Sound Source Localization Techniques Research Articles

Three-dimensional sound source localization system based on fiber optic sensor array with an adaptive algorithm

Sound Source Localization (SSL) based on microphone arrays nowadays is widely used in the fields of machinery rattle localization, weapon navigation, and robot hearing. However, there is also a need for SSL in some special environments, such as long-distance gas pipelines where there is a risk of leakage. In these environments, arrays built with ordinary microphones can be a safety hazard, which can be compensated by the External Fabry-Perot interferometer (EFPI) fiber-optic sensor's advantages of being uncharged, inexpensive, and highly sensitive. To expand the application areas of microphone array-based SSL techniques, this paper using four EFPI fiber-optic sensors to constructs a 1 m × 1 m × 1 m sensor array, which is used for real-time localization of sound sources based on three-wavelength adaptive intensity demodulation methods and Time Difference Of Arrival (TDOA) localization techniques. An improved method is proposed for the three-wavelength adaptive intensity demodulation method to avoid the occurrence of signal flips. Finally, formulas are provided for this sensor array to quickly calculate the sound source position based on the arrival delay. Several experiments show that the system can achieve sound source localization over a wide range with an average error of less than 3 cm.

Enhanced approach to fusing automatic characteristic frequency extraction and adaptive array signals weighting for abnormal machine sound localization

In this paper, an enhanced approach for sound localization is proposed, which fuses automatic extraction of array signal characteristic frequencies and adaptive weighting. The method refines the autoregressive power spectral estimation algorithm and improves density-based spatial clustering of applications with noise algorithm for characteristic frequency extraction. Adaptive weighting technique is introduced to alleviate the problem of frequency mismatch in the localization process. The initial weight of narrowband signals is calculated and normalized using the frequency domain amplitude integration of narrowband signals, followed by adaptive threshold correction to eliminate invalid narrowband signal weights. The adaptive weight vector improves the localization method’s accuracy and interference suppression. The effectiveness and universality of the proposed method are demonstrated with test data from dry transformers and pumps, and its applicability is shown to extend to various spatial spectrum estimation algorithms and deep learning-based sound source localization techniques.

Sound source localization of harmonic sources in entire 3D space using just 5 acoustic signals

Sound Source Localization (SSL) has numerous applications namely machinery noise control, machinery fault localization, localization of explosives and weaponry in defense, development of hearing aids and robot’s auditory system. Beamforming and holographic SSL methods require numerous microphones and measurements for predicting source location; hence are highly expensive and time consuming. The simpler alternatives require less microphones; but are inaccurate, require high SNR, and have space of non-convergence where source cannot be uniquely localized. There is also a lack of method to localize pure harmonic sources. This paper proposes a method that addresses these limitations by accurately and quickly identifying the exact location of a harmonic sound source in the complete 3D space in acoustic free-field using just five acoustic signals. The novelties of this method are: new array design, spatially averaged Time Delay Estimation (TDE), TDE-ER (Energy Ratio) based localization, ER to remove mirror points, spectral de-noising, and solution of non-convergence. The method is rigorously tested through numerous simulations and real-life free-field experiments. A performance comparison shows that proposed method is more accurate (6.7% error) compared with the existing 3D SSL technique (11.7% error). Overall, it accurately identifies the coordinates of dominant sound source with 9.17% mean error, in presence of inherent sensor noise (14 dB SNR onwards) and background noise (8 dB SNR onwards) in complete 3D space. Interestingly, the sound source location, and sound source frequency is found to effect the consistency of TDE-based SSL methods. The presented work paves way for further advancement in 3D SSL.

Multiarray Eigenbeam-ESPRIT for 3D Sound Source Localization With Multiple Spherical Microphone Arrays

A 3D sound source localization (SSL) technique for multiple spherical arrays is proposed. Spherical arrays have been popularly used for direction of arrival (DoA) estimation, which serves as important prior knowledge for source separation, dereverberation, and binaural rendering for VR applications. In this work, a parametric 3D SSL technique is proposed that can detect multiple sources using multiple spherical array recordings. From the enriched information by using multiple arrays, we show that the parametric estimation of 3D positions is possible without scanning candidate positions in 3D space. To mitigate the problem of the ambiguous association between DoAs estimated from different arrays, a total covariance matrix including both auto- and cross-covariance matrices between arrays is utilized such that DoAs from arrays are automatically paired to uniquely localize 3D source positions. The joint parameter estimation is based on the generalized joint Schur decomposition of multiple matrices combined with a geometric projection robustly guiding the convergence of the proposed iteration algorithm. Simulations conducted in anechoic and reverberant room conditions reveal that the proposed technique can accurately determine multiple source positions in 3D space without ambiguities.

A Review on Sound Source Localization Systems

Sound Source Localization (SSL) is an important function of a robotic ear. Various techniques are proposed for sound source localization which can be categorized into four main types. Out of which one type is SSL based on conventional algorithms like Generalized Cross Co-relation (GCC), Multiple Signal Classification (MUSIC), Time Difference of Arrival (TDOA), etc. using multiple microphones array configurations. The second type involves techniques based on binaural signal processing using conventional algorithms. SSL techniques that fall under the third and fourth types of categories are developed recently in the last decade with the rise of Convolutional Neural Network (CNN) algorithms. The third type is the SSL technique which makes use of multiple array configurations using CNN and the fourth type is the SSL technique based on binaural signal processing using CNN. The different SSL techniques based on multiaural and binaural signals using the conventional algorithm as well as CNN are presented in this review. Binaural signals are based on the principle of the human auditory system which makes use of two ears, the left ear and the right ear. Binaural SSL techniques that use CNN have recently evolved. We present a short review of different SSL techniques that use a different number of microphones and the algorithms used to perform SSL. The review suggests that out of all binaural SSL techniques, CNN is the emerging and optimized one.

Wideband sparse Bayesian learning for off-grid binaural sound source localization

Some traditional binaural sound source localization (BSSL) techniques estimate the azimuth using measured head-related transfer function (HRTF) databases, which is constrained by the discretely measured azimuths of HRTF databases. The azimuth localization performance of these HRTF-based BSSL methods may degrade significantly when the true azimuth is not included in the discretely measured azimuths, which is a typical off-grid problem. This paper proposes an off-grid BSSL method based on an off-grid wideband sparse Bayesian learning algorithm. An off-grid binaural sparse signal model is established first, which takes into account both the shadowing effects by the head and the impacts of off-grid problem. Based on the spatial sparsity of sound sources, the off-grid BSSL problem can be reduced to a convex optimization problem. An off-grid wideband sparse Bayesian learning algorithm is further derived to solve the convex optimization problem and thus improve the localization performance. Experimental results demonstrate that the proposed off-grid BSSL method can achieve higher localization accuracy than the state-of-the-art HRTF-based BSSL methods in various acoustic environments, especially in the off-grid situations.

Fast and Robust RBF Neural Network Based on Global K-means Clustering with Adaptive Selection Radius for Sound Source Angle Estimation

The sound source localization technique is widely applied to target detection and localization. However, the application of conventional sound source localization methods is limited in actual environment because of estimation accuracy, computational complexity, and flexibility of the environment. In order to improve the sound source localization performance in actual environment, a fast and robust radial basis function (RBF) neural network based on global K-means clustering with adaptive selection radius is proposed in this paper. In the proposed method, an adaptive selection radius is calculated according to the population density sampling method to remove unnecessary points around cluster centers during the global K-means clustering; thus, compared with the conventional neural network, a fast optimization of hidden layer neuron parameters can be achieved. Afterward, the RBF neural network is trained to locate the sound source by solving nonlinear equations of the time difference of arrival and sound source location. Because of the adoption of adaptive selection radius in global K-means clustering, the proposed method can provide desirable performance with low computational complexity. Based on the simulated and actual experimental data, the proposed method is verified and an improved performance is achieved compared with that of conventional neural network sound source localization methods.

A New Regional Localization Method for Indoor Sound Source Based on Convolutional Neural Networks

At present, the sound source localization methods based on microphone arrays can be roughly classified into three categories: the controllable beamforming technology based on a maximum output power, the high-resolution spectrogram estimation technique, and the sound source localization technique based on time difference of sound. However, an existing localization technology in unstructured indoor environment lacks of localization accuracy and adaptability. In some practical situations, the location of sound source is limited to predefined areas. In this paper, we propose a research method of source region location system based on convolutional neural networks (CNNs). Based on the characteristics of weighted values of CNN, we realize the regional of indoor single sound sources transforming the sound source signals into grammar diagrams and then inputting them into the CNN. The whole process is based on the characteristics of weighted values of CNN. Finally, this paper completes the training and testing for CNN by using the Tensorflow framework. Simulation experiments on the test sets show the effectiveness of the proposed method.

Localization of sound sources in robotics: A review

Sound source localization (SSL) in a robotic platform has been essential in the overall scheme of robot audition. It allows a robot to locate a sound source by sound alone. It has an important impact on other robot audition modules, such as source separation, and it enriches human–robot interaction by complementing the robot’s perceptual capabilities. The main objective of this review is to thoroughly map the current state of the SSL field for the reader and provide a starting point to SSL in robotics. To this effect, we present: the evolution and historical context of SSL in robotics; an extensive review and classification of SSL techniques and popular tracking methodologies; different facets of SSL as well as its state-of-the-art; evaluation methodologies used for SSL; and a set of challenges and research motivations.

Sound source localization based on sparse estimation and convex clustering

Sound source localization techniques in array signal processing have been interested for many years. A lot of direction-of-arrival (DOA) estimation methods have been proposed. Most of them use plane waves as model. On the other hand, 3D sound source localization techniques are important for many applications. They are required to estimate the DOA and the distance of sound sources. In the case of the techniques targeted to estimate the distances of sound sources, monopoles must be used for the model. The positions of monopoles are candidates for the sound source positions. However, it is difficult to handle a large number of monopoles placed at various distances. Moreover, it is difficult to decide the number of sound sources to be estimated from the candidates. The proposed method uses sparse estimation with a monopole dictionary in frequency domain, as is well known that the effectiveness of sparse representation for DOA estimation has been in late years. After the sparse estimation, the proposed method ...

Sound source localization technique using a seismic streamer and its extension for whale localization during seismic surveys.

Marine seismic surveys are under increasing scrutiny because of concern that they may disturb or otherwise harm marine mammals and impede their communications. Most of the energy from seismic surveys is low frequency, so concerns are particularly focused on baleen whales. Extensive mitigation efforts accompany seismic surveys, including visual and acoustic monitoring, but the possibility remains that not all animals in an area can be observed and located. One potential way to improve mitigation efforts is to utilize the seismic hydrophone streamer to detect and locate calling baleen whales. This study describes a method to localize low frequency sound sources with data recoded by a streamer. Beamforming is used to estimate the angle of arriving energy relative to sub-arrays of the streamer which constrains the horizontal propagation velocity to each sub-array for a given trial location. A grid search method is then used to minimize the time residual for relative arrival times along the streamer estimated by cross correlation. Results from both simulation and experiment are shown and data from the marine mammal observers and the passive acoustic monitoring conducted simultaneously with the seismic survey are used to verify the analysis.

Baleen whale localization using hydrophone streamers during seismic reflection surveys

Seismic reflection surveys use acoustic energy to image the structure beneath the seafloor, but concern has been raised about their potential impact on marine animals. Most of the energy from seismic surveys is low frequency, so the concern about their impact is focused on Baleen whales that communicate in the same frequency range. To better mitigate against this impact, safety radii are established based on the criteria defined by the National Marine Fisheries Service. Marine mammal observers use visual and acoustic techniques to monitor safety radii during each experiment. However, additional acoustic monitoring, in particular, locating marine mammals, could demonstrate the effectiveness of the observations, and help us understand animal responses to seismic experiments. A novel sound source localization technique using a seismic streamer has been developed. Data from seismic reflection surveys conducted with the R/V Langseth are being analyzed with this method to locate baleen whales and verify the accuracy of visual detections during experiments. The streamer is 8 km long with 636 hydrophones sampled at 500 Hz. The work focuses on time intervals when only a mitigation gun is firing because of marine mammal sightings. [Sponsored by NSF.]

Inverse problem with beamforming regularization matrix applied to sound source localization in closed wind-tunnel using microphone array

Microphone arrays have become a standard technique to localize and quantify source in aeroacoustics. The simplest approach is the beamforming that provides noise source maps with large main lobe and strong side lobes at low frequency. Since a decade, the focus is set on deconvolution techniques such as DAMAS or Clean-SC. While the source map is clearly improved, these methods require a large computation time. In this paper, we propose a sound source localization technique based on an inverse problem with beamforming regularization matrix called Hybrid Method. With synthetic data, we show that the side lobes are removed and the main lobe is narrower. Moreover, if the sound noise source map provided by this method is used as input in the DAMAS process, the number of DAMAS iterations is highly reduced. The Hybrid Method is applied to experimental data obtained in a closed wind-tunnel. In both cases of acoustic or aeroacoustic data, the source is correctly detected. The proposed Hybrid Method is found simple to implement and the computation time is low if the number of scan points is reasonable.

Three dimensional passive acoustic spatial identification of nucleations in superheated droplet detectors

A new method for spatial localization of bubble nucleations in superheated droplet detectors is presented which has two steps: validation and localization. Validation is accomplished through signal processing techniques serving the purpose of filtering out electromagnetic noise and gaseous micro-leaks. The 3D spatial localization uses a passive acoustic sound source localization technique using a microphone array (5 elements) and a simple generalised cross-correlation (GCC) time delay of arrival (TDOA) algorithm. The approach for the nucleation validation is new and regarding the localization, as of this writing, we are the first to endeavour its feasibility assessment. Experimental results are shown.

Microphone array techniques using cross-correlations.

Civilian noise complaints and damage claims have created a need to establish a detailed record of impulse noise generated at military training facilities. Wind noise is causing false positive impulse detections in the current noise monitoring systems. Multiple channel data methods were investigated in order to distinguish the characteristics of noise events. A microphone array was used to collect four simultaneous channels of military impulse and wind noise data. Cross‐correlation functions were then used to characterize the input waveforms. Three different analyses of microphone array data were developed. A new value, the min peak correlation coefficient, is defined from the minimum value of peaks of the cross‐correlation coefficient functions among the different channels. This value is a measure of the likelihood that a given waveform originated from a correlated noise source. The angle of incidence of the noise source is calculated using a sound source localization technique based on the geometry of th...

Acoustic source localization with eye array

A novel signal-processing algorithm and array for sound source localization (SSL) in three-dimensional space is presented. This method, which has similarity to the eye in localization of light rays, consists of a novel hemispherical microphone array with 26 microphones on the shell and one microphone in the sphere center. The microphones on the shell map a geodesic hemisphere called two-frequency icosahedron; hence, each microphone has at least four other orthogonal microphones. A signal-processing scheme utilizes parallel creation of a special closeness function for each microphone direction on the shell in the time domain. Each closeness function cell (lens cell) consists of center microphone, shell microphone, and one of its orthogonals. The closeness function output values are linearly proportional to spatial angular difference between the sound source direction and each of the shell microphone directions. By choosing microphone directions corresponding to the highest closeness function values and implementing a linear weighted spatial averaging on them, the sound source direction is estimated. Contrary to traditional SSL techniques, this method is based on simple parallel mathematical calculations in the time domain with low computational costs. The laboratory implementation of the array and algorithm shows reasonable accuracy in a reverberant room.

Adaptive Microphone Array System with Two-Stage Adaptation Mode Controller

In this paper, an adaptive microphone array system with a two-stage adaptation mode controller (AMC) is proposed for high-quality speech acquisition in real environments. The proposed system includes an adaptive array algorithm, a time-delay estimator and a newly proposed AMC. To ensure proper adaptation of the adaptive array algorithm, the proposed AMC uses not only temporal information, but also spatial information. The proposed AMC is constructed with two processing stages: an initialization stage and a running stage. In the initialization stage, a sound source localization technique is adopted, and a signal correlation characteristic is used in the running stage. For the adaptive array algorithm, a generalized sidelobe canceller with an adaptive blocking matrix is used. The proposed algorithm is implemented as a real-time man-machine interface module of a home-agent robot. Simulation results show 13 dB SINR improvement with the speaker sitting 2 m distance from the home-agent robot. The speech recognition rate is also enhanced by 32% when compared to the single channel acquisition system.

Sound source localization in multimedia applications

With the arrival of real-time 3-D digital audio processing, techniques for acoustic space modeling and sound source spatialization have become of great interest for implementation in various multimedia applications, ranging from CD-ROM authoring to video conferencing. In surround sound applications, where sound source localization is a primary objective, various digital signal processing (DSP) techniques can be employed to heighten the sense of source localization, often creating a stronger link between sound and image. In several multimedia applications, the degree of live interactivity can play a crucial role in the success of the application. For example, the underlying software in video games, acoustic space simulators, and air traffic monitoring systems (using auditory display) must render audio objects in a real-time acoustically modeled space, based on varying position data supplied to these programs by the user(s). This paper will focus on sound source localization techniques and their use in various multimedia applications, and will examine key issues in implementation.