Search

In this paper, we propose a learning-based approach to identify the type and position of sounds using a single microphone in a real-world building. We attempt to treat this problem as a joint classification problem in which we predict the exact positions of sounds while classifying the types that are assumed to be from pre-defined types of sounds. The most problematic issue is that while the types are readily classified under supervised learning frameworks with one-hot encoded labels, it is difficult to predict the exact positions of the sound from unseen positions during training. To address this potential discrepancy, we formulate the position identification problem as a zero-shot learning problem inspired by the human ability to perceive new concepts from previously learned concepts. We extract feature representations from audio data and vectorize the type and position of the sound source as ‘type/position-aware attributes’, instead of labeling each class with a simple one-hot vector. We then train a promising generative model to bridge the extracted features and the attributes by learning the class-invariant structure to transfer the knowledge from seen to unseen classes through their attributes; generative adversarial networks are conditioned on the class-embeddings. Our proposed methods are evaluated on an indoor noise dataset, SNU-B36-EX, a real-world dataset collected inside a building.

Ray-based blind deconvolution (RBD) is a method that estimates the source waveform and channel impulse response (CIR) using the ray arrival in an underwater environment. The RBD estimates the phase of the source waveform by using beamforming. However, low sampling, array shape deformation, and other factors can cause phase errors in the beamforming results. In this paper, phase correction is applied to the beamforming estimated source phase to improve RBD performance. The impulsiveness of the CIR was used as additional information to correct the initially estimated source phase. Kurtosis was used to measure impulsiveness, and the phase correction that maximized the kurtosis of the CIRs was calculated through optimization. The proposed approach is called ray-based blind deconvolution with maximum kurtosis phase correction (RBD-MKPC) and is based on a single-input multiple-output system. The RBD-MKPC was tested with several CIRs and source waveform combinations in the shallow-water acoustic variability experiment 2015 using broadband high-frequency pulses (11–31 kHz) as the source and a sparse vertical 16-element line array as receivers. The results indicate that the RBD-MKPC improves the estimation performance. In addition, from an optimization point of view and compared with other initialization methods, the proposed method showed superior convergence speed and estimation performance.

The sensitivity of underwater propagation models to acoustic and environmental variability increases with the signal frequency; therefore, realizing accurate acoustic propagation predictions is difficult. Owing to this mismatch between the model and actual scenarios, achieving high-frequency source localization using model-based methods is generally difficult. To address this issue, we propose a deep learning approach trained on real data. In this study, we focused on depth estimation. Several 18-layer residual neural networks were trained on a normalized log-scaled spectrogram that was measured using a single hydrophone. The algorithm was evaluated using measured data transmitted from the linear frequency modulation chirp probe (11-31 kHz) in the shallow-water acoustic variability experiment 2015. The signal was received through two vertical line arrays (VLAs). The proposed method was applied to all 16 sensors of the VLA to determine the estimation performance with respect to the receiver depth. Furthermore, frequency-difference matched field processing was applied to the experimental data for comparison. The results indicate that ResNet can determine complicated features of high-frequency signals and predict depths, regardless of the receiver depth, while exhibiting robust environmental and positional variability.

In many industrial systems involving pipe or duct structures, sound transmission modeling is critical to noise treatment and control. A typical noise treatment method applied to pipe structures is to use a silencer. In particular, a dissipative silencer with a sound absorbing material has a noise control ability to reduce noise through successive sound reflections by impedance mismatch while dissipating incident sound energy as heat. In this study, the acoustic performance of a dissipative silencer lined with multi-layered absorbing materials for fluid-filled pipe system is investigated. Herein, the absorbing materials are modeled applying poroelastic and viscoelastic theories. In the analysis, the eigenvalues and eigenfunctions in the dissipative silencer are determined from the rigid wall boundary condition and the interfaces between the absorbing materials, and then the acoustic performance is compared by obtaining the transmission loss using the mode matching method. Also, a parametric study is conducted to investigate the effect of the number and arrangement of layers on the acoustic performance of the silencer according to the absorbing material properties.

A convolutional neural network (CNN)-based inter-floor noise source type classifier and locator with input from a single microphone was proposed in [Appl. Sci. 9, 3735 (2019)] and validated in a campus building experiment. In this study, the following extensions are presented: (1) data collections of nearly 4700 inter-floor noise events that contain the same noise types as those in the previous work at source positions on the floors above/below in two actual apartment buildings with spatial diversity, (2) the CNN-based method for source type classification and localization of inter-floor noise samples in apartment buildings, (3) the limitations of the method as verified through several tasks considering actual application scenarios, and (4) source type and localization knowledge transfer between the two apartment buildings. These results reveal the generalizability of the CNN-based method to inter-floor noise classification and the feasibility of classification knowledge transfer between residential buildings. The use of a short and early part of event signal is shown as an important factor for localization knowledge transfer.

An approximate form of three-dimensional Cartesian split-step marching solution for the acoustic parabolic equation is derived in order to obtain the efficient algorithm for sound propagation in the three-dimensional ocean. The operator splitting method is used to split the full exponential operator into three exponential operators for depth, cross-range, and the combination of the two. The first two terms are implemented with the split-step Padé algorithm and the final term is implemented with the Taylor series expansion in depth and cross-range operator. In order to resolve the divergence of Taylor approximation out of the interval of convergence, the rational filter of rectangular type is applied to the depth and cross-range operator. The use of the filter improves the stability of the solution but requires extra numerical burdens. Numerical issues involving the accuracy, efficiency, and stability of the proposed model are discussed and illustrated in an ocean wedge environment.

In this study, the three-dimensional Cartesian acoustic parabolic equation is fully solved in the stair-step representation of an ocean environment based on the split-step Padé algorithm. The sum representation of the Padé approximation of a full exponential operator in the split-step marching solution is used for parallel computing. A stabilized self-starter solution is derived in the same way. This solver is benchmarked against reference solutions for a rectangular duct propagation, an ideal wedge problem, and an axisymmetric ocean. Additionally, a solution for the Gaussian canyon problem is provided and compared with that obtained from the approximate parabolic equation solver, considering a leading-order cross-term.

In an active towed array sonar, it is important to find the optimal operation depth. Generally, the optimal depth can be chosen via numerical simulations for all sonar depths and this imposes great burdens of time and cost.In this paper, an efficient approach is proposed to find the optimal depth using the optimisation technique. First, the sonar performance function is newly defined as a measure of how well the active sonar might perform. This function depends on the properties of the ocean environment and the positions of sonar and underwater target. Then, the simulated annealing to find an optimal solution for maximising sonar performance is used. The optimised depth agrees well with the depth obtained from direct searching for all depths of source and receiver combinations, but its computational time is largely reduced.

Synthetic aperture sonar (SAS) is a technique that acquires an underwater image by synthesizing the signal received by the sonar as it moves. By forming a synthetic aperture, the sonar overcomes physical limitations and shows superior resolution when compared with use of a side-scan sonar, which is another technique for obtaining underwater images. Conventional SAS algorithms require a high concentration of sampling in the time and space domains according to Nyquist theory. Because conventional SAS algorithms go through matched filtering, side lobes are generated, resulting in deterioration of imaging performance. To overcome the shortcomings of conventional SAS algorithms, such as the low imaging performance and the requirement for high-level sampling, this paper proposes SAS algorithms applying compressive sensing (CS). SAS imaging algorithms applying CS were formulated for a single sensor and uniform line array and were verified through simulation and experimental data. The simulation showed better resolution than the ω-k algorithms, one of the representative conventional SAS algorithms, with minimal performance degradation by side lobes. The experimental data confirmed that the proposed method is superior and robust with respect to sensor loss.

The International Maritime Organization has recently updated the ship emission standards to reduce atmospheric contamination. One technique for reducing emissions involves using liquefied natural gas (LNG). The tanks used for the transport and storage of LNG must have very low thermal expansion and high cryogenic toughness. For excellent cryogenic properties, high-Mn steel with a complete austenitic structure is used to design these tanks. We aim to determine the optimum welding conditions for performing Laser-MIG (Metal Inert Gas) hybrid welding through the MIG leading and laser following processes. A welding speed of 100 cm/min was used for welding a 15 mm thick high-Mn steel plate. The welding performance was evaluated through mechanical property tests (tensile and yield strength, low-temperature impact, hardness) of the welded joints after performing the experiment. As a result, it was confirmed that the tensile strength was slightly less than 818.4 MPa, and the yield strength was 30% higher than base material. The low-temperature impact values were equal to or greater than 58 J at all locations in the weld zone. The hardness test confirmed that the hardness did not exceed 292 HV. The results of this study indicate that it is possible to use laser-MIG hybrid welding on thick high-Mn steel plates.

Inter-floor noise propagates through a building structure from a noise source to neighbors on other floors. Identification of type/position of inter-floor noise in a building is difficult for human hearing. A convolutional neural network-based inter-floor noise type/position classification method was proposed in [Appl. Sci. 9, 3735 (2019)] to identify inter-floor noise. The method was evaluated against inter-floor noise collected in a single campus building as a feasibility test. In this work, the generalizability of the method was addressed through numerous tasks using new datasets collected in two real apartment buildings. These datasets contain inter-floor noise generated in rooms and at positions with three-dimensional spatial diversity, which was not studied in the previous work. Furthermore, type classification knowledge transfer between two individual apartment building domains was studied.

Ray-based blind deconvolution (RBD) has demonstrated remarkable performance on several in situ underwater acoustic data. RBD estimates the channel impulse response by compensating the source phase estimated by beamforming. However, when beamforming fails to achieve the desired signal-to-noise ratio, RBD fails to achieve the expected performance. To address this issue, we propose an improved RBD framework that uses the kurtosis of the channel impulse response to provide additional phase correction after beamforming. The proposed method is validated with the measurement data from the shallow-water acoustic variability experiment 2015 using a high-frequency (11–31 kHz) broadband source and a vertical line array of 16 sensors located approximately 3 km away. The results are compared with conventional RBD and discussed from an optimization point of view.

Noise induced by incipient-propeller tip vortex cavitation (TVC) has a few sources near the propeller tips, which radiate a broadband signal. This article describes a compressive sensing (CS)-based TVC localization technique for coherent multiple-frequency processing, which jointly processes the measured data at multiple frequencies. Block-sparse CS, which groups several single-frequency measurements into blocks, is adopted for coherent multiple-frequency processing. The coherent multiple-frequency processing improves localization performance over that of single-frequency processing. Unlike single-frequency processing using conventional CS, which combines independent single-frequency measurement treatments by averaging, coherent multiple-frequency processing produces accurate localization without requiring a sufficient number of treated frequencies, long-time-sampled data with a time-invariant signal assumption, or even a single cavitation event. The approach is demonstrated on experimental data from a transducer source experiment and a cavitation source experiment.

For a sound field observed on a planar sensor array, compressive sensing (CS) reconstructs the two-dimensional (2D) direction-of-arrival (DOA) of multiple sources using a sparsity constraint. Conventional compressive beamforming methods suffer from grid mismatch, where true DOAs do not fall on the discretized angular search grid. This paper adopts a CS-based model, which can reconstruct block-sparse signals, and the model treats DOAs and the off-grid DOA compensation parts as blocks to deal with the off-grid 2D beamforming. The method is illustrated by numerical simulations and shows high estimation accuracy. Also, the approach does not require a specific array configuration and is suitable for arbitrary planar array geometry, which is practically useful. Since propeller tip vortex cavitation induces noise sources located sparsely near the propeller tip, the high-resolution of the method is demonstrated with experimental data from cavitation tunnel experiments.

In wire arc additive manufacturing (AM), as in arc welding, arc heat thermally deforms substrates and articles. For industrial applications, deformation characteristics of various materials must be understood and appropriate materials and methods of reducing deformation must be devised. Therefore, angular distortions of different materials were investigated through bead-on-plate welding and finite element analysis. A model that simplifies temperature-dependent properties was developed to establish relationships between thermomechanical properties and angular distortion. A simplified model of temperature-dependent properties was used, and angular distortion characteristics were extensively investigated for different material properties and heat inputs. Coefficient of thermal expansion, density, and specific heat all notably affected angular distortion depending on heat input conditions. Results showed that during wire arc AM, flatness of both substrates and articles could vary depending on material properties, heat input, substrate thickness, and bead accumulation. Study findings can provide insight into deformation characteristics of new materials and how to mitigate thermal distortions.

Heat straightening of welding-induced distortion, which has mostly relied on shipyard experience, was systemized for computation. The process of geometrically flattening a curved surface with fixed edges was formulated step by step based on thermal forming mechanisms. The proposed algorithm divides the measured surface into patches, connects them with virtual springs, and optimally arranges the patches on a plane in a direction that minimizes the potential energy within the weld lines. The overlaps between the geometrically calculated patches become the amount to shrink. The relationship between spot heating distance, curvature, and shrinkage was investigated through experiments and thermal elastic-plastic analysis to convert shrinkages into process parameters. It showed that the plastic shrinkage increases as the spot heating distance narrows, but the rate decreases. Also, the plastic shrinkage keeps constant regardless of the curvature, whereas the total shrinkage decreases as the curvature decreases. For validation of the proposed methodology, flame straightening experiments were carried out. It has been shown that the case using the plastic shrinkage produced fewer heating spots with higher efficiency than that using the total shrinkage.

In order to assess the damage detectability of the structural health monitoring method using diffuse fields, a laboratory experiment is conducted on an aluminum plate where two accelerometers are mounted as receivers. Two types of damages, a nonlinear material attachment and a punctuated hole, are considered. A hand-held impact hammer is used for the excitation, whereby the hammer is moved over grid points drawn on the aluminum plate, and thus the diffuse fields are generated by superposing the wave fields by many excitations randomly sampled. From the cross-correlation of diffuse fields between two receivers, we extract the coherent wave field in cases with and without damages. To detect the damage, a novel damage detection algorithm using a support vector machine is suggested based on the reduced features, transformed from several statistical parameters of damaged and undamaged noise cross-correlation functions, aided by the principal component analysis. The performance of the proposed algorithm is analyzed for the number of sources and damage types.

Compressive sensing (CS) in acoustics has received significant attention in the last decade, and thus motivates this special issue. CS emerged from the signal processing and applied math community and has since generated compelling results in acoustics. This special issue primarily addresses the acoustics CS topics of compressive beamforming and holography. For a sound field observed on a sensor array, CS reconstructs the direction of arrival of multiple sources using a sparsity constraint. Similarly, in holography a sparsity constraint gives improved sound field reconstruction over conventional l2-regularization. Other topics in this issue include sparse array configurations (as co-arrays) and sparse sensing in acoustic communication.

The tip vortex cavitation (TVC) noise of marine propellers is of interest due to the environmental impacts from commercial ships as well as for the survivability of naval ships. Due to complicated flow and noise field around a marine propeller, a theoretical approach to the estimation of TVC noise is practically unrealizable. Thus, estimation of prototype TVC noise level is realized through extrapolation of the model TVC noise level measured in a cavitation tunnel. In this study, for the prediction of prototype TVC noise level from a model test, a novel scaling law reflecting the physical basis of TVC is derived from the Rayleigh-Plesset equation, the Rankine vortex model, the lifting surface theory, and other physical assumptions. Model and prototype noise data were provided by Samsung Heavy Industries (SHI) for verification. In applying the novel scaling law, similitude of the spectra of nuclei is applied to assume the same nuclei distribution in the tip vortex line of the model and the prototype. It was found that the prototype TVC noise level predicted by the novel scaling law has better agreement with the prototype TVC noise measurement than the prototype TVC noise level predicted by the modified ITTC noise estimation rule.

This paper proposes a roll-angle estimation method for the triplet towed array using the oceanic ambient noise recorded in the mid-frequency range of 1 to 10 kHz. Here, the dominance of oceanic surface-generated noises such as wind-driven and shipping noises is expected. The ratio between the approximated phase parts of the noise-coherence function regarding the triplet hydrophones is used for the roll estimation. Advantage of this technique is that explicit input regarding the oceanic environment is not required, especially for an ocean with an azimuthally uniform noise field or dominant noise sources at the end-fire direction of the horizontal array. In other cases, a roll-angle bias occurs depending on the horizontal directionality. This proposed technique is applied to ocean-ambient-noise measurements that were collected in the East Sea of Korea during August 2015. The roll angles of the triplet array are estimated from the measurements and compared with roll-sensor-measured data. The estimated roll angles agree with the measured ones within the root mean square error of 10°, and their difference tends to decrease with increasing wind speed.

This paper describes a time delay estimation (TDE) technique using compressive sensing (CS) off the grid, which estimates the channel impulse response in a continuous time domain. The TDE can be formulated into a sparse signal reconstruction problem where the CS technique can be applied. Previous works have used standard finite dimensional CS with evenly discretized grids. However, the actual time delays will not always lie on the discrete grid, and this mismatch between the actual and discretized time delays results in reconstruction degradation. To overcome the basis mismatch, a TDE technique using an off the grid CS framework is proposed by modifying the scheme in the off the grid direction of arrival (DOA) estimation [Xenaki and Gerstoft, J. Acoust. Soc. Am. 137(4), 1923-1935 (2015)]. The effectiveness of the suggested method is demonstrated on real data from a water tank experiment. The off the grid CS TDE is shown to have super-resolution, which enables close arrivals to be distinguished.

The time-domain Helmholtz-Kirchhoff (H-K) integral for surface scattering is derived for a refractive medium, which can handle shadowing effects. The starting point is the H-K integral in the frequency domain. In the high-frequency limit, the Green's function can be calculated by ray theory, while the normal derivative of the incident pressure from a point source is formulated using the ray geometry and ray-based Green's function. For a corrugated pressure-release surface, a stationary phase approximation can be applied to the H-K integral, reducing the surface integral to a line integral. Finally, a computationally-efficient, time-domain H-K integral is derived using an inverse Fourier transform. A broadband signal scattered from a sinusoidal surface in an upwardly refracting medium is evaluated with and without geometric shadow corrections, and compared to the result from a conventional ray model.

In this paper, a ray-based noise model is developed to calculate the two-point spatial coherence function for ocean-surface-generated noise in a hydrophone triplet, which consists of three hydrophones arranged as an equilateral triangle with side length $d$ . While existing ray-based models employ the semianalytic approach, this model is more generally formulated for the 3-D noise field with vertical and azimuthal directivity using a ray-tracing model. Moreover, we derive a Maclaurin series of the formulated spatial coherence function for the ratio of the hydrophone spacing and the acoustic wavelength $\lambda $ . This series explicitly reveals the feature of spatial coherence for the noise field as a function of the orientation of the hydrophone pair and the ocean environmental coefficients. At small $d/\lambda $ , the leading-order terms alone provide a good approximate solution for the spatial coherence function, which enables rapid computation of six coherence functions for the hydrophone triplet. Especially, for the ocean environment with weakly azimuth-dependent ocean bathymetry, the contribution from each term in the derived series is analyzed and a simpler formula for the spatial coherence function is suggested.

The relation between high-frequency broadband acoustic signal variability and two types of internal waves (short-period internal solitary waves; ISWs, and semidiurnal internal tides; ITs) is investigated using data collected during the shallow-water acoustic variability experiment 2015 in the northeastern East China Sea. In this flat (∼100 m depth) region, an underwater sound channel with sound speed profile (SSP) variability observed during the experiment significantly affects the acoustic variability induced by the ISW, and the arrival structure of the channel impulse response (CIR) modeled by ray tracing. To model the range-dependent SSP due to ISW, the location and characteristics of the mode-1 ISW of wavelength (0.5–1 km) are estimated and verified based on the two-layer Korteweq–de Vries theory and by analyzing the observed temperature fluctuations. It is found from comparison between the measured and modeled CIRs that the ISW scatters the arrival structures of refracted rays. Meanwhile, semidiurnal ITs change the channel size modeled as range-independent considering the wavelengths (15–40 km) longer than the model range (3 km). Higher centroid of acoustic arrival time is found with lower isotherm depressions owing to the multimode ITs, indicative of acoustic energy focusing at the lower channel region.

The use of accelerometer signals for early recognition of severe slugging is investigated in a pipeline-riser system conveying an air&ndash, water two-phase flow, where six accelerometers are installed from the bottom to the top of the riser. Twelve different environmental conditions are produced by changing water and gas superficial velocities, of which three conditions are stable states and the other conditions are related to severe slugging. For online recognition, simple parameters using statistics and linear prediction coefficients are employed to extract useful features. Binary classification to recognize stable flow and severe slugging is performed using a support vector machine and a neural network. In multiclass classification, the neural network is adopted to identify four flow patterns of stable state, two types of severe slugging, and an irregular transition state between severe slugging and dual-frequency severe slugging. The performance is compared and analyzed according to the signal length for three cases of sensor location: six accelerometers, one accelerometer at the riser base, and one accelerometer at the top of the riser.

A grid-free compressive sensing (CS) based method for extracting the normal modes of acoustic propagation in the ocean waveguide from vertical line array (VLA) data is presented. Extracting the normal modes involves the estimation of mode horizontal wavenumbers and the corresponding mode shapes. Sparse representation of the waveguide propagation using modes at discrete horizontal wavenumbers enables CS to be applied. Grid-free CS, based on group total-variation norm minimization, is adopted to mitigate the issues of the wavenumber search grid discretization in the conventional CS. In addition, the suggested method can process multiple sensor data jointly, which improves performance in estimation over single sensor data processing. The method here uses data on a VLA from a source at several ranges, and processes the multiple sensor data at different depths jointly. The grid-free CS extracts the mode wavenumbers and shapes even with no a priori environmental knowledge, a partial water column spanning array data, and without the mode orthogonality condition. The approach is illustrated by numerical simulations and experimental SWellEx-96 (shallow water evaluation cell experiment 1996) data.

The acoustic radiation efficiency of a cylindrical shell is an important parameter in estimating the sound radiation emitted by a structure. However, many of the relevant previous studies have approached acoustic radiation problem by employing numerical methods, requiring large computational resources. In this study, an analytical method is developed to calculate the acoustic radiation efficiency of a finite-length, submerged, periodically stiffened cylindrical shell excited by a radial line force. Using the Flugge shell equations and Helmholtz equation, the motions of the shell wall and the acoustic pressure field in the outer fluid can be expressed analytically. Furthermore, to investigate the effects of finite length on the acoustic radiation efficiency, a stiffened cylindrical shell of infinite length is assumed to have a finite-length vibration distribution and to be constrained by simply supported boundary conditions. Numerical simulations are performed to investigate the acoustic radiation efficiency’s dependency on the finite-length vibration and the presence of stiffeners.

The prediction of underwater noise emissions from impact pile driving during near-shore and offshore construction activities and its potential effect on the marine environment has been a major field of research for several years. A number of different modeling approaches have been suggested recently to predict the radiated sound pressure at different distances and depths from a driven pile. As there are no closed-form analytical solutions for this complex class of problems and for a lack of publicly available measurement data, the need for a benchmark case arises to compare the different approaches. Such a benchmark case was set up by the Institute of Modelling and Computation, Hamburg University of Technology (Hamburg, Germany) and the Organisation for Applied Scientific Research (TNO, The Hague, The Netherlands). Research groups from all over the world, who are involved in modeling sound emissions from offshore pile driving, were invited to contribute to the first so-called COMPILE (a portmanteau combining computation, comparison, and pile) workshop in Hamburg in June 2014. In this paper, the benchmark case is presented, alongside an overview of the seven models and the associated results contributed by the research groups from six different countries. The modeling results from the workshop are discussed, exhibiting a remarkable consistency in the provided levels out to several tens of kilometers. Additionally, possible future benchmark case extensions are proposed.

When a sound wave propagates in a water medium bounded by a smooth surface wave, reflection from a wave crest can lead to focusing and result in rapid variation of the received waveform as the surface wave moves [Tindle, Deane, and Preisig, J. Acoust. Soc. Am. 125, 66-72 (2009)]. In prior work, propagation paths have been constrained to be in a plane parallel to the direction of corrugated surface waves, i.e., a two-dimensional (2-D) propagation problem. In this paper, the azimuthal dependence of sound propagation as a three-dimensional (3-D) problem is investigated using an efficient, time-domain Helmholtz-Kirchhoff integral formulation. When the source and receiver are in the plane orthogonal to the surface wave direction, the surface wave curvature vanishes in conventional 2-D treatments and the flat surface simply moves up and down, resulting in minimal temporal variation of the reflected signal intensity. On the other hand, the 3-D propagation analysis reveals that a focusing phenomenon occurs in the reflected signal due to the surface wave curvature formed along the orthogonal plane, i.e., out-of-plane scattering.

This paper presents a practical technique to evaluate the Percus–Yevick (PY) radial distribution function (RDF) for a water-saturated granular medium with a high volume fraction of grains. This is based on the weighted average of the PY RDF׳s recursive formula (Goodwin et al., 1992) and the PY integral equation׳s numerical solution using the filtered fast Fourier transform. In a highly dense medium, the combination of both models produces more accurate and stable results. The error analysis for the volume fraction of scatterers and the distance is given for validation. Finally, using the proposed PY RDF result, we solve an acoustic multiple-scattering equation with quasi-crystalline approximation (QCA) to obtain the effective bulk modulus and attenuation of the water-saturated granular medium, and confirm that our method provides a reasonable result.

Localization of high-frequency sources in underwater is a difficult problem because the propagation model is increasingly sensitive to environmental parameters as the frequency increases. A deep learning approach is proposed to estimate the depth of a high-frequency underwater source using a single hydrophone trained on real data. A residual neural network is trained by a spectrogram of measured signal and estimates the depth of the source as a regression problem. The method is applied to data collected during the shallow water acoustic variability experiment 2015 in the northeastern East China Sea and compared with the results of the frequency difference matched field processing which uses 16 sensors. It was also validated that the trained model can display the generalized results for the signals measured at other times or at similar settings but different source-receiver location. As a result, the neural network is able to more accurately estimate the depth of high-frequency source and shows the features found by the network from real data are effective for localization, while the model often fails to generate accurate replicas.

In this paper, a depth-bistatic bottom reverberation model that employs the ray theory is presented. The model can be applied to an active towed array in the ocean. The reverberation time series are modeled under the depth-bistatic assumption and their Doppler shift is calculated based on the actual source&ndash, receiver geometry. This model can handle N &times, 2D range-dependent bathymetry, the geometry of a triplet array, and the Doppler motion of the source, targets, and receiver. The model predictions are compared with the mid-frequency reverberation data measured by an active triplet towed array during August 2015 in the East Sea, Korea. These data are collected with a variable depth source at mid-frequency and the triplet line array in a deep-water environment. Model predictions of the beam time series and its spectrogram are in good agreement with the measurement. In particular, we discuss the effects of the source and receiver depths on the reverberation in deep water observed in both the measured and modeled results.

This paper suggests a method for source location and type classification of noise between floors at an apartment complex, which is a serious social conflict issue in Korea. Pre-trained convolutional neural networks proposed by visual geometry group is adapted and used for the task. A dataset for evaluation of method is generated and gathered in a building. The dataset is converted to log scaled mel-spectrograms to be fed into the input of the networks. The method is evaluated via k-fold cross validation. For comparison of performance depending on network architecture, convolutional neural networks suggested by Salamon and Bello [IEEE Signal Process. Lett. 24, 279–283 (2017)] is employed and validated. Also, the effectiveness of pre-training is measured.