Carey M. Rappaport, Edwin Marengo, Timothy Duda
Date of Award
Doctor of Philosophy
Department or Academic Unit
College of Engineering. Department of Computer and Electrical Engineering
engineering, electronics and electrical, matched filter, modal dispersion, multiple scattering, population density imaging
Electrical and Electronics
Underwater acoustic imaging and surveillance systems often transmit broadband pulses and process the scattered returns with a correlation matched filter in order to achieve high spatial resolution in range localization. When imaging distributed groups of scatterers, such as schools or shoals of fish, bubble clouds and swarms of AUVs, large numbers of individuals, ranging from tens to hundreds of thousands, are often present within each resolution cell of the imaging system. To estimate population density, spatial distribution, scattering cross-section and other parameters of the scatterer group and their error bounds, it is essential to understand the statistical nature and composition of the broadband matched filtered scattered returns. Here, using a combination of theoretical model development and experimental data analysis, we investigate (1) the statistics of broadband acoustic propagation in temporally and spatially varying shallow water waveguides, (2) the statistics and mechanisms for scattering from distributed groups, and (3) the effects of multiple scattering and waveguide dispersion in predicting the level and duration of the broadband matched filtered scattered intensities. The analysis is applied to two types of underwater acoustic imaging systems for mapping fish distributions. (a) The ocean acoustic waveguide remote sensing (OAWRS) system that allows instantaneous wide area continental-shelf scale imaging by utilizing muti-modal waveguide propagation, and (b) the conventional ultrasonic echosounder using direct-path propagation to image scatterer groups over water depth.
Using experimental data acquired in the US east coast continental shelf, the scintillation statistics of one-way propagated broadband waveforms are quantified as a function of signal bandwidth, center frequency, and range. The received broadband intensity is shown to follow the Gamma distribution implying the central limit theorem has led to a fully saturated Gaussian field. An efficient numerical approach is developed to estimate the broadband transmission loss in a fluctuating, range-dependent ocean waveguide using spatial averaging of a time-harmonic stochastic propagation model. This approach is essential for rapidly detrending OAWRS imagery to make real time population density estimates.
Theoretical models are then developed to determine the statistical moments of the broadband matched filtered scattered returns from distributed groups. For the ultrasonic echosounder, a numerical Monte-Carlo model that includes multiple scattering is developed and implemented to determine the conditions for when multiple scattering is significant. The model is implemented for Atlantic herring schools found in the Gulf of Maine, imaged with a conventional fisheries echosounder. Our analysis indicates that the single-scatter approximation is valid even for dense Atlantic herring schools at ultrasound frequencies. For direct-path imaging systems, an analytic model is also developed to efficiently estimate the scattered field statistics from the group by applying a single scatter approximation. The model can be used to compare the levels of the coherently and incoherently scattered intensities. It provides an explanation for how high spatial resolution is acheived with incoherently scattered fields with the coherent matched filter.
We examine the effects of multiple scattering, attenuation, and waveguide dispersion on population density imaging with the OAWRS imaging system by implementing a numerical Monte-Carlo model. The model determines the statistical moments of the matched filtered multiply scattered returns from a distributed group of discrete scatterers, and incorporates propagation through a fluctuating range-dependent ocean environment. Results of the model are applied to interpret OAWRS imagery of shoaling Atlantic herring populations acquired during the NOPP-funded 2006 Experiment in the Gulf of Maine. Our analysis indicates that multiple scattering and attenuation were mostly negligible at OAWRS frequencies employed and for herring densities observed.
Andrews, Mark, "Enabling high resolution population density imaging of random scatterer groups in a fluctuating range-dependent ocean waveguide with the matched filter variance" (2009). Electrical Engineering Dissertations. Paper 24. http://hdl.handle.net/2047/d20000666
Click button above to open, or right-click to save.