Search

Extensive ocean noise records have kurtoses markedly different from the Gaussian distribution and therefore exhibit non-Gaussianity, which influences the performance of many sonar signal processing methods. To model the amplitude distribution, this paper studies a Bayesian Gaussian mixture model (BGMM) and its associated learning algorithm, which exploits the variational inference method. The most compelling feature of the BGMM is that it automatically selects a suitable number of effective components and then can approximate a sophisticated distribution in practical applications. The probability density functions (PDFs) of three types of noise in different frequency bands collected in the South China Sea-ambient noise, ship noise, and typhoon noise-are modeled and the goodness of fit is examined by applying the one-sample Kolmogorov-Smirnov test. The results demonstrate that: (i) Ambient noise in the low-frequency band may be slightly non-Gaussian, ship noise in each considered band is apparently non-Gaussian, and typhoons affect the noise in the low-frequency band to make it apparently non-Gaussian, while the noise in the high-frequency band is less affected and appears to be Gaussian. (ii) BGMM has higher goodness of fit than the Gaussian or Gaussian mixture model. (iii) In the non-Gaussian case, despite some components having small mixing coefficients, they are of great significance for describing the PDF.

When a beam emitted from an active monostatic sensor system sweeps across a volume, the echoes from scatterers present will fluctuate from ping to ping due to various interference phenomena and statistical processes. Observations of these fluctuations can be used, in combination with models, to infer properties of the scatterers such as numerical density. Modeling the fluctuations can also help predict system performance and associated uncertainties in expected echoes. This tutorial focuses on "physics-based statistics," which is a predictive form of modeling the fluctuations. The modeling is based principally on the physics of the scattering by individual scatterers, addition of echoes from randomized multiple scatterers, system effects involving the beampattern and signal type, and signal theory including matched filter processing. Some consideration is also given to environment-specific effects such as the presence of boundaries and heterogeneities in the medium. Although the modeling was inspired by applications of sonar in the field of underwater acoustics, the material is presented in a general form, and involving only scalar fields. Therefore, it is broadly applicable to other areas such as medical ultrasound, non-destructive acoustic testing, in-air acoustics, as well as radar and lasers.

Explainable artificial intelligence is gaining wider traction within the machine learning community and its application domains at large for its inherent motivation to explain and interpret large-scale machine interpretation of experimental datasets. We will present cognitive sampling as new way to implement explainable AI for acoustical signal processing. In particular, methods based on geometric signal processing and sparse sensing will be harnessed with machine cognition to interpret, classify and predict information autonomously from large-scale acoustic datasets spanning a wide variety of applications. We will compare the difference in performance between traditional supervised and semi-supervised learning architectures such as deep learning, and ensemble approaches, unsupervised learning networks. We will also present preliminary research on implementing cognitive sampling in machine-directed inverse problem-solving techniques such as autoencoders. The end goal is to discover efficient data encodings that enable hitherto unforeseen feature spaces using optimal or close to optimal sampling strategies. Specific applications will include a variety of acoustical environmental sensing applications involving spectral feature generation and interpretation such as sonar signal processing, undersea multipath channel sensing, as well as feature extraction in complex melodic structures in Indian classical music. [Work funded partially by ONR under Grants N00014-19-1-2436, N00014-19-1-2609, N00174-20-1-0016 and N00014-20-1-2626.]

Multipath propagation poses many challenges for broadband sonar signal processing in underwater acoustic channels. The waveguide invariant (WI) provides a useful relation between the temporal and frequency fading characteristics in shallow-water channels that can been exploited to improve broadband target detection in both active and passive sonars. In active sonar, the WI motivates an alternative to conventional constant-false-alarm rate (CFAR) normalization of the background by averaging over observed time-frequency striations to estimate the reverberation level. A WI-based generalized likelihood ratio test is reviewed and shown to outperform conventional cell-averaged and frequency-invariant CFAR detectors in both simulation and with real Mediterranean data. In passive towed-array sonar, targets near endfire are often masked by angularly spread broadband multipath interference whose statistics must be estimated with limited training data for effective suppression. A WI relationship characterizing the frequency-dependence of wavenumber differences permits averaging of narrowband array snapshots to form a low-rank broadband “focused” covariance matrix. Simulation experiments in a realistic ocean environment are reviewed showing that adaptive beamforming using WI focused covariance matrices can provide a significant array gain improvement over conventional adaptive methods with limited observation time. Finally, the potential for waveguide invariant-based blind source separation is discussed. [Work supported by ONR.]

The biosonar signals of two free-swimming Atlantic bottlenose dolphins performing a complex sonar search for a bottom target in San Diego Bay were compared with the biosonar signals of a dolphin performing a target discrimination task in a net pen in the same bay. A bite-plate device carried by the free-swimming dolphins supported a hydrophone that extended directly in front of the dolphin. A biosonar measuring tool attached to the bite plate measured the outgoing biosonar signals while the dolphins conducted sonar searches. Each of the free-swimming dolphins used different biosonar search strategy in solving the problem and the dolphins' biosonar signals reflect the difference in strategy. The dolphin in the pen stationed in a hoop while echolocating on a target 6 m away and reported if the indentation on a spherical target was directed toward it. The signals were parameterized by determining the peak-to-peak source levels, source energy flux density, peak frequency, center frequency, root-mean-square (rms) bandwidth, rms duration, and the Q of the signals. Some parameters were similar for the free-swimming and stationary dolphins while some were significantly different, suggesting biosonar signals used by free-swimming animals may be different than signals used by dolphins in a pen.

Visual displays in passive sonar based on the Fourier spectrogram are underpinned by detection models that rely on signal and noise power statistics. Time-frequency representations specialised for sparse signals achieve a sharper signal representation, either by reassigning signal energy based on temporal structure or by conveying temporal structure directly. However, temporal representations involve nonlinear transformations that make it difficult to reason about how they respond to additive noise. This article analyses the effect of noise on temporal fine structure measurements such as zero crossings and instantaneous frequency. Detectors that rely on zero crossing intervals, intervals and peak amplitudes, and instantaneous frequency measurements are developed, and evaluated for the detection of a sinusoid in Gaussian noise, using the power detector as a baseline. Detectors that rely on fine structure outperform the power detector under certain circumstances; and detectors that rely on both fine structure and power measurements are superior. Reassigned spectrograms assume that the statistics used to reassign energy are reliable, but the derivation of the fine structure detectors indicates the opposite. The article closes by proposing and demonstrating the concept of a doubly reassigned spectrogram, wherein temporal measurements are reassigned according to a statistical model of the noise background.

Active sonar systems are used to detect underwater man-made objects of interest (targets) that are too quiet to be reliably detected with passive sonar. Performance of active sonar can be degraded by false alarms caused by echoes returned from geological seabed structures (clutter) in shallow regions. To reduce false alarms, a method of distinguishing target echoes from clutter echoes is required. Research has demonstrated that perceptual-based signal features similar to those employed in the human auditory system can be used to automatically discriminate between target and clutter echoes, thereby reducing the number of false alarms and improving sonar performance. An active sonar experiment on the Malta Plateau in the Mediterranean Sea was conducted during the Clutter07 sea trial and repeated during the Clutter09 sea trial. The dataset consists of more than 95,000 pulse-compressed echoes returned from two targets and many geological clutter objects. These echoes were processed using an automatic classifier that quantifies the timbre of each echo using a number of perceptual signal features. Using echoes from 2007, the aural classifier was trained to establish a boundary between targets and clutter in the feature space. Temporal robustness was then investigated by testing the classifier on echoes from the 2009 experiment.

Approaches for instantaneous passive source localization using a towed horizontal receiver array in a random range-dependent ocean waveguide are examined. They include: (1) Moving array triangulation, (2) array invariant, (3) bearings-only target motion analysis in modified polar coordinates via the extended Kalman filter, and (4) bearings-migration minimum mean-square error. These methods are applied to localize and track a vertical source array deployed in the far-field of a towed horizontal receiver array during the Gulf of Maine 2006 Experiment. The source transmitted intermittent broadband pulses in the 300 to 1200 Hz frequency range. A nonlinear matched-filter kernel designed to replicate the acoustic signal measured by the receiver array is applied to enhance the signal-to-noise ratio. The source localization accuracy is found to be highly dependent on source-receiver geometry and the localization approach. For a relatively stationary source drifting at speeds much slower than the receiver array tow-speed, the mean source position can be estimated by moving array triangulation with less than 3% error near broadside direction. For a moving source, the Kalman filter method gives the best performance with 5.5% error. The array invariant is the best approach for localizing sources within the endfire beam of the receiver array with 7% error.

Clutter is related to false alarms for active sonar. It is demonstrated that, in shallow water, target-like clutter in reverberation signals can be caused by nonlinear internal waves. A nonlinear internal wave is modeled using measured stratification on the New Jersey shelf. Reverberation in the presence of the internal wave is modeled numerically. Calculations show that acoustic energy propagating near a sound speed minimum is deflected as a high intensity, higher angle beam into the bottom, where it is backscattered along the reciprocal path. The interaction of sound with the internal wave is isolated in space, hence resulting in a target-like clutter, which is found to be greater than 10 dB above the mean reverberation level.

Multistatic active sonar systems involve the transmission and reception of multiple probing sequences and can achieve significantly enhanced performance of target detection and localization through exploiting spatial diversity. This paper mainly focuses on two signal processing aspects of such systems, namely, enhanced range-Doppler imaging and improved target parameter estimation. The main contributions of this paper are (1) a hybrid dense-sparse method is proposed to generate range-Doppler images with both low sidelobe levels and high accuracy; (2) a generalized K-Means clustering (GKC) method for target association is developed to associate the range measurements from different transmitter-receiver pairs, which is actually a range fitting procedure; (3) the extended invariance principle-based weighted least-squares method is developed for accurate target position and velocity estimation. The effectiveness of the proposed multistatic active sonar signal processing techniques is verified using numerical examples.

Target localization can be accomplished through an ultrasonic sonar system equipped with an emitter and two receivers. Time of flight of the sonar echoes allows the calculation of the distance of the target. The orientation can be estimated from knowledge of the beam pattern of the receivers and the ratio, in the frequency domain, between the emitted and the received signals after compensation for distance effects and air absorption. The localization method is described and, as its performance strongly depends on the beam pattern, the search of the most appropriate sonar receiver in order to ensure the highest accuracy of target orientation estimations is developed in this paper. The structure designs considered are inspired by the ear shapes of some bat species. Parameters like flare rate, truncation angle, and tragus are considered in the design of the receiver structures. Simulations of the localization method allow us to state which combination of those parameters could provide the best real world implementation. Simulation results show the estimates of target orientations are, in the worst case, 2° with SNR = 50 dB using the receiver structure chosen for a potential practical implementation of a sonar system.

A high frequency experiment was conducted in the Woods Hole Harbor in Massachusetts to evaluate the effectiveness of Doppler sonar for discriminating targets from reverberation. Using a pulsed linear frequency modulated signal, one finds that the matched filtered outputs are filled with high-level discrete backscattered returns, referred to as clutter, which are often confused with the target echo. The high level non-target returns have an amplitude distribution that is heavy-tailed. Using a Doppler-sensitive binary-phase-shift-keying signal coded with an m-sequence, the target echo and clutter can be separated by Doppler and delay, and tracked using the Doppler spectrogram (Dopplergram). The Doppler filtered time series show a background reverberation with a Rayleigh-like amplitude distribution, with an improved signal-to-(peak) reverberation ratio compared with that without Doppler filtering. The reduced reverberation level with Doppler processing decreases the probability of false alarm (Pfa) for a given threshold level. Conversely, for a given Pfa, the higher signal-to-(peak) reverberation ratio implies a higher probability of detection. Transmission loss measurement was conducted to estimate some of the system parameters, e.g., the source level and target strength relative to the noise level.

Extraction of Doppler signatures that characterize human motion has attracted a growing interest in recent years. These Doppler signatures are generated by various components of the human body while walking, and contain unique features that can be used for human detection and recognition. Although, a significant amount of research has been done in radio frequency regime for human Doppler signature extraction, considerably less has been done in acoustics. In this work, 40 kHz ultrasonic sonar is employed to measure the Doppler signature generated by the motion of body segments using different electronic and signal processing schemes. These schemes are based on both analog and digital demodulation with homodyne and heterodyne receiver circuitry. The results and analyses from these different schemes are presented.

Biosonar signals radiated along the beam axis of an Atlantic bottlenose dolphin resemble short transient oscillations. As the azimuth of the measuring hydrophones in the horizontal plane progressively increases with respect to the beam axis the signals become progressively distorted. At approximately ±45°, the signals begin to divide into two components with the time difference between the components increasing with increasing angles. At ±90° or normal to the longitudinal axis of the animal, the time difference between the two pulses measured by the hydrophone on the right side of the dolphin's head is, on average, ∼11.9 μs larger than the time differences observed by the hydrophone on the left side of the dolphin's head. The center frequency of the first pulse is generally lower, by 33-47 kHz, than the center frequency of the second pulse. When considering the relative locations of the two phonic lips, the data suggest that the signals are being produced by one of the phonic lips and the second pulse resulting from a reflection within the head of the animal. The generation of biosonar signals is a complex process and the propagation pathways through the dolphin's head are not well understood.

Measurement of acoustic backscattering properties of targets requires removal of the range dependence of echoes. This process is called range compensation. For conventional sonars making measurements in the transducer farfield, the compensation removes effects of geometrical spreading and absorption. For parametric sonars consisting of a parametric acoustic transmitter and a conventional-sonar receiver, two additional range dependences require compensation when making measurements in the nonlinearly generated difference-frequency nearfield: an apparently increasing source level and a changing beamwidth. General expressions are derived for range compensation functions in the difference-frequency nearfield of parametric sonars. These are evaluated numerically for a parametric sonar whose difference-frequency band, effectively 1-6 kHz, is being used to observe Atlantic herring (Clupea harengus) in situ. Range compensation functions for this sonar are compared with corresponding functions for conventional sonars for the cases of single and multiple scatterers. Dependences of these range compensation functions on the parametric sonar transducer shape, size, acoustic power density, and hydrography are investigated. Parametric range compensation functions, when applied with calibration data, will enable difference-frequency echoes to be expressed in physical units of volume backscattering, and backscattering spectra, including fish-swimbladder-resonances, to be analyzed.

Iterative, single-channel time reversal is employed to isolate backscattering resonances of an air-filled spherical shell in a frequency range of 0.5-20 kHz. Numerical simulations of free-field target scattering suggest improved isolation of the dominant target response frequency in the presence of varying levels of stochastic noise, compared to processing returns from a single transmission and also coherent averaging. To test the efficacy of the technique in a realistic littoral environment, monostatic scattering experiments are conducted in the Gulf of Mexico near Panama City, Florida. The time reversal technique is applied to returns from a hollow spherical shell target sitting proud on a sandy bottom in 14 m deep water. Distinct resonances in the scattering response of the target are isolated, depending upon the bandwidth of the sonar system utilized.

Compressive sensing (CS) is compared with conventional beamforming using horizontal beamforming of at-sea, towed-array data. They are compared qualitatively using bearing time records and quantitatively using signal-to-interference ratio. Qualitatively, CS exhibits lower levels of background interference than conventional beamforming. Furthermore, bearing time records show increasing, but tolerable, levels of background interference when the number of elements is decreased. For the full array, CS generates signal-to-interference ratio of 12 dB, but conventional beamforming only 8 dB. The superiority of CS over conventional beamforming is much more pronounced with undersampling.

[EN] The phase angle component of the complex frequency response of a sonar system operating near transducer resonance is usually distorted. Interpretation and classification of the received sonar signal benefits from the preservation of waveform fidelity over the full bandwidth. A calibration process that measures the phase response in addition to the amplitude response is thus required. This paper describes an extension to the standard-target calibration method to include phase angle, without affecting the experimental apparatus, by using dual-frequency transmission pulses and frequency-domain data processing. This approach reduces the impact of unknown range and sound speed parameters upon phase calibration accuracy, as target phase is determined from the relationship of the two frequency components instead of relying on a local phase reference. Tungsten carbide spheres of various sizes were used to simultaneously calibrate the amplitude and phase response of an active sonar system in a laboratory tank. Experimental measurements of target phase spectra are in good agreement with values predicted from a theoretical model based upon full-wave analysis, over an operating frequency of 50-125 kHz. (C) 2011 Acoustical Society of America., The authors would like to thank Dr. Trevor Francis for his insight and valuable help with the computer script of the theoretical model. A.I.-C. gratefully acknowledges funding provided by the Mexican National Committee for Science and Technology (CONACYT).

Doppler analysis has been extensively used in active radar and sonar sensing to estimate the speed and direction of a single target within an imaging system resolution cell following deterministic theory. For target swarms, such as fish and plankton in the ocean, and raindrops, birds and bats in the atmosphere, multiple randomly moving targets typically occupy a single resolution cell, making single-target theory inadequate. Here, a method is developed for simultaneously estimating the instantaneous mean velocity and position of a group of randomly moving targets within a resolution cell, as well as the respective standard deviations across the group by Doppler analysis in free-space and in a stratified ocean waveguide. While the variance of the field scattered from the swarm is shown to typically dominate over the mean in the range-velocity ambiguity function, cross-spectral coherence remains and maintains high Doppler velocity and position resolution even for coherent signal processing algorithms such as the matched filter. For pseudo-random signals, the mean and variance of the swarms' velocity and position can be expressed in terms of the first two moments of the measured range-velocity ambiguity function. This is shown analytically for free-space and with Monte-Carlo simulations for an ocean waveguide.

This paper investigates the extraction of acoustic signatures from small boats using a passive sonar system. Noise radiated from a small boats consists of broadband noise and harmonically related tones that correspond to engine and propeller specifications. A signal processing method to automatically extract the harmonic structure of noise radiated from small boats is developed. The Harmonic Extraction and Analysis Tool (HEAT) estimates the instantaneous fundamental frequency of the harmonic tones, refines the fundamental frequency estimate using a Kalman filter, and automatically extracts the amplitudes of the harmonic tonals to generate a harmonic signature for the boat. Results are presented that show the HEAT algorithms ability to extract these signatures.

Many attempts have been made to model the signal transformations that occur in the mammalian cochlea, in particular with respect to modeling human hearing. Cochlea models that have been developed in the last few decades fall into several categories, e.g., electrical analogs, digital filter models, and incorporate different levels of complexity from entirely linear to strongly nonlinear. Similarly, the implementations used have included various numerical approaches as well as analog and digital hardware architectures. Mammalian hearing systems are known to be highly capable when it comes to analyzing sound from complex natural environments and soundscapes. It may hence be hypothesized that the signal transformations that occur in the cochlea play a critical role in facilitating these capabilities since the cochlea is positioned at the critical first stage of auditory signal encoding. The goal of the research presented here is to evaluate the signal transformations of the mammalian cochlea in the context of sonar signal processing, especially for automatic target recognition in difficult environments such as reverberant shallow water. The cochlea of bats could proof a suitable model for these applications since many bat species are known to be capable of solving demanding sonar tasks under difficult conditions and cluttered environments.Many attempts have been made to model the signal transformations that occur in the mammalian cochlea, in particular with respect to modeling human hearing. Cochlea models that have been developed in the last few decades fall into several categories, e.g., electrical analogs, digital filter models, and incorporate different levels of complexity from entirely linear to strongly nonlinear. Similarly, the implementations used have included various numerical approaches as well as analog and digital hardware architectures. Mammalian hearing systems are known to be highly capable when it comes to analyzing sound from complex natural environments and soundscapes. It may hence be hypothesized that the signal transformations that occur in the cochlea play a critical role in facilitating these capabilities since the cochlea is positioned at the critical first stage of auditory signal encoding. The goal of the research presented here is to evaluate the signal transformations of the mammalian cochlea in the context of...

This paper addresses the problem of field directionality mapping (FDM) or spatial spectrum estimation in dynamic environments with a maneuverable towed acoustic array. Array processing algorithms for towed arrays are typically designed assuming the array is straight, and are thus degraded during tow-ship maneuvers. In this paper, maneuvering the array is treated as a feature allowing for left and right disambiguation as well as improved resolution toward endfire. The Cramér-Rao lower bound is used to motivate the improvement in source localization which can be theoretically achieved by exploiting array maneuverability. Two methods for estimating time-varying field directionality with a maneuvering array are presented: (1) Maximum likelihood (ML) estimation solved using the expectation maximization algorithm and (2) a non-negative least squares (NNLS) approach. The NNLS method is designed to compute the field directionality from beamformed power outputs, while the ML algorithm uses raw sensor data. A multi-source simulation is used to illustrate both the proposed algorithms' ability to suppress ambiguous towed array backlobes and resolve closely spaced interferers near endfire which pose challenges for conventional beamforming approaches especially during array maneuvers. Receiver operating characteristics are presented to evaluate the algorithms' detection performance versus signal-to-noise ratio. The results indicate that both FDM algorithms offer the potential to provide superior detection performance when compared to conventional beamforming with a maneuverable array.

Wideband direction of arrival (DOA) estimation plays an important role in passive sonar signal processing. Recently, sparsity-based DOA estimation method has attracted considerable attention because of its high resolution in the condition of few snapshots and low signal-to-noise ratio. However, the localization accuracy is seriously affected by the interferences. Matrix filter (MF) has been widely used in passive sonar systems as a useful tool to passband the targets-of-interest while attenuating the interferences, but the output of the MF seriously affected subsequent DOA estimation when the power of the interferences after filtering is still stronger than the weak targets. In this paper, a method based on MF and sparse asymptotic minimum variance (SAMV) is given to localize the weak targets in a strong interference environment. The given method improves the ability of SAMV on weak targets localization and achieves high localization accuracy even in the condition that the power of the interferences after filtering remains stronger than the weak targets, which is verified by simulation and experimental results.

The coherent backscattering (or backscattering enhancement) effect results from persistent interferences between multiply scattered waves propagating along reciprocal paths. This effect results in a doubling of the recorded backscattered intensity reflected exactly in the backward direction with respect to any other direction, after averaging intensity measurements collected over multiple emission angles. This letter presents experimental observations of the coherent backscattering effect from mid-frequency (3-4 kHz) shallow water reverberation measurements collected using a two-dimensional hydrophone array (8.4 × 1.5 m(2)). These results demonstrate the presence of coherent effects in shallow water reverberation not commonly accounted for.

This article her is posted with by permission of Acoustical Society of America, for personal use, not for redistribution. Journal homepage: The difference-frequency band of the Kongsberg TOPAS PS18 parametric sub-bottom profiling sonar, nominally 1–6 kHz, is being used to observe Atlantic herring. Representative TOPAS echograms of herring layers and schools observed in situ in December 2008 and November 2009 are presented. These agree well with echograms of volume backscattering strength derived simultaneously with the narrowband Simrad EK60/18- and 38-kHz scientific echo sounder, also giving insight into herring avoidance behavior in relation to survey vessel passage. Progress in rendering the TOPAS echograms quantitative is described.

The scintillation statistics of broadband acoustic transmissions are determined as a function of signal bandwidth B, center frequency f(c), and range with experimental data in the New Jersey continental shelf. The received signal intensity is shown to follow the Gamma distribution implying that the central limit theorem has led to a fully saturated field from independent multimodal propagation contributions. The Gamma distribution depends on the mean intensity and the number of independent statistical fluctuations or coherent cells micro of the received signal. The latter is calculated for the matched filter, the Parseval sum, and the bandpassed center frequency, all of which are standard ocean acoustic receivers. The number of fluctuations mu of the received signal is found to be an order of magnitude smaller than the time-bandwidth product TB of the transmitted signal, and to increase monotonically with relative bandwidth Bfc. A computationally efficient numerical approach is developed to predict the mean intensity and the corresponding broadband transmission loss of a fluctuating, range-dependent ocean waveguide by range and depth averaging the output of a time-harmonic stochastic propagation model. This model enables efficient and accurate estimation of transmission loss over wide areas, which has become essential in wide-area sonar imaging applications.

Acoustic signals propagating through the ocean are refracted, scattered, and attenuated by the ocean volume and boundaries. Many aspects of how the ocean affects acoustic propagation are understood, such that the characteristics of a received signal can often be predicted with some degree of certainty. However, acoustic ocean parameters vary with time and location in a manner that is not, and cannot be, precisely known; some uncertainty will always remain. For this reason, the characteristics of the received signal can never be precisely predicted and must be described in probabilistic terms. A signal processing structure recently developed relies on knowledge of the ocean environment to predict the statistical characteristics of the received signal, and incorporates this description into the processor in order to detect and classify targets. Acoustic measurements at 250 Hz from the 1996 Strait of Gibraltar Acoustic Monitoring Experiment are used to illustrate how the processor utilizes environmental data to classify source depth and to underscore the importance of environmental model fidelity and completeness.

A major problem in moving platform active sonar systems is the detection of targets in spatially distributed and Doppler-spread reverberation. This paper presents a novel space time adaptive prewhitener for reverberation based on a two-dimensional autoregressive model. The space time adaptive prewhitener jointly processes received data in angle and Doppler to improve the separation of a target from reverberation. The detector using the space time adaptive prewhitener is shown to yield better detection performance than previously known schemes when operating in a reverberation background containing target echoes.

In underwater acoustics, the accuracy of computational field predictions is commonly limited by uncertainty in environmental parameters. An approximate technique for determining the probability density function (PDF) of computed field amplitude, A, from known environmental uncertainties is presented here. The technique can be applied to several, N, uncertain parameters simultaneously, requires N+1 field calculations, and can be used with any acoustic field model. The technique implicitly assumes independent input parameters and is based on finding the optimum spatial shift between field calculations completed at two different values of each uncertain parameter. This shift information is used to convert uncertain-environmental-parameter distributions into PDF(A). The technique's accuracy is good when the shifted fields match well. Its accuracy is evaluated in range-independent underwater sound channels via an L(1) error-norm defined between approximate and numerically converged results for PDF(A). In 50-m- and 100-m-deep sound channels with 0.5% uncertainty in depth (N=1) at frequencies between 100 and 800 Hz, and for ranges from 1 to 8 km, 95% of the approximate field-amplitude distributions generated L(1) values less than 0.52 using only two field calculations. Obtaining comparable accuracy from traditional methods requires of order 10 field calculations and up to 10(N) when N1.

The performance of broadband sonar array processing can degrade significantly in shallow-water environments when interference becomes angularly spread due to multipath propagation. Particularly for towed line arrays near endfire, elevation angle spreading of multipath interference often results in masking of weaker sources of interest. While adaptive beamforming in a series of narrow frequency bands can suppress coherent multipath interference, this approach requires long observation times to estimate the required narrowband covariance matrices. To form wideband covariance matrices which can be estimated with less observation time, plane-wave focusing methods have been used to avoid interference covariance matrix rank inflation. This paper extends wideband focusing to the case of coherent multipath interference. The approach presented here, called waveguide invariant focusing (WIF), exploits a robust relationship for the frequency dependence of horizontal wave number differences. Unlike matched-field methods, WIF does not model multipath wave fronts but rather makes the interference appear to occupy the same rank-one subspace across frequency. This permits formation of wideband covariance matrices without interference rank inflation. Simulation experiments in a realistic ocean environment indicate that adaptive beamforming using WIF covariance matrices can provide a significant array gain improvement over conventional adaptive methods with limited observation time.

Acoustic communication channel capacity determines the maximum data rate that can be supported by an acoustic channel for a given source power and source/receiver configuration. In this paper, broadband acoustic propagation modeling is applied to estimate the channel capacity for a time-invariant shallow-water waveguide for a single source-receiver pair and for vertical source and receiver arrays. Without bandwidth constraints, estimated single-input, single-output (SISO) capacities approach 10 megabitss at 1 km range, but beyond 2 km range they decay at a rate consistent with previous estimates by Peloquin and Leinhos (unpublished, 1997), which were based on a sonar equation calculation. Channel capacities subject to source bandwidth constraints are approximately 30-90% lower than for the unconstrained case, and exhibit a significant wind speed dependence. Channel capacity is investigated for single-input, multi-output (SIMO) and multi-input, multi-output (MIMO) systems, both for finite arrays and in the limit of a dense array spanning the entire water column. The limiting values of the SIMO and MIMO channel capacities for the modeled environment are found to be about four times higher and up to 200-400 times higher, respectively, than for the SISO case. Implications for underwater acoustic communication systems are discussed.

Multipath propagation in shallow water can lead to mismatch losses when single-path replicas are used for horizontal array beamforming. Matched field processing (MFP) seeks to remedy this by using full-field acoustic propagation models to predict the multipath arrival structure. Ideally, MFP can give source localization in range and depth as well as detection gains, but robustly estimating range and depth is difficult in practice. The approach described here seeks to collapse full-field replica outputs to bearing, which is robustly estimated, while retaining any signal gains provided by the full-field model. Cluster analysis is used to group together full-field replicas with similar responses. This yields a less redundant “sampled field” describing a set of representative multipath structures for each bearing. A detection algorithm is introduced that uses clustering to collapse beamformer outputs to bearing such that signal gains are retained while increases in the noise floor are minimized. Horizontal ar...

In theory, matched field processing offers the significant benefit of higher signal gains and increased localization capability. However, this has not been robustly observed in practice because of inherent uncertainties about details of the shallow water propagation environments which limit the prediction of the channel response. The use of guide sources to directly measure the transfer function between source and receiver arrays has been proposed as a means for reducing mismatch. However, the guide source measurement only provides a measured transfer function at the guide source location. In this paper a method of depth-shifting guide source observations is proposed, making it possible to estimate transfer functions for points in the ocean other than the guide source location. The proposed depth-shifting process does not require knowledge of environmental parameters. The theoretical background for the technique is developed below and its range of applicability is examined

The concept of environmentally adaptive reverberation nulling using a time reversal mirror (TRM) recently has been described [Song et al., J. Acoust. Soc. Am. 116, 762–768 (2004)]. In this paper, monostatic reverberation nulling is demonstrated experimentally at 850 and 3500Hz using data from a shallow water experiment conducted off the west coast of Italy in April 2003. The active transmission of a seafloor spatial null from a vertical source array is shown to result in the attenuation by 3–5dB of prominent reverberation features with their levels being reduced to that of the more diffuse reverberation background.

The three-dimensional beam pattern of a sperm whale (Physeter macrocephalus) tagged in the Ligurian Sea was derived using data on regular clicks from the tag and from hydrophones towed behind a ship circling the tagged whale. The tag defined the orientation of the whale, while sightings and beamformer data were used to locate the whale with respect to the ship. The existence of a narrow, forward-directed P1 beam with source levels exceeding 210 dBpeak re: 1 microPa at 1 m is confirmed. A modeled forward-beam pattern, that matches clicks20 degrees off-axis, predicts a directivity index of 26.7 dB and source levels of up to 229 dBpeak re: 1 microPa at 1 m. A broader backward-directed beam is produced by the P0 pulse with source levels near 200 dBpeak re: 1 microPa at 1 m and a directivity index of 7.4 dB. A low-frequency component with source levels near 190 dBpeak re: 1 microPa at 1 m is generated at the onset of the P0 pulse by air resonance. The results support the bent-horn model of sound production in sperm whales. While the sperm whale nose appears primarily adapted to produce an intense forward-directed sonar signal, less-directional click components convey information to conspecifics, and give rise to echoes from the seafloor and the surface, which may be useful for orientation during dives.

The performance of passive acoustic signal-processing techniques can become severely degraded when the acoustic source of interest is obscured by strong interference. The application of matrix filters to suppress interference while passing a signal of interest with minimal distortion is presented. An algorithm for single-frequency matrix filter design is developed by converting a constrained convex optimization problem into a sequence of unconstrained problems. The approach is extended to broadband data by incoherently combining the responses of matrix filters designed at frequencies across a band of interest. The responses of single-frequency and multifrequency matrix filters are shown. Examples are given which demonstrate the effectiveness of matrix filtering applied to matched-field localization of a weak source in the presence of a strong interferer and noise. These examples show the matrix filter effectively suppressing the interference, thereby enabling the localization of the weak source. Standard matched-field processing, without matrix filtering, is not effective in localizing the weak source.

Active acoustic time reversal may be accomplished by recording sounds with an array of transducers--a time-reversing array (TRA) or time-reversal mirror (TRM)--and then replaying the recorded and time-reversed sounds from the same array to produce back-propagating waves that converge at the location(s) of the remote sound source(s). Future active sonar and underwater communication systems suitable for use in unknown shallow ocean waters may be developed from the automatic spatial and temporal focusing properties of TRAs. However, ocean currents affect time reversal because they alter acoustic reciprocity in the environment. This paper presents a theoretical and computational investigation into how ocean currents influence TRA retrofocusing in shallow ocean environments for various array orientations. The case of TRA retrofocusing in a three-dimensional range-independent sound channel with a steady horizontal ocean current is covered here, based on a normal-mode propagation model valid for low Mach number currents. The main finding is that in the presence of ocean currents (typically1 m/s), a TRA performs well (the associated retrofocus amplitude decay is less than 1 dB) except that a retrofocus shift (up to a few wavelengths at 500 Hz at a range of 2.5 km) may occur due to the differing interaction between the ocean current profile and each acoustic normal mode. In addition, TRA performance is predicted to depend on the array orientation relative to the ocean current direction, especially for horizontal arrays.

Adaptive beamformers automatically adjust to the input data on the sensors in an attempt to maximize the bearing resolution of a signal and minimize the effects of noise or interfering signals. To the author's knowledge, all adaptive beamformers currently available in the literature assume that the sensors are stationary over the time required to collect the data. This assumption is invalid on arrays of sensors mounted on nonstationary platforms, and results in poor beamforming performance. In this paper we present an adaptive beamformer that has been designed to operate on nonstationary arrays. The beamformer directly incorporates any changes in array shape or heading that may occur during the time required to collect the data. The output of the adaptive beamformer is shown for both synthetic data and for real data collected on a towed array. Results show that signal detection, bearing accuracy, bearing resolution, and interference suppression all increase when the array shape and track are incorporated into the beamformer if the sensor platform is not stationary.

An inversion method based on the concept of back wave propagation (BWP) is described in this paper for estimation of geoacoustic parameters from acoustic field data. A phase-regulation technique is introduced to increase the sensitivity of the method for geoacoustic model parameters having low sensitivity. The case of data consisting of signal plus additive noise is also addressed. It is shown theoretically that the sensitivity can be increased by a factor alpha using the phase regulation procedure, and that the spatial resolution of signal energy that is concentrated by BWP at the known source position is increased when a increases. This result suggests an effective criterion for use in the inversion, based on the spatial distribution of signal energy around the true source location. The basis for the criterion is the spatial variance of the back-propagated pressure field in a window around the known source location. A multistep search process is proposed to avoid using a complicated multidimensional search process. Inversion results from both simulations and experimental data are given. The real data were taken from the Pacific Shelf experiment carried out in shallow water off the West Coast of Vancouver Island in the Northeast Pacific Ocean.

The talk will highlight a few of the many important contributions Professor Dyer made to underwater acoustics through his interactions with ONR as a researcher, advisor, and educator. His wide ranging expertise included sound and vibration in complex structures that led the way for the U.S. Navy to develop ultra-quiet submarines, as well as the statistics of sound propagation in the ocean, submarine sonar design, and advances in sonar signal processing. Beginning in 1978, he led and participated in six Arctic field programs, including the Canadian Basin Arctic Reverberation Experiment that imaged the entire Arctic basin with acoustics. Over the course of these efforts, that he really enjoyed talking about, he and his students developed a systematic understanding of sound propagation and noise in the Arctic.

This paper presents improvements to a previously described biomimetic system to make it more robust in terms of localizing objects and increasing the types of objects that can be classified. Time-of-flight determination employs a large threshold and searches the echo envelope backward in time to find a smaller threshold crossing point, rather than using a single small threshold. This change eliminates small nuisance echoes, which can cause instabilities in object positioning. Second, the echo envelope is also searched forward in time to find the threshold crossing point, thus defining a window containing significant echo data. This changes the analysis from a fixed-duration data window to one that is variable and allows echo duration to be an object feature. Third, separate feature vectors were extracted from right and left receivers rather than employing a concatenated feature vector. The new system can then classify objects that are not isolated. The major significance of these changes is that it allows...

The method of matched-field source tracking by ambiguity surface averaging is demonstrated using low-frequency shallow-water acoustic data recorded on a vertical array. In this method, source motion is taken into account by averaging values from individual, short time-average matched-field ambiguity functions; each resulting average value corresponds to a trial source track. The result is a new ambiguity function which depends on six parameters: the three-dimensional coordinates of the initial and final source positions. A simplification to four parameters is made by assuming the source remains at a constant depth. Only mild restrictions are made on trial tracks considered, and the resulting ambiguity surfaces can be plotted as a function of final coordinates by holding the initial coordinates fixed at the optimal position. When applied to experimental data, the number of high sidelobes present in individual matched-field ambiguity surfaces is greatly reduced. Also, the technique appears robust to uncertainties in replica computation (due to various mismatch problems) and periods when the signal-to-noise ratio is low.

An algorithm is introduced for detecting signal and echo onsets and estimating their arrival times in multipath acoustic data. The algorithm accommodates superimposed narrow-band signals with continuous onsets, unknown envelopes, and arrival-time separations of less than one cycle at the dominant tonal frequency. It does not require the multipath components to be shifted and weighted replicas of each other or of a template signal. The separation of closely spaced signals is enabled by applying a prefilter that isolates spectral “tail” energy in which the onsets are discernible, and by operating on the filtered data using a matched-subspace detector that is conditioned on previously detected onsets. This algorithm has particular value for laboratory measurements, as it utilizes the high signal-to-noise ratios available in a laboratory to solve a previously intractable detection and estimation problem.

Current research on classification of submerged objects is concerned with using broadband sonar signals to insonify the targets, and applying signal-processing techniques to the backscattered signals. One characteristic of target echoes which may provide classification clues is the so-called resonance scattering response, the characteristics of which depend on target elastic properties. This paper presents algorithms for extracting resonance information based on autoregressive (AR) spectral estimation techniques. The AR-based representation is useful for detecting and accurately localizing resonances in the frequency domain. The extracted resonance frequencies are grouped into identified wave families, and processed in order to characterize the scatterer in terms of elastic and geometrical parameters on the basis of equations derived from resonance scattering theory. The targets considered are water-loaded elastic, cylindrical, thin-walled shells immersed in salt water under free-field conditions. Analysis was performed on data collected at sea at low-intermediate frequencies [ka∈(0,50)] and provided good results. The proposed approach is a first step towards the characterization of more complex targets either lying proud on the seabed or buried in bottom sediments.

Realistic simulations with spatially coherent noise have been run in order to compare the performance of adaptive beamforming (ABF), inverse beamforming (IBF), and conventional beamforming (CBF) for the case of finite-averaging times, where the actual spatial coherence of the acoustic field, or covariance matrix, is not known a priori, but must be estimated. These estimation errors cause large errors in the ABF estimate of the directionality of the acoustic field, partly because ABF is a highly nonlinear algorithm. In addition, it is shown that ABF is fundamentally limited in its suppression capability at very low frequency (VLF), based on the sidelobe level of the conventional beampattern in the direction of the noise interferer [G. L. Mohnkern, "Effects of Errors and Limitations on Interference Suppression," NOSC Technical Document 1478, Naval Ocean Systems Center (1989)]. The simulations include a low-level plane wave signal of interest, a stronger noise plane wave interferer, and spatially random background noise. Both IBF and ABF performed significantly better than CBF, and IBF's performance was slightly better than ABF's performance. The performances of IBF and the ABF algorithm, the minimum variance distortionless response (MVDR) [A. H. Nuttall and D. W. Hyde, "Unified Approach to Optimum and Suboptimum Processing for Arrays," USL Report Number 992, Naval Underwater Systems Center, New London, CT (22 April 1969)] were recently compared independently [J. S. D. Solomon, A. J. Knight, and M. V. Greening, "Sonar Array Signal Processing for Sparse Linear Arrays," Defense Science and Technology Organization (DSTO) Technical Report (June 1999)] using measured data, with the result that IBF outperformed MVDR. This result is significant because MVDR requires orders of magnitude more processing power than IBF or CBF.

Passive broadband detection of target signals by an array of hydrophones in the presence of multiple discrete interferers is analyzed under Gaussian statistics and low signal-to-noise ratio conditions. A nonhomogeneous Poisson-modulated interference process is used to model the ensemble of possible arrival directions of the discrete interferers. Closed-form expressions are derived for the recognition differential of the passive-sonar equation in the presence of Poisson-modulated interference. The interference-compensated recognition differential differs from the classical recognition differential by an additive positive term that depend on the interference-to-noise ratio, the directionality of the Poisson-modulated interference, and the array beam pattern.

Coherent multi-frequency matched-field processing is investigated using a matched-phase coherent matched-field processor. Its main difference from previous coherent processors is that the relative phases of the Fourier components contained within the recorded signal are not assumed to be known a priori. Rather they are considered free parameters that can be determined using a global functional minimization algorithm. Additionally, this processor uses only the cross-frequency terms, making it less susceptible to the detrimental effects of ambient noise; in one example, this processor shows a five decibel improvement over a similar coherent processor. Along with its increased sensitivity with respect to the broadcast source levels, this coherent processor exhibits superior range resolution as compared with multi-frequency incoherent processors, due to the cross-frequency interference of the vertical eigenmodes. Within this work we explore the efficacy of the algorithms used to determine the relative phases along with the performance of the matched-phase coherent processor itself, performed within the context of data collected during an event from the SWellEx-96 experiment. Performance comparisons between this processor, an incoherent processor, and another coherent processor are demonstrated using this data set.

The transmission beam pattern of an echolocating harbor porpoise (Phocoena phocoena) was measured in both the vertical and horizontal planes. An array of seven Brüel and Kjaer 8103 hydrophones connected to an amplifier-line driver module was used to measure the beam patterns. The porpoise was trained to station in a hoop and echolocate a cylindrical target located at a range between 7 and 9 m while the array was located 2 m in front of the hoop. The 3-dB beamwidth in both the vertical and horizontal planes was the same at approximately 16 degrees and the beam was pointed toward the forward direction. The individual hydrophones in both the vertical and horizontal arrays measured signal waveforms that were similar throughout the 40-degree span of the array. The porpoise emitted signals with intervals that were 20 to 35 ms longer than the round trip travel time between the animal and the target. The average source level, peak frequency, and bandwidth were 157 dB, 128 kHz, and 16 kHz, respectively.

Matched-field inversion (MFI) techniques have been applied for effective and efficient estimation of geoacoustic parameters of the ocean bottom. This paper presents a new tomographic MFI method for use in range-dependent environments. The MFI correlates modeled data with measured data and uses a search algorithm to determine model parameter values that maximize the correlator. In the present method, a parabolic equation propagation model is used to compute replica fields to account for mode coupling in the environment. The search algorithm is a two-stage hybrid method that combines an initial global component and a final local component. The first stage employs a random search to determine N+1 parameter sets with the best correlations, where N is the number of parameters being determined. In the next stage the N+1 sets are used as inputs to the local downhill simplex algorithm. The algorithm is shown to perform well for simulated vertical line array data for an environment representative of the SWellEX-96 experimental site. The inversion technique is then applied to measured data obtained from a radial track in SWellEX-96. The geometric parameters of the experiment and dominant geoacoustic parameters were successfully recovered for data with relatively low signal-to-noise ratio.

The principal aim of this work is to estimate the complex resonant frequencies of scattering returns from the Mock-0 target. The experiments were conducted at the Naval Undersea Warfare Center. First, the singular value decomposition approach is employed for noise suppressing prefiltering where the initial estimate of the system order is obtained. Second, the eigen-matrix pencil method is employed to approximate the complex resonant frequencies. Finally, the system order and the complex resonances are determined by a systematic procedure in which the estimated results obtained from various system orders and aspect angles are compared. Echoes observed at the same aspect angle but excited by different incident pulses with overlapping frequency spectra can also be employed to obtain reliable estimations of resonant frequencies. The key feature of the proposed approach is robustness.