Performance study of ray-based ocean acoustic tomography methods for estimating submesoscale variability in the upper ocean.

Ocean acoustic tomography (OAT) methods aim at estimating variations of sound speed profiles (SSP) based on acoustic measurements between multiple source-receiver pairs (e.g., eigenray travel times). This study investigates the estimation of range-dependent SSPs in the upper ocean over short ranges (<5 km) using the classical ray-based OAT formulation as well as iterative or adaptive OAT formulations (i.e., when the sources and receivers configuration can evolve across successive iterations of this inverse problem). A regional ocean circulation model for the DeSoto Canyon in the Gulf of Mexico is used to simulate three-dimensional sound speed variations spanning a month-long period, which exhibits significant submesoscale variability of variable intensity. OAT performance is investigated in this simulated environment in terms of (1) the selected source-receivers configuration and effective ray coverage, (2) the selected OAT estimator formulations, linearized forward model accuracy, and the parameterization of the expected SSP variability in terms of empirical orthogonal functions, and (3) the duration over which the OAT inversion is performed. Practical implications for the design of future OAT experiments for monitoring submesoscale variability in the upper ocean with moving autonomous platforms are discussed.

Self-localization of mobile underwater vector sensor platforms using a source of opportunity.

Using a network of a few compact mobile underwater platforms, each equipped with a single acoustic sensor, as a distributed sensing array is attractive but requires precise positioning of each mobile sensor. However, traditional accurate underwater positioning tools rely on active acoustic sources (e.g., acoustic pingers), which implies additional hardware and operational complexity. Hence, self-localization (i.e., totally passive) methods using only acoustic sources of opportunity (such as surface vessels) for locating the mobile sensors of a distributed array appear as a simpler alternative. Existing underwater self-localization methods have mainly been developed for mobile platforms equipped with time-synchronized hydrophones and rely only on the time-differences of arrival between multiple pairwise combinations of the mobile hydrophones as inputs for a complex non-linear inversion procedure. Instead, this article introduces a self-localization method, which uses a linear least-square formulation, for two mobile time-synchronized vector sensor platforms based on their acoustic recordings of a distant surface vessel and their inertial navigation system (INS) measurements. This method can be generalized to multiple vector sensor pairs to provide additional robustness toward input parameter errors (e.g., due to a faulty INS) as demonstrated experimentally using drifting buoys with inertial vector sensors deployed ∼100 m apart in shallow water.

Machine learning approaches for ray-based ocean acoustic tomography.

Underwater sound propagation is primarily driven by a nonlinear forward model relating variability of the ocean sound speed profile (SSP) to the acoustic observations (e.g., eigenray arrival times). Ocean acoustic tomography (OAT) methods aim at reconstructing SSP variations (with respect to a reference environment) from changes of the acoustic measurements between multiple source-receiver pairs. This article investigates the performance of three different OAT methods: (1) model-based methods (i.e., classical ray-based OAT using a linearized forward model), (2) data-driven methods (such as deep learning) to directly learn the inverse model, and (3) a hybrid solution [i.e., the neural adjoint (NA) method], which combines deep learning of the forward model with a standard recursive optimization to estimate SSPs. Additionally, synthetic SSPs were generated to augment the variability of the training set. These methods were tested with modeled ray arrivals calculated for a downward refracting environment with mild fluctuations of the thermocline. Idealized towed and fixed source configurations are considered. Results indicate that merging data-driven and model-based methods can benefit OAT predictions depending on the selected sensing configurations and actual ray coverage of the water column. But ultimately, the robustness of OAT predictions depends on the dynamics of the SSP variations.

Passive underwater acoustic identification tags using multi-layered shells.

The development of pre-deployed underwater infrastructures to aid in autonomous underwater vehicle (AUV) navigation is of keen interest, with the increased use of AUVs for undersea operations. Previous literature has introduced a class of passive underwater acoustic markers, termed acoustic identification (AID) tags [Satish, Trivett, and Sabra, J. Acoust. Soc. Am. 147(6), EL517-EL522 (2020)], which are inexpensive to construct, simple to deploy, and reflect unique engineered acoustic signatures that can be detected by an AUV instrumented with high-frequency sonar systems. An AID tag is built of multi-layer shells with different acoustic properties and thicknesses to generate a unique acoustic signature, composed of the multiple reflections created by the layer interfaces, thus akin to an "acoustic barcode." AID tags can be used as geospatial markers to highlight checkpoints in AUV trajectories or mark areas of interest underwater. This article investigates the optimization of the AID tag's design using energy based metrics and evaluates the detectability of an AID tag in the presence of interfering signals, such as clutter using matched-filter based techniques. Furthermore, experimental results of AID tags interrogated by a standard high-frequency sonar are presented to provide proof of concept of AID tag detection in a reverberant water tank.

Omnidirectional passive acoustic identification tags for underwater navigation.

A class of passive acoustic identification (AID) tags with curved symmetry for underwater navigation is presented. These AID tags are composed of radially stratified shells designed to backscatter a unique specular reflection pattern independent of the incidence orientation in a monostatic configuration, thus acting as acoustic bar-codes. The AID tag's response can be uniquely engineered by selecting the thicknesses and material properties of the individual constitutive shells. Furthermore, in the high-frequency regime, the specular component of the AID tag's response can be simply predicted numerically assuming horizontally stratified layers. This approach is demonstrated using scaled experiments with an AID tag constructed from 3D printed hemispherical shells.

Analysis of the ray-based blind deconvolution algorithm for shipping sources.

The ray-based blind deconvolution (RBD) technique for ocean waveguides estimates both the unknown waveform radiated by some source of opportunity and the channel impulse response (CIR) between the source and the receiving elements of an array of hydrophones using only measured signals, knowledge of the array geometry, and the local sound speed. Previous studies have investigated the applicability of this method for shipping sources in a shallow, nearly range-independent waveguide (∼200 m depth), but using a limited set of shipping vessels (typically only the research vessel itself) and operating within a small domain of RBD processing parameters (e.g., integration time and frequency band). This study systematically investigates the performance of the RBD method for estimating the CIR for a large set of shipping vessels recorded on short aperture, bottom-mounted, vertical arrays deployed in the Santa Barbara channel across different frequency bands and integration times, and also in comparison to CIR measured using active sources. Furthermore, the influence of the source motion on the RBD algorithm is quantified both numerically and experimentally.

Geoacoustic inversion using ray-based blind deconvolution of shipping sources.

The ray-based blind deconvolution algorithm can provide an estimate of the channel impulse responses (CIRs) between a shipping source of opportunity and the elements of a receiving array by estimating the unknown phase of this random source through wideband beamforming along a well-resolved ray path. However, due to the shallow effective depth (typically <10 m) and low frequency content (typically less than a few kHz) associated with shipping sources, the interfering direct and surface arriving pair and subsequent bottom and surface-bottom arrival pair cannot always be resolved in the CIR arrival-time structure. Nevertheless, this study demonstrates that the bottom reflection loss can be inferred from the ratio of the magnitude spectra of these two arrival pairs if a frequency-dependent correction (which can be purely data based) is applied to correct for the dipole source effect. The feasibility of the proposed approach is demonstrated to invert for the geoacoustic parameters of a soft-layer covering the ocean floor using a nonlinear least-square algorithm.

Passive underwater acoustic tags using layered media.

Autonomous Underwater Vehicle (AUV) navigation requires accurate positioning information from the environment. Existing underwater navigation paradigms employ active acoustic transponders that assist in this task, but these more complex and costly systems require maintenance and power. This paper presents instead a passive underwater marker made of different horizontally stacked acoustically reflective materials that is cost effective and relatively simple to service. A marker's characteristic acoustic signature can be detected by AUVs as acoustic backscattering upon tag insonification, and hence be used for navigation purposes.

Enhancing ambient noise correlation processing using vector sensors.

Ambient noise cross-correlations between separated sensors can yield estimates of the Green's function between them. Vector sensors (which record both pressure and acoustic velocity vector components) can leverage their directionality to reject ambient noise sources that do not contribute to the emergence of the Green's function, thus improving performance over standard omnidirectional hydrophones. To quantify this performance gain, a time-domain analytical expression for the correlation between each component of a vector sensor in the presence of an isotropic ambient noise field is derived. Improvement of the velocity channel correlations relative to pressure channel correlations is examined for varying bandwidth, sensor separation distance, and additive channel noise levels. Last, the experimentally measured reduction in variance for the velocity channels correlations vs pressure correlations, using drifting vector sensors deployed in the Long Island Sound, were found to be comparable to the theoretical prediction. Overall, both theoretical and experimental results indicate modest gains are obtained when extracting the Green's function from velocity correlations over using pressure correlations. Thus, vector sensors can be used to reduce the required averaging time for this noise correlation processing, which may be especially useful, for instance, in a fluctuating environment or for drifting sensors.

Ray-based blind deconvolution of shipping sources using multiple beams separated by alternating projection.

This article presents a method for improving the performance of the ray-based blind deconvolution (RBD) algorithm, which was first proposed by Sabra, Song, and Dowling [J. Acoust. Soc. Am. 127(2), EL42-EL47 (2010)]. In order to retrieve the channel impulse response (CIR), the original RBD algorithm uses the source signal phase from a selected single beam output. However, when the impinging multipath signals have low coherence, the channel estimate from a selected beam may not show all paths correctly. In this research, the maximum likelihood estimator, which is called the alternating projection, is applied to separate multipath signals. Then the multiple CIRs obtained from those separated signals are coherently combined. This results in more robust detection of existing multipaths. The performance of the proposed method is verified using Noise09 sea experiment data, where the proposed method better resolves the multipath arrival structure.

Quantifying the influence of respiration and cardiac pulsations on cerebrospinal fluid dynamics using real-time phase-contrast MRI.

PURPOSE: To validate a real-time phase contrast magnetic resonance imaging (RT-PCMRI) sequence in a controlled phantom model, and to quantify the relative contributions of respiration and cardiac pulsations on cerebrospinal fluid (CSF) velocity at the level of the foramen magnum (FM). MATERIALS AND METHODS: To validate the 3T MRI techniques, in vitro studies used a realistic model of the spinal subarachnoid space driven by pulsatile flow waveforms mimicking the respiratory and cardiac components of CSF flow. Subsequently, CSF flow was measured continuously during 1-minute RT-PCMRI acquisitions at the FM while healthy subjects (N = 20) performed natural breathing, deep breathing, breath-holding, and coughing. Conventional cardiac-gated PCMRI was obtained for comparison. A frequency domain power ratio analysis determined the relative contribution of respiration versus cardiac ([r/c]) components of CSF velocity. RESULTS: In vitro studies demonstrating the accuracy of RT-PCMRI within 5% of input values showed that conventional PCMRI measures only the cardiac component of CSF velocity (0.42 ± 0.02 cm/s), averages out respiratory effects, and underestimates the magnitude of CSF velocity (0.96 ± 0.07 cm/s). In vivo RT-PCMRI measurements indicated the ratio of respiratory to cardiac velocity pulsations averaged over all subjects as [r/c = 0.14 ± 0.27] and [r/c = 0.40 ± 0.47] for natural and deep breathing, respectively. During coughing, the peak CSF velocity increased by a factor of 2.27 ± 1.40. CONCLUSION: RT-PCMRI can noninvasively measure instantaneous CSF velocity driven by cardiac pulsations, respiration, and coughing in real time. A comparable contribution of respiration and cardiac pulsations on CSF velocity was found during deep breathing but not during natural breathing. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 1 J. MAGN. RESON. IMAGING 2017;46:431-439.

Passive underwater acoustic markers using Bragg backscattering.

This letter demonstrates the feasibility of a passive underwater acoustic marker technology (or "AcoustiCode") for use in underwater navigation. An AcoustiCode tag is a planar surface with machined periodic patterns capable of producing Bragg backscattering beampatterns with engineered spatial and frequency variations, thus having a unique three-dimensional acoustic signature over a selected frequency band. Hence, these AcoustiCodes enable three-dimensional navigation and information signaling in a totally passive manner for existing high-frequency SONAR systems (potentially mounted on autonomous underwater vehicles), which naturally operate in a narrow frequency band and can also be used over significantly longer ranges compared to optically-based systems.

Blind deconvolution of shipping sources in an ocean waveguide.

This paper investigates the applicability of a ray-based blind deconvolution (RBD) method for underwater acoustic sources of opportunity such as ships recorded on a receiver array. The RBD relies on first estimating the unknown phase of the random source by beamforming along a well-resolved ray path, and then matched-filtering each received signal using the knowledge of this random phase to estimate the full channel impulse responses (CIRs) between the unknown source and the array elements (up to an arbitrary time-shift) as well as recovering the radiated signal by the random source. The performance of this RBD is investigated using both numerical simulation and experimental recordings of shipping noise in the frequency band [300-800 Hz] for ranges up to several kilometers. The ray amplitudes of the estimated CIRs are shown to be consistent with known bottom properties in the area. Furthermore, CIRs obtained for an arbitrarily selected shipping track are used as data-derived replicas to perform broadband matched-field processing to locate another shipping source recorded at a later time in the vicinity of the selected track.

Acoustic scattering from phononic crystals with complex geometry.

This work introduces a formalism for computing external acoustic scattering from phononic crystals (PCs) with arbitrary exterior shape using a Bloch wave expansion technique coupled with the Helmholtz-Kirchhoff integral (HKI). Similar to a Kirchhoff approximation, a geometrically complex PC's surface is broken into a set of facets in which the scattering from each facet is calculated as if it was a semi-infinite plane interface in the short wavelength limit. When excited by incident radiation, these facets introduce wave modes into the interior of the PC. Incorporation of these modes in the HKI, summed over all facets, then determines the externally scattered acoustic field. In particular, for frequencies in a complete bandgap (the usual operating frequency regime of many PC-based devices and the requisite operating regime of the presented theory), no need exists to solve for internal reflections from oppositely facing edges and, thus, the total scattered field can be computed without the need to consider internal multiple scattering. Several numerical examples are provided to verify the presented approach. Both harmonic and transient results are considered for spherical and bean-shaped PCs, each containing over 100 000 inclusions. This facet formalism is validated by comparison to an existing self-consistent scattering technique.

Cross-coherent vector sensor processing for spatially distributed glider networks.

Autonomous underwater gliders fitted with vector sensors can be used as a spatially distributed sensor array to passively locate underwater sources. However, to date, the positional accuracy required for robust array processing (especially coherent processing) is not achievable using dead-reckoning while the gliders remain submerged. To obtain such accuracy, the gliders can be temporarily surfaced to allow for global positioning system contact, but the acoustically active sea surface introduces locally additional sensor noise. This letter demonstrates that cross-coherent array processing, which inherently mitigates the effects of local noise, outperforms traditional incoherent processing source localization methods for this spatially distributed vector sensor network.

Variability of the coherent arrivals extracted from low-frequency deep-ocean ambient noise correlations.

Correlation processing of ocean noise can be used to develop totally passive ocean monitoring methods. Using various hydrophone pair orientations, this study investigates the frequency dependence, seasonal variability, and emergence rate of coherent arrivals from cross-correlations of low frequency ambient noise (f < 40 Hz) recorded on triangular hydrophones arrays. These arrays are located at five existing hydroacoustic stations of the International Monitoring System (IMS), situated in the deep-sound channel, and distributed across the Atlantic, Pacific, and Indian Ocean basins. For the majority of studied sites, persistent and fast-emerging coherent arrivals are reliably obtained if the axis connecting the selected hydrophone pair has a direct line-of-sight with regions of the globe containing stable and diffuse noise sources (e.g., polar-ice or seismic noise). Furthermore, for this favorable orientation, the emergence rate of coherent arrivals extracted between hydrophone pairs separated by long ranges (here ∼130 km) can be approximated based on measurements made between hydrophone pairs separated by short ranges (∼2 km) in the Atlantic Ocean. Hence, results from this study, obtained using existing hydrophone configurations of the IMS hydroacoustic stations, could be used to guide the placement of other hydrophone arrays over the globe for future long-range passive ocean monitoring experiments.

A three-dimensional Bloch wave expansion to determine external scattering from finite phononic crystals.

External scattering from a finite phononic crystal (PC) is studied using the Helmholtz-Kirchhoff integral theorem integrated with a Bloch wave expansion (BWE). The BWE technique is used to describe the internal pressure field of a semi-infinite or layered PC subject to an incident monochromatic plane wave. Following the BWE solution, the Helmholtz-Kirchhoff integral is used to determine the external scattered field. For cubic PCs, the scattered results are compared to numerical treatments in both the frequency and time domain. The presented approach is expected to be valid when the PC size is larger than the acoustic wavelength. However, very good agreement in the spatial beam pattern is also documented for both large and small (with respect to the wavelength) PCs. The result of this work is a fully-analytical, efficient, and verified approach for accurately predicting external scattering from finite, three-dimensional PCs.

Optimized extraction of coherent arrivals from ambient noise correlations in a rapidly fluctuating medium.

Ambient noise correlations can be used to estimate Green's functions for passive monitoring purposes. However, this method traditionally relies on sufficient time-averaging of the noise-correlations to extract coherent arrivals (i.e., Green's function estimates), and is thus limited by rapid environmental fluctuations occurring on short time scales while the averaging takes place. This letter demonstrates with simulation and data that the use of a stochastic search algorithm to correct and track these rapid environmental fluctuations can significantly reduce the required averaging time to extract coherent arrivals from noise correlations in a fluctuating medium.

Assessment of muscle stiffness using a continuously scanning laser-Doppler vibrometer.

INTRODUCTION: A stand-alone and low-cost elastography technique has been developed using a single continuously scanning laser Doppler vibrometer. METHODS: This elastography technique is used to measure the propagation velocity of surface vibrations over superficial skeletal muscles to assess muscle stiffness. RESULTS: Systematic variations in propagation velocity depending on the contraction level and joint position of the biceps brachii were demonstrated in 10 subjects. CONCLUSIONS: This technique may assist clinicians in characterizing muscle stiffness (or tone) changes due to neuromuscular disorders.

Subspace array processing using spatial time-frequency distributions: applications for denoising structural echoes of elastic targets.

Structural echoes of underwater elastic targets, used for detection and classification purposes, can be highly localized in the time-frequency domain and can be aspect-dependent. Hence such structural echoes recorded along a distributed (synthetic) aperture, e.g., using a moving receiver platform, would not meet the stationarity and multiple snapshots requirements of common subspace array processing methods used for denoising array data based on their estimated covariance matrix. To address this issue, this article introduces a subspace array processing method based on the space-time-frequency distribution (STFD) of single-snapshots of non-stationary signals. This STFD is obtained by computing Cohen's class time-frequency distributions between all pairwise combination of the recorded signals along an arbitrary aperture array. This STFD is interpreted as a generalized array covariance matrix which automatically accounts for the inherent coherence across the time-frequency plane of the received nonstationary echoes emanating from the same target. Hence, identifying the signal's subspace from the eigenstructure of this STFD provides a means for denoising these non-stationary structural echoes by spreading the clutter and noise power in the time-frequency domain; as demonstrated here numerically and experimentally using the structural echoes of a thin steel spherical shell measured along a synthetic aperture.