Processing For Source Localization Research Articles

Matched-field source localization with multiple small-aperture arrays

This paper considers combining information from multiple small-aperture arrays in matched-field processing (MFP) for source localization. Assuming individual arrays are comprised of calibrated sensors which are synchronized in time, conventional MFP can be applied for each array and the resulting Bartlett processors summed over arrays. However, if the relative calibration and/or time synchronization is known between some or all arrays, more informative multiple-array processors can be derived by maximum-likelihood methods. For example, if the relative calibration between arrays is known, the observed amplitude variations between arrays provide additional information for source localization; if synchronization is known, phase variations provide localization information. Various multiple-array processors are derived and evaluated in terms of the probability of correct localization from Monte-Carlo analyses for a range of signal-to-noise ratios and number of frequencies for simulated shallow-water scenarios with vertical and horizontal arrays. Effects of environmental mismatch in seabed geoacoustic parameters and water depth are also considered. The analysis indicates that, dependent on array configurations, substantial improvements in source localization performance can be achieved when including relative amplitude and/or phase information in the multiple-array processor. The improvement is reduced by environmental mismatch; this degradation can be partially mitigated by including additional frequencies in the processing.

Coherent vector sensor processing for autonomous underwater glider networks

A distributed array of autonomous underwater gliders, each fitted with a vector sensor measuring acoustic pressure and velocity, form an autonomous sensor network theoretically capable of detecting and tracking objects in an ocean environment. However, uncertainties in sensor positions impede the ability of this glider network to perform optimally. Our work aims to compare the performance of coherent and incoherent processing for acoustic source localization using an array of underwater gliders. Data used in the study were obtained from numerical simulations as well as experimental data collected using the research vessel as a source for localization purposes. By estimating the vessel position with a single glider’s data (incoherent) and comparing to the location estimated with both gliders’ data (coherent), it was determined that location estimation accuracy could be improved using coherent processing, provided the gliders’ positions could be measured with sufficient precision. The results of this study could potentially aid the design and navigation strategies of future glider networks with a large number of elements.

Information-theoretic quantification of underwater acoustic source localization performance

Metrics that historically have been applied to quantify the performance of signal processing for source localization are algorithm-dependent. For example, performance of conventional beamforming or matched-field processing is usually quantified by main-peak width and secondary-peak levels of the beam response or spatial ambiguity function, while performance of Bayesian localization may be quantified by measures of the statistical dispersion of the a posteriori pdf of source location. While algorithm-dependent performance metrics permit comparisons within a given class of signal processing algorithms, they do not provide comparability across algorithm classes. The present work identifies fundamental information-theoretic quantities that can be used as metrics to quantify the source localization performance of diverse signal processing algorithms and thus provide for performance comparisons across signal-processor classes. These quantities include conditional entropy of source location given processor output, mutual information of source location and processor output, and cross-entropy of actual and posterior source-location probability distributions. Applications of these information-theoretic metrics are illustrated in examples of Bayesian localization, conventional beamforming, and matched-field processing of a time-harmonic source in a range-independent shallow-water acoustic waveguide. The results are interpreted in the light of the data processing inequality of information theory. [Work supported by ONR.]

Source localization and tracking in a dispersive medium using wireless sensor network

In this paper, we address the issue of collaborative information processing for diffusive source localization and tracking using wireless sensor networks capable of sensing in dispersive medium/environment. We first determine the space-time concentration distribution of the dispersion from physical modeling and mathematical formulations of an underwater oil spill scenario, considering the effect of laminar water velocity as an external force. For static diffusive source localization, we propose two parametric estimation techniques based on maximum-likelihood (ML) and best linear unbiased estimator for the special case of our physical dispersion model. We prove the consistency and asymptotic normality of the obtained ML solution when the number of sensor nodes and samples approach infinity, and derive the Cramer-Rao lower bound on its performance. We also propose a particle filter-based target tracking scheme for moving diffusive source and derive the posterior Cramer-Rao lower bound for the moving source state estimates as a theoretical performance bound. The performance of the proposed schemes are shown through numerical simulations and compared with the derived theoretical bounds.

Matched Field Source Localization via Environmental Focalization

In a realistic application of matched field processing for source localization, the ability to localize an acoustic source is strongly affected by the uncertainty of the environment. However, accurate environment measurement is a difficult task in large regions of the ocean. We employ a Bayesian approach, referred to here as focalization, to overcome the mismatch by including the environment in the parameter search space. Focalization maximizes the posterior probability density over the unknown source and environmental parameters. The broadband signals recorded by a vertical line array in a Yellow Sea experiment are presented to demonstrate the feasibility of focalization.

Impact source localization in plate utilizing multiple signal classification

This article proposed a multiple signal classification method based on array signal processing for impact source localization in a plate. For source localization, the direction of arrival of the wave caused by an impact on a plate and the distance between the impact position and sensor should be estimated. The direction of arrival can be estimated accurately using the multiple signal classification method; the distance can be obtained using the time delay of arrival and the group velocity of the Lamb wave in a plate. The time delay of arrival is experimentally estimated using the continuous wavelet transform for the wave. The group velocity is theoretically obtained based on the elastodynamic theory.

Odor Source Localization Using Multiple Robots in Complicated City-Like Environments

The detection of a dangerous emission source location has the potential to be enhanced by using plume-tracing mobile robots, without endangering human life during the detection and source localization process. So far, many researchers focus on odor source localization in simple & laboratory based environments. The present study focuses on more real life odor source localization scenarios. In this study, multiple robots were used and coordinated by a supervisory program to locate an odor source in complicated city-like environments. A series of simulations has been conducted and the results demonstrated the potential of the supervisory program to effectively control a number of robots to locate a dangerous odor source in real life scenarios.

Compressing the computational burden of matched‐field processing for sound source localization.

Sound source localization in shallow water environment is done using matched field processing (MFP), also referred to as time‐reversal imaging or backprojection method. Standard MFP is usually implemented by matching (using a cross‐correlation operation) the received acoustic data from the sought source with modeled data (or replicas) for point source located at multiple test locations over the a priori search area. Consequently, a direct implementation of MFP (i.e., brute force search) over a large search area can be highly computationally demanding especially when attempting to locate repeatedly several stationary or moving unknown sources in a complex environment. We formulated instead a compressive MFP approach allowing for significant computation time savings with a computational cost that scales with the logarithm of the number of independent‐or uncorrelated‐ test locations. This approach leverages the key concepts behind group testing and compressed sensing by computing instead the expected value of multiple ambiguity surface realizations obtained by backprojection of orthogonal random signals from the receiver locations. Thus compressive MFP allows to estimate the actual ambiguity function using few replica computations by evaluating hypothetical random groups of test locations instead of individual test locations (as done in conventional MFP).

Evidence of binaural directional computations in the auditory pathway of a teleost fish.

Although binaural processing is known to be critical for sound source localization in terrestrial vertebrates, binaural processing for sound source localization in fishes has not been demonstrated. Previous physiological and neuroanatomical studies have indicated that interaural convergence is present in auditory pathway of oyster toadfish (Opsanus tau). In this study, we introduce a reversible method, otolith tipping, to evaluate the presence of binaural inputs. The normal directional response pattern (DRP) was recorded extracellularly for auditory cells in the first‐order descending octaval nucleus (DON) or the midbrain torus semicircularis (TS) using directional stimuli in the horizontal and midsagittal planes. The DRP was re‐evaluated while gently tipping the ipsilateral or contralateral otolith of the saccule to alter the normal relationship with the auditory epithelium. The DRP was also evaluated post‐tipping to ensure no damage had occurred. In the DON, ipsilateral or contralateral tipping most often altered or eliminated spike activity; in the TS, tipping most often caused a change in the shape or orientation of the DRP. The data indicate complex interactions of excitatory and inhibitory inputs, which appear to be associated with the convergence of binaural inputs in the DON and TS. [Work supported by the NIH.]

Source amplitude spectral information in matched-field localization.

This paper examines a variety of approaches to make use of knowledge of the relative amplitude spectrum of an acoustic source (but no knowledge of the phase spectrum) in multifrequency matched-field processing for source localization. A common example of this procedure involves cases where the source amplitude spectrum can be considered flat over the frequency band of interest. The primary issue is how to combine the information of complex acoustic fields at multiple frequencies, given the unknown phase spectrum. Approaches examined include maximum-likelihood phase estimation, pair-wise processing, and phase rotation to zero the phase at a specific sensor or to zero the mean phase over the array. The performance of the various approaches (processors) is quantified in terms of the probability of localizing the source within an acceptable range-depth region, as computed via Monte Carlo sampling over a large number of random realizations of noise and of environmental parameters. Processor performance is compared as a function of signal-to-noise ratio, number of frequencies, number of sensors, and number of time samples (snapshots) included in the signal averaging.

Effect of array element location on coherent inter‐array processing

In the SWellEx‐96 experiment, two 25‐elements horizontal arrays, each of length 250 m, were moored 3.5 km apart [N.O. Booth et al, IEEE, JOE, 25, no 3, July 2000]. A broadband source (50 to 400 Hz tones) was towed from 1 km to 10 km away from the arrays. The data is processed using the simple conventional plane‐wave beamformer for all the cases covering incoherent frequency and incoherent inter‐array processing to coherent frequency and coherent inter‐array processing as a function of source position relative to the arrays. The effect of noise and array element location mismatch on source localization and coherent processing gain is investigated. An attempt to improve array element localization is made using a self‐focusing technique. [Work supported by ONR.]

A neural network for mode space source depth estimation

A great deal of research in mode space source localization processing exists. One technique proposed by Shang performs mode filtering of the observed pressure field on a vertical line array, using the calculated mode amplitude distributions and phase differences to estimate source location. Ozard et al. proposed a feed forward neural network using input vectors of mode energy distribution for source localization, a dedicated neuron for each possible source location was employed in their study. Instead of employing a dedicated neuron for each possible search location, the interest of this study is to determine an efficient setup of neurons to properly sample the search space. In MATLAB neural network toolbox, the self-organizing-map function learns to classify input vectors according to how they are grouped in the input space, specifically their distribution and topology. We used this function to examine how input vectors of mode energy distribution for signals at different depths are grouped to determine the sampling grids of the search space in depth. We also compared the neural network performance using input vectors of mode amplitude distribution with those using input vectors of mode energy distribution.

Validation of SOBI components from high-density EEG

Second-order blind identification (SOBI) is a blind source separation (BSS) algorithm that can be used to decompose mixtures of signals into a set of components or putative recovered sources. Previously, SOBI, as well as other BSS algorithms, has been applied to magnetoencephalography (MEG) and electroencephalography (EEG) data. These BSS algorithms have been shown to recover components that appear to be physiologically and neuroanatomically interpretable. While some proponents of these algorithms suggest that fundamental discoveries about the human brain might be made through the application of these techniques, validation of BSS components has not yet received sufficient attention. Here we present two experiments for validating SOBI-recovered components. The first takes advantage of the fact that noise sources associated with individual sensors can be objectively validated independently from the SOBI process. The second utilizes the fact that the time course and location of primary somatosensory (SI) cortex activation by median nerve stimulation have been extensively characterized using converging imaging methods. In this paper, using both known noise sources and highly constrained and well-characterized neuronal sources, we provide validation for SOBI decomposition of high-density EEG data. We show that SOBI is able to (1) recover known noise sources that were either spontaneously occurring or artificially induced; (2) recover neuronal sources activated by median nerve stimulation that were spatially and temporally consistent with estimates obtained from previous EEG, MEG, and fMRI studies; (3) improve the signal-to-noise ratio (SNR) of somatosensory-evoked potentials (SEPs); and (4) reduce the level of subjectivity involved in the source localization process.

Regularized matched-mode processing for source localization.

This paper develops a new approach to matched-mode processing (MMP) for ocean acoustic source localization. MMP consists of decomposing far-field acoustic data measured at an array of sensors to obtain the excitations of the propagating modes, then matching these with modeled replica excitations computed for a grid of possible source locations. However, modal decomposition can be ill-posed and unstable if the sensor array does not provide an adequate spatial sampling of the acoustic field (i.e., the problem is underdetermined). For such cases, standard decomposition methods yield minimum-norm solutions that are biased towards zero. Although these methods provide a mathematical solution (i.e., a stable solution that fits the data), they may not represent the most physically meaningful solution. The new approach of regularized matched-mode processing (RMMP) carries out an independent modal decomposition prior to comparison with the replica excitations for each grid point, using the replica itself as the a priori estimate in a regularized inversion. For grid points at or near the source location, this should provide a more physically meaningful decomposition; at other points, the procedure provides a stable inversion. In this paper, RMMP is compared to standard MMP and matched-field processing for a series of realistic synthetic test cases, including a variety of noise levels and sensor array configurations, as well as the effects of environmental mismatch.

Regularized matched-mode localization with environmental mismatch

This paper considers a new approach to matched-mode processing (MMP) for source localization. The MMP consists of decomposing far-field acoustic data to obtain the modal excitations, then matching these with modeled replica excitations. A potential advantage of MMP over matched-field processing (MFP) is that subsets of the complete mode set can be considered. For example, if geoacoustic properties are poorly known, the matching can be applied only to low-order modes that interact minimally with the seabed. However, modal decomposition can be ill posed and unstable if the sensor array does not adequately sample the acoustic field. For such cases, standard decomposition methods yield minimum-norm solutions that are biased towards zero. Although these methods provide mathematical solutions (stable solutions that fit the data), they may not represent physically meaningful solutions. The new approach of regularized MMP (RMMP) carries out an independent decomposition prior to comparison with the replica excitations for each grid point, using the replica itself as the prior estimate in a regularized inversion. This provides a more meaningful decomposition near the actual source location. In this paper, RMMP, MMP, and MFP are compared for realistic test cases, including various sensor array configurations, as well as environmental mismatch in seabed properties.

Adaptive matched beam processing for source localization

Matched beam processing (MBP) has been previously applied to a horizontal line array for localizing a moving source in bearing and range, and for differentiating a submerged target from a surface clutter [T. C. Yang and T. Yates, J. Acoust. Soc. Am. 104 (1998)]. MBP is matched field processing in the beam domain. Using the Bartlett processor, bearing and depth resolutions are often inadequate for rejecting surface clutter and other noise sources. High-resolution processors may be required. We note that in situations where many clutters exist, adaptive (plane-wave) beamforming has effectively rejected clutter by either high-angular resolution or by steering a null to the interference source. Motivated by this approach, an adaptive MBP will be investigated in this paper. The new processor will apply adaptive beamforming to the data and replica fields such that interference has already been removed before the matched field correlation. This approach differs from the adaptive matched field processing where interference remains in the data field. The latter approach is known to be sensitive to environmental mismatch, as a result the correlation between the replica and the interference fields may be present as substantial clutter. Initial numerical results will be presented using simulated data with interference sources. [Work supported by the Office of Naval Research.]

Brain Topography TodayBrain Topography Today Edited by KogaY.NagataK. and HirataK.. Available from Elsevier Science BV, P.O. Box 211, 1000 AE Amsterdam, The Netherlands, or from Elsevier Science Inc., P.O. Box 945, Madison Square Station, New York NY 10160-0757, USA. Price: US$293.00. ISBN 0-444-82778-2.

Book ReviewsBrain Topography Today Zuheir Al-KawiMD, FAAN Zuheir Al-Kawi Department of Neurosciences, MBC-76, King Faisal Specialist Hospital and Research Centre, P.O. Box 3354, Riyadh 11211, Saudi Arabia Search for more papers by this author Published Online:1 Nov 1998https://doi.org/10.5144/0256-4947.1998.571SectionsPDF ToolsAdd to favoritesDownload citationTrack citations ShareShare onFacebookTwitterLinked InRedditEmail AboutIntroductionThis book is a report on the proceedings of the third Pan Pacific Conference on Brain Topography held in Tokyo Bay, Japan, on 1-4 April 1997.The advances in computer technology have made it possible to crunch large amounts of numeric data into visual images that are easy for the human brain to understand. The sources of data are multiple and include recorded spontaneous or evoked cerebral electrical activity (EEG and evoked potentials), recorded minute magnetic fields from the surface of the brain (MEG, or magneto-encephalogram), received signals that are emitted from protons within the brain in response to radio-frequency stimulation (MRI and fMRI), and energy generated by annihilation of positrons contained in special radiopharmaceuticals that are delivered to the entire brain or targeted to specific areas (PET), etc. Sophisticated computer algorithms have made it possible to localize such sources with precision and to display them as maps or images. The subject of the book is, therefore, a multidisciplinary look at the functions of the brain through diverse techniques unified by their ability to produce a visual end product.The book contains over 140 articles contributed by nearly triple that number of authors. Titles range from “Event-related potentials in alternation of perceived images in binocular rivalry” to “Topographical analysis of CNV during listening to background music: rock vs. classic.” Articles are grouped under 15 headings covering basic areas, such as source localization and visual information processing, or applied areas such as epilepsy, psychiatric and neurological disorders.Proceedings are usually prepared in a hurry, with less emphasis on the quality of the print. The publisher has done an exceptional job in the quality of the print and image reproduction.The subject is sophisticated and can be of interest to subspecialists in neurosciences, such as neurophysio-logists, neuroradiologists and neuropsychologists, especially those involved in research. Considering the limited audience, the book price is not unreasonable. It should be a useful reference for researchers on the subject. Previous article Next article FiguresReferencesRelatedDetails Volume 18, Issue 6November 1998 Metrics History Published online1 November 1998 InformationCopyright © 1998, Annals of Saudi MedicinePDF download

Matched-field processing for multitonal source localization

Recent studies with simulations and real data have shown that multiple frequency information can lead to increased robustness in matched-field estimation. This observation generates the question of how this information should be combined for best results. In this work an estimator is proposed that is suitable for matched-field inversion involving multitonal sources. The processor operates on the basis of an interfrequency spatially coherent framework. Designed for passive problems, it does not require knowledge of the source spectrum, an advantage over other coherent schemes. Moreover, the processor has the additional advantage of a simple implementation. Using synthetic data, the new coherent processor was compared to the conventional Bartlett processor with incoherent averaging across frequencies. The results indicate that the new processor is, in general, preferable to the conventional Bartlett processor. Interestingly, the superiority of the coherent processor is most pronounced when the receiving array undersamples the propagation medium, making the processor attractive for its cost effectiveness. [Work supported by ONR.]

Analytical and experimental comparison of azimuth/range/depth bias errors in MFP source localization using vertical and tilted arrays

During a recent shallow-water (200-m) experiment west of San Diego, two 120-m aperture hydrophone arrays were deployed from the R/P FLIP. One array was oriented in the vertical direction, while the second was deployed at a 45-deg tilt. A set of 13 tones, from 49 to 388 Hz, were transmitted from a 60-m-deep towed source traveling cross slope over a 6- km track. At CPA, the source was at endfire to the horizontal aperture of the tilted array. The data from both arrays were used to perform matched- field processing (MFP) for source localization in range and depth using replica vectors generated from a range-independent environmental model. In addition, the data from the tilted array were used to simultaneously localize the source in azimuth. The range and depth MFP biases of the two arrays agree quite well with one another and with analytical predictions from previous work. Here, those results are extended to successfully predict the offset bias in the MFP azimuth results, which reaches a maximum of 20 deg at endfire. [Work supported by ONR, Code 321(US).]

Matched-field inversion for source location and optimal equivalent bathymetry

Matched-field processing (MFP) for source localization can be highly sensitive to mismatch between the true bathymetry and the bathymetry model used to compute the replica fields. This paper presents an efficient algorithm to invert for source location and an optimal representation of the (range-dependent) bathymetry which is applicable whenever the adiabatic normal-mode model applies. In the adiabatic mode formulation, the range dependence of the bathymetry between source and receiver is completely contained in range integrals of the wave numbers. Parametrizing these integrals by a single-pole expansion maps all bathymetries which have an equivalent effect on the propagation into as little as a single parameter, resulting in a tremendous reduction in the dimensionality of the inverse problem. The resulting parameter space can then be searched for the bathymetry parameters which provide the best match with the measured acoustic fields. This search can be carried out very efficiently, and applied to optimize the bathymetry for each grid point of an ambiguity surface. The optimization procedure has the effect of broadening the peaks of the ambiguity surface. This results in a loss in resolution, but allows the use of a coarser search grid.