Estimated Source Signals Research Articles

Direction-of-Arrival Estimation of Circular and Noncircular Wideband Source Signals via Augmented Envelope Alignment

In practical applications of array system, the incident signals may come from either circular or noncircular sources, or both. The problem of direction-of-arrival (DOA) estimation is herein addressed for both circular and noncircular wideband source signals. Based on the concept of effective rank, the aligned augmented effective noise subspace, spanned by the subordinate eigenvectors of the aligned array output augmented covariance matrix, is constructed for each scanning angle. It is shown that the aligned augmented effective noise subspace corresponding to a true DOA of signal is: 1) approximately orthogonal to a frequency-invariant one-dimensional subspace spanned by the aligned augmented transfer vector of the signal in the presence of multiple strictly noncircular sources; and 2) approximately orthogonal to a frequency-invariant two-dimensional subspace spanned by the aligned augmented transfer matrix of the signal in the presence of multiple circular or partially noncircular sources. This enables the noncircularity exploited wideband DOA estimation based on rank-reduction and norm-nulling. The performance of the proposed method has been evaluated and compared with some existing popular wideband DOA estimators by extensive simulations.

Cross-Correlation-Based Identification of Rotating Sources

Noise-source identification using microphone arrays has become well-established in wind-tunnel testing. When the sources are moving, the accepted approach is first to “de-Dopplerize” the microphone signals, producing estimated source signals for subsequent analysis. A straightforward time-domain average of these signals is vulnerable to flow-noise contamination, and so an alternative approach is necessary for wind-tunnel applications. This method is equivalent to the basic, “diagonal-removal,” cross-spectral technique employed with fixed sources, but, unlike the fixed-source case, cannot be generalized to account simultaneously for multiple sources. The alternative presented here, which works with the time-domain cross-correlations between microphones, is not subject to this limitation. It is illustrated using simulated data from an example rotating-source configuration, and the results are compared against a de-Dopplerization-based analysis. For three equal-strength sources, the current method eliminates the contamination evident in de-Dopplerization estimates. Equally, it is able to locate and accurately characterize weak secondary sources that are masked by primary-source contributions in de-Dopplerization beamforming maps. Hence it extends the option of higher-fidelity wind-tunnel beamforming, currently only available for fixed sources, to rotating-source configurations.

Separating Multiple Moving Sources by Microphone Array Signals for Wayside Acoustic Fault Diagnosis

Abstract Wayside acoustic defective bearing detector (ADBD) system is a potential technique in ensuring the safety of traveling vehicles. However, Doppler distortion and multiple moving sources aliasing in the acquired acoustic signals decrease the accuracy of defective bearing fault diagnosis. Currently, the method of constructing time-frequency (TF) masks for source separation was limited by an empirical threshold setting. To overcome this limitation, this study proposed a dynamic Doppler multisource separation model and constructed a time domain-separating matrix (TDSM) to realize multiple moving sources separation in the time domain. The TDSM was designed with two steps of (1) constructing separating curves and time domain remapping matrix (TDRM) and (2) remapping each element of separating curves to its corresponding time according to the TDRM. Both TDSM and TDRM were driven by geometrical and motion parameters, which would be estimated by Doppler feature matching pursuit (DFMP) algorithm. After gaining the source components from the observed signals, correlation operation was carried out to estimate source signals. Moreover, fault diagnosis could be carried out by envelope spectrum analysis. Compared with the method of constructing TF masks, the proposed strategy could avoid setting thresholds empirically. Finally, the effectiveness of the proposed technique was validated by simulation and experimental cases. Results indicated the potential of this method for improving the performance of the ADBD system.

Underdetermined convolutive blind separation of sources integrating tensor factorization and expectation maximization

Underdetermined convolutive blind separation of sources is a challenging topic in speech and audio processing. In plenty of conventional separation algorithms, the source signal estimation is performed in the frequency-domain, thus leading to permutation problems and poor separation results. In this paper, to solve the permutation alignment and obtain better source separation results, we exploit the algebraic structure of tensor factorization model and expectation–maximization (EM) to provide a new time-frequency algorithm. This is because tensor factorization has an advantage to estimate the channel for the underdetermined convolutive mixture case, and EM algorithm is conducive to faster converge to the desired solution and better source separating property. In the proposed algorithm, we first estimate the mixing matrix by using tensor decomposition, while permutation alignment algorithm is used to deal with the permutation problems. Then, the model parameters are updated using EM algorithm for improving source separation performance. At the same time, the spatial images of source signals are obtained using Wiener filters constructed from the estimated parameters. Furthermore, the time-domain source signals can be obtained through inverse short-time Fourier transform. Finally, a series of simulations show that compared with some existing separation algorithms, the proposed algorithm achieves better separation performance.

Aligned Propagator Scanning Approach to DOA Estimation of Circular and Noncircular Wideband Source Signals

The problem addressed in this paper is focused on the direction-of-arrival (DOA) estimation of the circular and noncircular wideband signals. The propagator scanning scheme is applied to the array output augmented cross covariance matrix by aligning first the baseband array observation. When the scanning angle equals a true DOA of signal, the complex envelope of the corresponding signal is perfectly aligned, and then, the associated propagator can be used to fulfill subspace separation based wideband direction finding. This method involves no data focusing and eigendecomposition. The performance of the proposed method is analyzed in terms of the mean-square errors of the DOA estimates. Simulation results are included to evaluate the new technique and to verify the performance analysis.

Estimating source signatures from source-over-spread marine seismic data

Recent developments in marine seismic acquisition include deploying a source vessel above a towed-streamer spread. We have developed an inversion algorithm to estimate source signatures for such acquisition configurations, by minimizing the difference between the recorded and a modeled direct wave. The forward modeling is based upon a physical modeling of the air bubble created by each air gun in the source array, and a damped Gauss-Newton approach is used for the optimization. Typical inversion parameters are empirical damping factors for the bubble oscillations and firing time delays for each air gun. Variations in streamer depth are taken into account, and a constant sea-surface reflection coefficient is also estimated as a by-product of the inversion. For data acquired in shallow waters, we have developed an extension of the forward modeling to include reflections from the water bottom to stabilize the inversion. The algorithm is tested on synthetic- and field-data examples, and the estimated source signature for the field-data example is used in a designature processing flow.

Number of Source Signal Estimation by the Mean Squared Eigenvalue Error

Detection of the number of source signals (NoSS) in the presence of additive noise is considered. We present a new approach denoted by the mean squared eigenvalue error (MSEE). The MSEE is the mean squared error between the desired noise-free eigenvalues and the available estimated eigenvalues. The approach investigates and analyzes the probabilistic distribution of the available eigenvalue estimates and revisits proper thresholding of these sorted values. The optimum NoSS is provided by minimizing the MSEE. A probabilistic worst-case technique is proposed to estimate the value of the MSEE by using only the available data. It is shown that the proposed method is consistent as the data length increases. It is also shown that the method is consistent as the signal-to-noise ratio (SNR) increases. Simulation results illustrate advantages of the MSEE over competing approaches and confirm effectiveness and robustness of the MSEE even in low-SNR or small sample size scenarios.

Efficient blind source extraction of noisy mixture utilising a class of parallel linear predictor filters

This study presents a novel blind source extraction of a noisy mixture using a class of parallel linear predictor filters. Analysis of a noisy mixture equation is carried out to address new autoregressive source signal model based on the covariance matrix of the whitened data. A method of interchanging the rules of filter inputs is proposed such that this matrix becomes the filter input while the estimated source signals are considered as the parallel filter coefficients. As the matrix has unity norm and unity eigenvalues, the filter becomes independent on the mixture signal norm and eigenvalues variations, thus solving drastically the ambiguity due to the dependency of the filter on the mixture power levels if the mixture is considered as the filter input. Furthermore, the unity eigenvalues of the matrix result in a very fast convergence in two iterations. Simulation results show that the model is capable of extracting the unknown source signals and removing noise when the input signal-to-noise ratio is varied from −20 to 80 dB.

Frame-Based Variational Bayesian Learning for Independent or Dependent Source Separation.

Variational Bayesian (VB) learning has been successfully applied to instantaneous blind source separation. However, the traditional VB learning is restricted to the separation of independent source signals. Moreover, it has the difficulty to recover source signals with a sizable number of samples because of its rapidly increasing computational requirement. To overcome such shortcomings, frame-based VB (FVB) learning is proposed to address both independent and dependent source separation with a large number of samples in this paper. Specifically, a Gaussian process (GP) is employed to model independent or dependent source signals. To our knowledge, GP has been only used to model each of independent source signals. For dependent source signals, this paper proposes a novel modeling process: initial source signals are zigzag concatenated into a long serial and GP is then used to model it. In order to obtain a reliable covariance function for GP, first, we apply singular value decomposition to give initial estimated source signals and then we select an appropriate covariance function with which GP can perfectly fit them. In order to alleviate the computational burden of VB learning, we split observed signals into frames, and then model and infer source signals for each frame. Compared with the state-of-the-art algorithms, the experimental results show that the FVB learning has potential to provide improvement in separation performance not only for independent source signals but also for dependent ones, especially for long data records.

Joint Alignment of Multiple Point Sets with Batch and Incremental Expectation-Maximization.

This paper addresses the problem of registering multiple point sets. Solutions to this problem are often approximated by repeatedly solving for pairwise registration, which results in an uneven treatment of the sets forming a pair: a model set and a data set. The main drawback of this strategy is that the model set may contain noise and outliers, which negatively affects the estimation of the registration parameters. In contrast, the proposed formulation treats all the point sets on an equal footing. Indeed, all the points are drawn from a central Gaussian mixture, hence the registration is cast into a clustering problem. We formally derive batch and incremental EM algorithms that robustly estimate both the GMM parameters and the rotations and translations that optimally align the sets. Moreover, the mixture's means play the role of the registered set of points while the variances provide rich information about the contribution of each component to the alignment. We thoroughly test the proposed algorithms on simulated data and on challenging real data collected with range sensors. We compare them with several state-of-the-art algorithms, and we show their potential for surface reconstruction from depth data.

Cortical Source Localization and Signal Estimation Without Exact Knowledge of the Leadfield Matrix

This paper develops an iterative algorithm for cerebral activity localization and signal estimation when exact knowledge of the leadfield matrix is not available. The algorithm exploits a high resolution localization technique that can determine the number of sources and their locations. A beamformer is designed in order to estimate the dipole source amplitudes, and the EEG signal is reconstructed via a calibration matrix. The performance of the algorithm is quantified in terms of the localization error and the accuracy is validated by applying it to experimental data.

Distributed Blind Estimation Over Sensor Networks

Distributed estimation, where a set of nodes collaboratively estimate some parameters of interest from noisy measurements, has received much attention in both science and engineering. Recent studies mainly focus on distributed non-blind or training-based estimation, that is, a training signal and its desired output are both known to the distributed receivers. However, in some applications, it is physically difficult, if not impossible, to get a training signal in prior. Besides, the use of training signals consumes much channel bandwidth. So, it is more preferable for the receivers to perform distributed estimation without the assistance and expense of training sequences, i.e., distributed blind estimation . In this paper, the problem of distributed blind estimation over sensor networks is considered, and a kind of distributed diffusion generalized Sato algorithm is proposed to design a blind equalizer for channel equalization and source signal estimation. The stability of the proposed method in mean and mean-square senses is analyzed theoretically, and its performance is verified numerically by a series of simulations.

Variable Step-Size Method Based on a Reference Separation System for Source Separation

Traditional variable step-size methods are effective to solve the problem of choosing step-size in adaptive blind source separation process. But the initial setting of learning rate is vital, and the convergence speed is still low. This paper proposes a novel variable step-size method based on reference separation system for online blind source separation. The correlation between the estimated source signals and original source signals increases along with iteration. Therefore, we introduce a reference separation system to approximately estimate the correlation in terms of mean square error (MSE), which is utilized to update the step-size. The use of “minibatches” for the computation of MSE can reduce the complexity of the algorithm to some extent. Moreover, simulations demonstrate that the proposed method exhibits superior convergence and better steady-state performance over the fixed step-size method in the noise-free case, while converging faster than classical variable step-size methods in both stationary and nonstationary environments.

Blind Channel and Source Estimation in Networked Systems

In this paper, we study the blind channel and source estimation in sensor networks, where the channels are modeled by FIR filters and the source signal is deterministic. Distributed estimation algorithms for networked systems under noise-free and noisy measurements are developed, which blindly identify the multiple channels, followed by the source signal estimation. The key to the proposed algorithms lies in the adaptation of the blind system identification technique for the distributed channel estimation. In the presence of measurement noises, conventional blind identification methods cannot be straightforwardly realized in distributed environments. Instead, two stable distributed algorithms are introduced, which can avoid trivial solutions for the blind identification problem. Convergence properties of the proposed algorithms are provided, and simulation examples are given to show the performances of the proposed algorithms.

Speaker adaptive voice source modeling with applications to speech coding and processing

We discuss the use of low-dimensional physical models of the voice source for speech coding and processing applications. A class of waveform-adaptive dynamic glottal models and parameter identification procedures are illustrated. The model and the identification procedures are assessed by addressing signal transformations on recorded speech, achievable by fitting the model to the data, and then acting on the physically oriented parameters of the voice source. The class of models proposed provides in principle a tool for both the estimation of glottal source signals, and the encoding of the speech signal for transformation purposes. The application of this model to time stretching and to fundamental frequency control (pitch shifting) is also illustrated. The experiments show that copy synthesis is perceptually very similar to the target, and that time stretching and “pitch extrapolation” effects can be obtained by simple control strategies.

TWO UNIFORM LINEAR ARRAYS FOR NON-COHERENT AND COHERENT SOURCES FOR TWO DIMENSIONAL SOURCE LOCALIZATION

This paper presents a novel method for the two-dimensional direction of arrival (DOA) estimation based on QR decomposition. A conflguration with two uniform linear antenna arrays (ULAs) is employed for the joint estimation of elevation (µ) and azimuth (`) angles. Q data matrix will estimate the azimuth angle while R data matrix will estimate the elevation angle. The proposed method utilizes only a single snapshot of the received data and constructs a Toeplitz data matrix. This reduces the computational complexity of the proposed method to O((N + 1) 2 ) from O(N 3 ) for SVD based methods. The structure of the data matrix also favors the 2D DOA estimation for both coherent and non-coherent source signals. Simulation results are presented, and performance of the proposed method is compared with the Matrix Pencil method for 2D DOA estimation of multiple incident source signals.

Wiener estimation of the Green’s function

I consider the problem of finding the impulse response, or Green’s function, from a measured response including noise, given an estimate of the source time function. This process is usually known as signature deconvolution. Classical signature deconvolution provides no measure of the quality of the result and does not separate signal from noise. Recovery of the earth impulse response is here formulated as the calculation of a Wiener filter in which the estimated source signature is the input and the measured response is the desired output. Convolution of this filter with the estimated source signature is the part of the measured response that is correlated with the estimated signature. Subtraction of the correlated part from the measured response yields the estimated noise, or the uncorrelated part. The fraction of energy not contained in this uncorrelated component is defined as the quality of the filter. If the estimated source signature contains errors, the estimated earth impulse response is incomplete, and the estimated noise contains signal, recognizable as trace-to-trace correlation. The method can be applied to many types of geophysical data, including earthquake seismic data, exploration seismic data, and controlled source electromagnetic data; it is illustrated here with examples of marine seismic and marine transient electromagnetic data.

Blind Source Separation by Nuclear Norm Minimization and Local Recoverability Analysis

We propose a new blind source separation (BSS) algorithm that is effective when Hankel matrices constructed from individual source signals are near low-rank and satisfy a certain near-orthogonality condition. Source separation is achieved by finding a nonsingular reverse-mixing operation that minimizes nuclear norms of Hankel matrices constructed from estimated source signals. The new formulation results in a non-convex optimization problem involving a reverse-mixing matrix. Preliminary analysis of local recoverability of source signals as well as few numerical simulations are presented in this letter.

An iterative model-based approach to cochannel speech separation

Cochannel speech separation aims to separate two speech signals from a single mixture. In a supervised scenario, the identities of two speakers are given, and current methods use pre-trained speaker models for separation. One issue in model-based methods is the mismatch between training and test signal levels. We propose an iterative algorithm to adapt speaker models to match the signal levels in testing. Our algorithm first obtains initial estimates of source signals using unadapted speaker models and then detects the input signal-to-noise ratio (SNR) of the mixture. The input SNR is then used to adapt the speaker models for more accurate estimation. The two steps iterate until convergence. Compared to search-based SNR detection methods, our method is not limited to given SNR levels. Evaluations demonstrate that the iterative procedure converges quickly in a considerable range of SNRs and improves separation results significantly. Comparisons show that the proposed system performs significantly better than related model-based systems.

HARDWARE IMPLEMENTATION OF A PROPOSED QR-TLS DOA ESTIMATION METHOD AND MUSIC, ESPRIT ALGORITHMS ON NI-PXI PLATFORM

In this paper, we present an experimental veriflcation of a novel QR-TLS algorithm. Two other algorithms for direction of arrival (DOA) estimation of multiple incident source signals called multiple signal classiflcation (MUSIC) and estimation of signal parameters via rotational invariance techniques (ESPRIT) are implemented on a National Instruments (NI) PXI platform. The proposed method is based on subspace decomposition of a received data into a signal and a noise space using QR decomposition. The angle of the signal arrival information is extracted from the signal subspace by using the method of total least squares (TLS). The algorithms are implemented in LabView NI hardware. The experimental procedures are discussed in details which includes interfacing of the uniform linear array (ULA) of antennas with the NI-PXI platform, calibrating phase difierences between the RF receivers, and selecting transmitter and receiver parameters, for determining the DOAs of the multiple incident source signals. The experimental results are shown for a single and two sources lying at arbitrary angles from the array reference to verify the successful real time implementation of the proposed and other DOA estimation algorithms.