Performance Enhancement of Drone Acoustic Source Localization Through Distributed Microphone Arrays.

This paper presents a microphone array sound source localization system based on deep learning algorithms. Currently, the most popular acoustic source localization algorithms are based on the traditional array signal processing methods. These methods have good localization performance in the ideal acoustic environment. However, the performances degraded significantly in low signal-to-noise ratio (SNR) and strong reverberation environments. To deal with this problem, this study developed a deep neural network (DNN) based system. Unlike the traditional algorithms with poor adaptability to environmental conditions, the proposed system can automatically learn the spatial information of sound sources under various conditions through training a large amount of data. Furthermore, it can fully utilize all the information of the original data without additional feature extraction. A set of experiments are carried out to evaluate the performance of the proposed system in comparison with the generalized cross correlation phase transform (GCC-PHAT) method. Results verify that the DNN based system achieves higher accuracy under low SNR conditions.

Mixed source localization plays an important role in passive radars. Aiming at the problem of low accuracy via phase difference method under a uniform circular array, this paper proposes a matrix differencing method for mixed far-field and near-field source localization. First, a two-dimensional MUltiple Signal Classification (MUSIC) method was utilized to estimate the azimuth and elevation angles of far-field sources. Thereafter, the covariance matrix difference method was exploited to extract the difference matrix of near-field sources. The azimuth and elevation angles of the far-field sources were estimated using the Estimation of Signal Parameters via Rotational Invariance Techniques-like (ESPRIT-like) method. Furthermore, the distance of the near-field sources was obtained by the one-dimensional MUSIC method. Finally, simulations were performed to verify the performance of the proposed algorithm. The proposed algorithm could effectively identify the mixed source when the two-dimensional Direction-Of-Arrival (DOA) of the far-field and near-field sources were the same. Moreover, the proposed algorithm could improve the accuracy of the mixed source parameter estimation. Results show that when the signal-to-noise ratio was set to 20 dB, the 2-D DOA estimation error of the near-field source was approximately 0.01°, and the distance error of the near-field source was approximately 0.1 m.

Sound source localization based on discrimination of cross-correlation functions

This study presents an application of a feedforward neural network (NN) structure to the bearing estimation problem. Using N snapshots from M sensors, the NN estimates the sensor-to-sensor propagation delays, which yield the far-field source location. The proposed network has only one output, which is the direction-of-arrival (DOA) angle. Thus, the network does not require any preprocessing. The NN buffers the sensor data, treats them as multidimensional delayed patterns and gives the location of a sinusoidal signal source in a noisy environment as output. Networks with various hidden nodes are tried with various sensor and snapshot numbers to find the best performance network structure. The effect of intersensor spacing on the performance is investigated. Using the best performance giving structure, the network is trained with various signal to noise ratios (SNRs) and then tested for various SNR levels.

In this paper, we address the multiple sound source localization problem using time differences of arrival (TDOAs) of sound sources to a microphone array. Typically, TDOAs are estimated based on the peak extraction of the generalized crosscorrelation function. In multi-source cases, for any given microphone pair, it is hard to tell the correspondence between the sound sources and the extracted peaks. In this work, we develop a novel localization approach based on data association which combines multiple TDOAs from the same source across different microphone pairs. Firstly, the generalized cross correlation-phase transform (GCC-PHAT) function is evaluated and multiple peaks of the GCC function indicating candidate TDOAs are extracted for each pair of microphones. Next, we employ the multi-dimensional assignment algorithm to associate multiple TDOAs from the same source. Finally, multiple sound source localization is carried out based on the obtained TDOA associations across different microphone pairs. Experimental results show the proposed method achieves superior performance for multiple sound source localization compared to the competing algorithm, especially in noisy environments.

Joint estimation of azimuth and elevation angles is of great importance in source localization and channel characterization. Firstly, some basic knowledge of three typical parametric estimation algorithms are introduced in this paper, i.e., multiple signal classification (MUSIC), Unitary estimation of signal parameter via rotational invariance technique (ESPRIT), and space-alternating generalized expectation-maximization (SAGE) algorithms. Each algorithm is capable of extracting both azimuth angle of arrival (AAoA) and elevation angle of arrival (EAoA) of multi-paths jointly. It is pointed out that the SAGE and MUSIC algorithms have higher complexity than the Unitary ESPRIT algorithm due to the iteration/angle searching procedure. Secondly, impacts of antenna number, closely spaced paths, and signal-to-noise ratio (SNR) on estimation performance of three algorithms are analyzed. Results show that the Unitary ESPRIT algorithm has lower accuracy in comparison with the MUSIC and SAGE algorithms when antenna number and SNR are large. Finally, three algorithms are applied to estimate multipath parameters in 16 GHz massive MIMO channel measurements. It is shown that the Unitary ESPRIT algorithm performs less satisfactory in MPCs extraction, while MUSIC can provide comparable results with the SAGE algorithm.

소리를 이용한 위치 추적은 마이크로폰을 이용하여 신호를 수집하고 수집된 신호로 부터 마이크로폰 간의 신호 도달 시간차를 추정한 뒤 추정된 시간차를 이용하여 소리의 발생 위치를 추정하는 과정을 거치게 된다. 실내 환경에서 이를 활용하기 위해서는 잡음과 반향음에 대한 강건성을 확보해야만 하는 제약이 따른다. 특히 실시간으로 구현하기 위해서는 계산의 효율성까지 고려되어야 한다. 본 논문에서는 네 개의 저가 콘덴서 마이크로폰을 이용하여 비용적인 측면과 계산량에서의 효율성을 모두 추구하였다. 네 개의 마이크로폰을 이용하여 마이크로폰 간의 소리 도달 시간차를 구하는 계산량을 줄였고 GCC-PHAT(Generalized Cross Correlation-Phase Transform) 알고리즘을 이용해서 강건성을 높였으며 iterative least square 방식을 이용하여 높은 정확도의 위치 데이터를 얻을 수 있었다. Sound localization algorithm usually adapts three step process: sampling sound signals, estimating time difference of arrival between microphones, estimate location of sound source. To apply this process in indoor environment, sound localization algorithm must be strong enough against reverberant and noisy condition. Additionally, calculation efficiency must be considered in implementing real-time sound localization system. To implement real-time robust sound localization system we adapt four low cost condenser microphones which reduce the cost and total calculation load. And to get TDOA(Time Differences of Arrival) of microphones we adapt GCC-PHAT(Generalized Cross Correlation-Phase Transform) which is robust algorithm to the reverberant and noise environment. The position of sound source was calculated by using iterative least square algorithm which produce highly accurate position data.

The probability density function (pdf) of an electromyography (EMG) signal provides useful information for choosing an appropriate feature extraction technique. The pdf is influenced by many factors, including the level of contraction force, muscle type, and noise. In this paper, we investigated the pdfs of noisy EMG signals artificially contaminated with five different noise types: 1) Electrocardiography (ECG) interference; 2) many spurious background spikes; 3) white Gaussian noise; 4) motion artifact; and 5) power line interference at various levels of signal-to-noise ratio (SNR). In addition, we evaluated a set of statistical descriptors for identifying a noisy EMG signal from its pdf, specifically kurtosis, negentropy, L-kurtosis, and robust measures of kurtosis (KR1 and KR2). The results show that at low SNR (<5 dB), all noise types affect the statistical descriptors for the pdf of a noisy EMG signal. In addition, KR2 performs the best among these descriptors in identifying a noisy EMG signal from its pdf, because it is computed based on the quantiles of the data. As a result, it can avoid the effects of outliers resulting in the correct identification of pdf shape of noisy EMGs with all contamination types and all levels of SNR.

A microphone clustering and back propagation (BP) neural network based acoustic source localization method using distributed microphone arrays in an intelligent meeting room is proposed. In the proposed method, a novel clustering algorithm is first used to divide all microphones into several clusters where each one corresponds to a specified BP network. Afterwards, the energy-based cluster selecting scheme is applied to select clusters which are small and close to the source. In each chosen cluster, the time difference of arrival of each microphone pair is estimated, and then all estimated time delays act as input of the corresponding BP network for position estimation. Finally, all estimated positions from the chosen clusters are fused for global position estimation. Only subsets rather than all the microphones are responsible for acoustic source localization, which leads to less computational cost; moreover, the local estimation in each selected cluster can be processed in parallel, which expects to improve the localization speed potentially. Simulation results from comparison with other related localization approaches confirm the validity of the proposed method.

During the last decade, distributed microphone arrays have been proposed in order to increase accuracy and spatial coverage of speaker localization systems operating in large and reverberant rooms. In principle, the framework provided by a distributed microphone network can also be applied effectively when using Blind Source Separation (BSS). Separation is commonly performed by processing the signals sampled at closely spaced microphones in a single adaptation step, for example by means of Independent Component Analysis (ICA). When the microphone spacing or the distance between source and microphones increase, the separation performance reduces due to spatial aliasing effects and to a reduced spatial coherence at microphones. In this paper we propose a new method, here referred to as Cooperative Wiener ICA (CW-ICA), which is able to apply BSS to signals acquired by a network of distributed microphone arrays. Different ICA adaptations are applied to the signals recorded by each array and are interconnected in order to constrain each adaptation to converge to a solution related to the same physical interpretation. A preliminary analysis on a network of two arrays shows that the proposed method can be applied successfully to source separation and localization tasks.

Passive source localization plays a critical role in understanding subsurface stress dynamics during hydraulic fracturing, assessing seismic risks, and exploring mineral resources. However, existing passive source localization methods often face challenges such as high computational costs, reliance on high‐quality data, and the need for large‐scale training data sets, which limits their practical application. To overcome these limitations, we propose a physics‐constrained deep operator network model (PhysLoc‐DeepONet) designed to address performance degradation in scenarios with sparse or low‐quality data. The model integrates physical constraints with deep learning techniques to achieve efficient and accurate source localization using limited seismic data. Compared with existing methods, PhysLoc‐DeepONet accounts for data sparsity in its network architecture and introduces innovations in training data set construction and loss function optimization, thereby maintaining high localization accuracy even under limited data conditions. Multiple numerical simulations and field tests are conducted based on microseismic monitoring scenarios to validate the effectiveness of the proposed method. Experimental results demonstrate that the model not only achieves fast ( s), high‐precision source localization in complex geological environments but also exhibits strong adaptability to various monitoring conditions and data quality levels. Further analysis indicates that incorporating additional physical information and data constraints can further enhance localization accuracy, yielding satisfactory results even under resource‐limited conditions. With its solid theoretical foundation and broad applicability, PhysLoc‐DeepONet enables efficient, accurate, and cost‐effective real‐time source localization, showing strong promise for practical implementation.

A data complementary method for thunderstorm point charge localization based on atmospheric electric field apparatus array group

The orthogonal imaging method is commonly used for source localization in brachytherapy. In some cases, however, difficulty is encountered in determining the dummy sources because of the presence of either contrast materials or bony structures. We here offer a novel method for source localization utilizing a dual-energy, radiographic technique. In this approach, two sets of orthogonal radiographic images (anterior-posterior and lateral views) are obtained using two different x-ray energies. Image processing (i.e., subtraction between two image sets) is carried out to enhance the source image. In a study performed using a laboratory developed pelvic phantom, it was demonstrated that the dual-energy method could significantly enhance the image quality of the dummy sources, and improve the achievable precision and relative accuracy in localization of source positions. When directly combined with digital imaging modalities, the dual-energy method can be a useful technique to improve the accuracy in brachytherapy source localization from planar radiographs.

Previous work demonstrates the importance of a high signal to noise ratio (SNR) when using transient evoked otoacoustic emissions (TEOAEs) to assay the medial olivocochlear reflex (MOCR). Increasing stimulus level provides one means to increase TEOAE SNR. However, this may come at the expense of a smaller MOCR effect. It is not clear whether the gain in SNR associated with the use of higher stimulus levels outweighs the disadvantage of a potentially smaller MOCR effect. The present study investigated the strength and detectability of the MOCR when assayed using TEOAEs at different stimulus levels. The hypothesis was that although the strength of the MOCR decreases with increasing stimulus level, the occurrence of statistically significant MOCR effects increases due to an increase in TEOAE SNR. Twenty-five young adult females with normal hearing participated in the study. TEOAEs were measured in the right ear with and without broadband noise presented in the left ear. The strength of the MOCR was quantified as the percent difference in the TEOAE between the contralateral noise and quiet conditions. Statistical bootstrapping was used to detect significant MOCR effects in individual subjects across different frequency bands and stimulus levels. The relationship between a detectable MOCR (response variable) and frequency, stimulus level, TEOAE SNR, MOCR strength, and subject (predictor variables) was evaluated using generalized linear mixed-effect models. The number of statistically significant MOCR effects increased with stimulus level at all frequencies. Occurrence was highest for the 2-kHz TEOAE frequency band and lowest for the 4-kHz frequency band. The strength of the MOCR decreased with increasing click level. TEOAE SNR, MOCR strength, and stimulus level were significant predictors of a detectable MOCR: The likelihood of a detectable MOCR increased with TEOAE SNR, MOCR strength, and stimulus level. Despite a reduction in the strength of the MOCR with increasing stimulus level, the detectability of the MOCR increased. This is due, in part, to an increase in TEOAE SNR with stimulus level. For clinical implementation of TEOAE-based MOCR assays, achieving a high SNR is necessary to permit the detection of the MOCR in individual patients.

This paper proposes a new speech enhancement (SE) algorithm utilizing constraints to the Wiener gain function which is capable of working at 10 dB and lower signal-to-noise ratios (SNRs). The wavelet thresholded multitaper spectrum was taken as the clean spectrum for the constraints. The proposed algorithm was evaluated under eight types of noises and seven SNR levels in NOIZEUS database and was predicted by the composite measures and the SNRLOSS measure to improve subjective quality and speech intelligibility in various noisy environments. Comparisons with two other algorithms (KLT and wavelet thresholding (WT)) demonstrate that in terms of signal distortion, overall quality, and the SNRLOSS measure, our proposed constrained SE algorithm outperforms the KLT and WT schemes for most conditions considered.