Stephen D. Lynch

Active control of passive acoustic fields: Passive synthetic aperture/Doppler beamforming with data from an autonomous vehicle

The maneuverability of autonomous underwater vehicles (AUVs) equipped with hull-mounted arrays provides the opportunity to actively modify received acoustic fields to optimize extraction of information. This paper uses ocean acoustic data collected by an AUV-mounted two-dimensional hydrophone array, with overall dimension one-tenth wavelength at 200-500 Hz, to demonstrate aspects of this control through vehicle motion. Source localization is performed using Doppler shifts measured at a set of receiver velocities by both single elements and a physical array. Results show that a source in the presence of a 10-dB higher-level interferer having exactly the same frequency content (as measured by a stationary receiver) is properly localized and that white-noise-constrained adaptive beamforming applied to the physical aperture data in combination with Doppler beamforming provides greater spatial resolution than physical-aperture-alone beamforming and significantly lower sidelobes than single element Doppler beamforming. A new broadband beamformer that adjusts for variations in vehicle velocity on a sample by sample basis is demonstrated with data collected during a high-acceleration maneuver. The importance of including the cost of energy expenditure in determining optimal vehicle motion is demonstrated through simulation, further illustrating how the vehicle characteristics are an integral part of the signal/array processing structure.

Estimating the horizontal and vertical direction-of-arrival of water-borne seismic signals in the northern Philippine Sea

Conventional and adaptive plane-wave beamforming with simultaneous recordings by large-aperture horizontal and vertical line arrays during the 2009 Philippine Sea Engineering Test (PhilSea09) reveal the rate of occurrence and the two-dimensional arrival structure of seismic phases that couple into the deep ocean. A ship-deployed, controlled acoustic source was used to evaluate performance of the horizontal array for a range of beamformer adaptiveness levels. Ninety T-phases from unique azimuths were recorded between Yeardays 107 to 119. T-phase azimuth and S-minus-P-phase time-of-arrival range estimates were validated using United States Geological Survey seismic monitoring network data. Analysis of phases from a seismic event that occurred on Yearday 112 near the east coast of Taiwan approximately 450 km from the arrays revealed a 22° clockwise evolution of T-phase azimuth over 90 s. Two hypotheses to explain such evolution-body wave excitation of multiple sources or in-water scattering-are presented based on T-phase origin sites at the intersection of azimuthal great circle paths and ridge/coastal bathymetry. Propagation timing between the source, scattering region, and array position suggests the mechanism behind the evolution involved scattering of the T-phase from the Ryukyu Ridge and a T-phase formation/scattering location estimation error of approximately 3.2 km.

Evidence of Doppler-shifted Bragg scattering in the vertical plane by ocean surface waves

A set of narrowband tones (280, 370, 535, and 695 Hz) were transmitted by an acoustic source mounted on the ocean floor in 10 m deep water and received by a 64-element hydrophone line array lying on the ocean bottom 1.25 km away. Beamformer output in the vertical plane for the received acoustic tones shows evidence of Doppler-shifted Bragg scattering of the transmitted acoustic signals by the ocean surface waves. The received, scattered signals show dependence on the ocean surface wave frequencies and wavenumber vectors, as well as on acoustic frequencies and acoustic mode wavenumbers. Sidebands in the beamformer output are offset in frequency by amounts corresponding to ocean surface wave frequencies. Deviations in vertical arrival angle from specular reflection agree with those predicted by the Bragg condition through first-order perturbation theory using measured directional surface wave spectra and acoustic modes measured by the horizontal hydrophone array.

Investigating sources of variability of the range and structure of the low frequency shallow convergence zone

During an experiment in the northern Philippine Sea in 2009, a ship towing Penn States Five-Octave Research Array (FORA) at approximately 120 m depth drove counter-clockwise in an arc, maintaining constant range at one convergence zone (CZ) from a second ship holding station with an acoustic source deployed at 15 and 60 m. In addition, the FORA was towed at various depths in a star pattern about the station-keeping source ship, thereby sampling the first CZ in range, depth, and azimuth. Throughout the experiment, sound speed profiles were measured using expendable bathy-thermographs, expendable sound velocimeters, and conductivity/temperature versus depth sensors, and detailed bathymetric data were collected using the multibeam systems aboard these and other ships. By incorporating this extensive environmental information into numerical models, variability observed in these measurements of the range and structure and asymmetry of the distribution of received levels of the first CZ resulting from a shallo...

Time-evolving T-phase arrival structure using simultaneous recordings by large-aperture horizontal and vertical line arrays in PhilSea09

A number of models have been proposed to explain the mechanisms by which seismic phases couple to the deep ocean sound channel in order to create water-borne acoustic tertiary (T) phases. Beamforming conducted on simultaneous recordings by large-aperture horizontal towed and vertical moored line arrays during PhilSea09 shows the temporal evolution of a T-phase arrival consistent with the down-slope modal conversion/propagation model. Towed array calibration is conducted using ship-deployed, controlled multi-tone acoustic sources. Conventional, minimum variance distortionless response, white noise constrained, and dominant mode rejection beamformers are compared in their ability to minimize bias and variance in estimating the azimuthal arrival directions of signals from both the controlled source and the seismic phases recorded by the horizontal array. Horizontal array beamformer-derived azimuth and time-of-arrival range estimates from P, S, and T-phase arrivals at towed and moored receivers indicate the e...

Deep‐water propagation measurements from a towed‐array.

Towed array measurements using the Office of Naval Research Five Octave Research Array (FORA) were made of narrowband and broadband transmissions as part of the Philippine Sea 2009 experiment (PhilSea09). Geometries between the tow‐ship and the receive ship were established to investigate the spatial and temporal structure of convergence zone (CZ) propagation, as well as short range propagation, dominated by a single bottom bounce path. Estimates of the CZ width as a function of geography and local bathymetry will be given as well as coherence scales for a constant range arc (within a CZ). For the bottom bounce path, significant bottom scattering (beam‐broadening) was observed and has been interpreted as out‐of‐plane scattering from the rough sub‐surface seafloor. Numerical modeling to begin to investigate the nature of 3‐D rough seafloor scattering work has been begun and will be presented. Implications for the detection processing for CZ and bottom‐bounce detection will be discussed.

Dependence of the structure of the shallow convergence zone on deep ocean bathymetry.

During an experiment in the northern Philippine Sea in 2009, low‐frequency tones were transmitted from a shallow (15‐ and 60‐m) source deployed from R/V Melville keeping station to a shallow (250‐m) horizontal receiver array towed by R/V Kilo Moana approximately one convergence zone (CZ) away. Recordings were made during events in which the receiver ship maintained constant range in the convergence zone and during events in which the receiver ship transited radially through the CZ. The shallow CZ exhibits strong dependence on the bathymetry mid‐way between the source and receiver array. In fact, the variability of the structure of the first CZ in this environment is significantly more strongly affected by the heterogeneous character of the bottom than water column fluctuations. Numerical modeling with a parabolic equation code is used to support the conclusions from the data analysis.

A simplified model of frequency distortion in ocean acoustic signals by multiple interactions with a moving ocean surface and its use in oceanographic inversions.

Received omnidirectional spectra of low frequency narrowband tones measured in a shallow ocean waveguide exhibit Doppler‐shifted sidebands at frequencies whose offsets are integer multiples of the dominant surface wave frequencies. These observations suggest that higher order scattering is important. However, higher‐order perturbation theory involving only a single surface interaction does not predict the observed relative sideband levels. In particular, the higher order sidebands were often observed to increase in relative level to the main spectral peak during smaller surface wave, and, therefore, smaller kh, conditions. A simplified model is presented that simulates the surface wave‐induced frequency distortions as path‐length modulation caused by the vertical motions of the surface. The model incorporates multiple interactions with the moving surface, and quantitatively predicts the effects on the received spectra of the spatial coherence of the surface waves along the propagation path. The model resu...

Underwater acoustic measurements of Doppler‐shifted Bragg scattering in the vertical plane from interaction with the ocean surface.

High resolution beamformer output in the vertical plane for received fields from narrowband acoustic tones transmitted at 280, 370, 535, and 695 Hz shows evidence of Doppler‐shifted Bragg scattering by the ocean surface waves. The sidebands in the beamformer output are offset in frequency by amounts corresponding to ocean surface wave frequencies, while deviations in vertical arrival angle agree with those predicted by the Bragg condition through first‐order perturbation theory using measured directional surface wave spectra and measured acoustic mode wave numbers. The acoustic measurements were made by a horizontal hydrophone array on the ocean bottom in a shallow water waveguide under a variety of ocean surface conditions. [Work supported by ONR, Code 321(US).]

Temporal variability of narrow‐band tones in very shallow coastal waveguides

During an experiment near San Clemente, CA, a source moored on the bottom just outside the surf zone (500 m offshore, approximately 9‐m depth) transmitted eight tones between 70 and 700 Hz to two nearly perpendicular, 64‐element bottom‐lying hydrophone arrays located approximately 1.2 km downcoast and 1 km further offshore (approximately 12‐m depth). The acoustic tones were transmitted for 5 min every half hour over a 22‐h period. Surface wave data obtained from a depth sensor array located approximately 40 km upcoast, as well as data from a PUV sensor data deployed during the experiment, are used to compute concurrent directional surface wave spectra. Long fast Fourier transforms of the acoustic recordings reveal temporal and spatial variability of the width and shape of the received tones in temporal frequency space. Surface‐wave‐related mechanisms of acoustic frequency distortion are proposed and their relative significance is investigated through the comparison of the data with results from physics‐ba...