kurtosis] to distinguish between spatially compact and spatially extended clutter objects. As illustrated with the isolation of several anthropogenic objects from widespread topographic clutter on the Malta Plateau, the new Poisson-Rayleigh (P-R)-based method appears to offer some promise in regards to reducing the false alarm rate.

Measurements and modeling of low-frequency near-surface scattering statistics

High-resolution acoustic measurements of low-frequency near-surface backscattering at low grazing angles have been made in the open ocean using vertical arrays of coherent sources. Over the range of wind speeds (4-18 m/s) encountered, the normalized data amplitudes exhibited variable non-Rayleigh behavior, from near Rayleigh in the highest sea states to near lognormal in low-to-moderate sea states. Seven probability density function (pdf) models were fit to the data, with the three-component Rayleigh mixture providing the most consistent fits and the least errors. One pdf model, the Poisson-Rayleigh, provided not only good fits to many data sets, but also physical insights into the scattering process. This models estimates of the expected number of discrete scatterers ranged from 200/km/sup 2/ at low wind speeds to 2000/km/sup 2/ at high wind speeds, consistent with the expected densities of fish and subsurface bubble clouds, respectively. These results are encouraging with regard to developing physical models capable of using local results (such as these) to accurately predict long-range reverberation and clutter statistics.

The Influence of the Sea Surface and Fish on Long‐Range Reverberation

Acoustic detection for active sonars involves identifying target signatures in the presence of environmental effects, such as acoustic scattering from the ocean boundaries and fish. The Naval Research Laboratory has recently developed 3D broadband models that provide physics‐based estimates of the dependence of scattering from the sea surface, bubble clouds and near‐boundary fish (including boundary‐interference effects) on the incident and scattered angles, and physical/biological descriptors of the environment. In this paper, these models and a surface‐loss model are used as kernels in reverberation models, which in turn are used to assess the sensitivity at 3.5 kHz of long‐range reverberation to environmental variables. It is shown that the acoustic field in shallow water waveguides could be quite sensitive to the values of sea surface (wind speed) and fish (density, size, depth) parameters, and that physics‐based models are needed for accurate field characterization.

Assessing the Variability of Near-Boundary Surface and Volume Reverberation Using Physics-Based Scattering Models

The increased importance of responding to regional conflicts has focused Navy attention on littoral waters, with active sonar expected to be a favored mode of operation. Major performance drivers of such systems are the acoustic interactions with the ocean boundaries and fish. The vicinity of the air-sea interface is in particular a complex mix of scattering by surface roughness and scattering from bubble clouds and fish, coupled with boundary-interference effects. The Naval Research Laboratory has recently developed broadband, physics-based scattering strength models that both unify and advance our understanding of boundary scattering at low frequencies (< 5 kHz) by providing a physical basis for isolating scattering mechanisms. In this paper, these models are used to assess both the sensitivity of scattering strength to environmental variables and their utility as tools for estimating these variables. These efforts are supported by a series of data-model comparisons that demonstrate both the environmental variability of acoustic response with frequency and scattering angle, and the importance of using physics-based tools to predict these responses.

Measurements of the spectral characteristics of low‐frequency, low‐grazing‐angle surface reverberation

Direct‐path measurements of low‐frequency (200–1000 Hz) and low‐grazing‐angle (≤10 deg) acoustic surface scattering were made in the Gulf of Alaska in April of 1990 and in March of 1992. Short‐duration (0.6–2.4 s) cw and PRN waveforms were used to quantify the spectral character of surface/near‐surface reverberation as a function of frequency and environmental conditions. Measures include spreading and peak‐Doppler‐shift statistics versus azimuth. Results over the range of wind speeds (10–35 kns) have revealed a dominant zero‐Doppler component and a weaker‐than‐expected dependence of spread on both sea conditions (wind speed and relative direction of the seas) and frequency. The results are consistent with sub‐surface bubbles as the primary mechanism for surface reverberation when white caps are present, and give insight as to the nature of the bubble clouds that could give rise to the observed acoustic scattering.

Clutter statistics of long-range wideband echoes from fish aggregations off the Oregon coast

Echoes from fish can be the dominant source of reverberation over a range of important sonar frequencies and grazing angles. Moreover, fish echoes from broadband signals often retain coherent structure (generate clutter) after undergoing normalized match-filter processing. Coupled with their inherent spatiotemporal variability, fish can thus be a significant clutter problem for active sonars. Using a towed source and horizontal line-array receiver, measurements of mid-frequency (1.5-11 kHz) backscattering from aggregations of fish were made from the R/V New Horizon in five shallow-water and shelf-break areas off the coast of Oregon (Astoria Canyon to Heceta Bank) during July and August 2012. The experiment and the frequency-dependent echo statistics in relation to the observed distribution and behavior of the two primary resident fish species (Pacific hake and Pacific sardines) are discussed. For example, the short-time echo variability and spatial patchiness were characteristic of the mid-water (hake) an...

Spatiotemporal variability of clutter due to fish aggregations in the vicinity of Heceta Bank

Acoustic interactions with fish can be a significant source of clutter to mid-frequency (MF; 1–10 kHz) active sonars. To develop robust schemes for reducing sonar false alarm rates, it is thus important to accurately characterize the spatiotemporal nature of MF echoes from fish. In the summer of 2012, long-range MF measurements of coherent backscattering from fish aggregations were made at several sites in the vicinity of Heceta Bank off the coast of Oregon. The dominant fish species included Pacific sardines that typically school near the surface and Pacific hake that typically are above the bottom in loose aggregations or layers, and so represent spatially different echo classes. Measured spatiotemporal fish echo statistics are discussed in terms of the observed in-situ distribution and behavior of these two species. NRLs moment-based Poisson-Rayleigh clutter model is used to physically relate the measured amplitude distributions to scatterer attributes (spatial density and dispersion). The temporal pe...

Observations of region-specific fish behavior using long- and short-range broadband (1.5—6 + kHz) active acoustic systems

Two broadband active acoustic systems, in concert with traditional narrowband systems and nets, were used to study distributions of fish in three regions within the Gulf of Maine. The long-range multi-beam broadband system detected fish out to 15 km range and the downward-looking short-range broadband system detected fish throughout the water column close behind the ship. The multi-year (2007–2011) study revealed distinct spatial patterns of fish and corresponding echo statistics in each region—diffusely distributed, sparsely distributed compact patches, and long (continuous) shoals. The broadband capabilities of the sonar systems (each spanning 1.5—6 + kHz) uniquely allow observations of resonance phenomena of the local swimbladder-bearing fish. The observed resonances were consistent with the fish species, sizes, and depths that were concurrently sampled in each area from a second research vessel. Spectral peak analysis also interestingly revealed the presence of distinct modes, which may be useful indi...

A physical statistical clutter model for active sonar scenarios with variable signal-to-noise ratios

Active sonar classification algorithms need to be robust in preventing operator overload while not being misled by false targets. This talk describes a new 3-parameter statistical sonar clutter model that not only provides a physical context for relating the characteristics of normalized matched-filter echo-data distributions to scatterer attributes, but scatterer information that is largely independent of its peak signal-to-noise ratio (SNR) value. It extends our 2-parameter Poisson-Rayleigh model (Fialkowski and Gauss, IEEE JOE, 2010) by adding a quantitative measure of scatterer spatial dispersion to its measures of scatterer density and relative strength. Maximum likelihood estimates of the clutter model’s 3 parameters were derived from mid-frequency (1-5 kHz) shallow-water active sonar data containing returns from biologic, geologic and anthropogenic objects with differing spatial and scattering characteristics. The resulting clutter model’s probability density functions not only fit the non-Rayleigh...

A moment-based technique to palliate pervasive clutter while preserving objects of interest

High false alarm rates are a persistent problem for ASW active sonars. This talk describes a recently developed sonar clutter characterization and control method, the “Poisson-Rayleigh Inspired Method” (PRIM). The closed-form PRIM is based on NRLs two-parameter Poisson-Rayleigh (P-R) model that, like the popular K-type model, provides a physical context for relating the characteristics of data distributions to scatterer attributes (density and relative strength). However, with its extra degree of freedom, the P-R model offers the potential to exploit more information through higher-order (4th and 6th) data moments, and thus do a better job of characterizing clutter. The technique is demonstrated on normalized, broadband sonar clutter data collected in two range-dependent shallow-water environments, one bottom dominated and one fish dominated. The data results suggest that in contrast to the popular K-distribution, the power of the discrete-scatterer component in the P-R distribution can provide feature i...