Blind Source Extraction Method Research Articles

Efficient independent vector extraction of dominant source (L).

The complete decomposition performed by blind source separation is computationally demanding and superfluous when only the speech of one specific target speaker is desired. This letter proposes a computationally efficient blind source extraction method based on the fast fixed-point optimization algorithm under the mild assumption that the average power of the source of interest outweighs the interfering sources. Moreover, a one-unit scaling operation is designed to solve the scaling ambiguity for source extraction. Experiments validate the efficacy of the proposed method in extracting the dominant source.

Auxiliary function-based algorithm for blind extraction of a moving speaker

In this paper, we propose a novel algorithm for blind source extraction (BSE) of a moving acoustic source recorded by multiple microphones. The algorithm is based on independent vector extraction (IVE) where the contrast function is optimized using the auxiliary function-based technique and where the recently proposed constant separating vector (CSV) mixing model is assumed. CSV allows for movements of the extracted source within the analyzed batch of recordings. We provide a practical explanation of how the CSV model works when extracting a moving acoustic source. Then, the proposed algorithm is experimentally verified on the task of blind extraction of a moving speaker. The algorithm is compared with state-of-the-art blind methods and with an adaptive BSE algorithm which processes data in a sequential manner. The results confirm that the proposed algorithm can extract the moving speaker better than the BSE methods based on the conventional mixing model and that it achieves improved extraction accuracy than the adaptive method.

Blind source extraction based on EMD and temporal correlation for rolling element bearing fault diagnosis

Purpose Fault diagnosis methods based on blind source separation (BSS) for rolling element bearings are necessary tools to prevent any unexpected accidents. In the field application, the actual signal acquisition is usually hindered by certain restrictions, such as the limited number of signal channels. The purpose of this study is to fulfill the weakness of the existed BSS method. Design/methodology/approach To deal with this problem, this paper proposes a blind source extraction (BSE) method for bearing fault diagnosis based on empirical mode decomposition (EMD) and temporal correlation. First, a single-channel undetermined BSS problem is transformed into a determined BSS problem using the EMD algorithm. Then, the desired fault signal is extracted from selected intrinsic mode functions with a multi-shift correlation method. Findings Experimental results prove the extracted fault signal can be easily identified through the envelope spectrum. The application of the proposed method is validated using simulated signals and rolling element bearing signals of the train axle. Originality/value This paper proposes an underdetermined BSE method based on the EMD and the temporal correlation method for rolling element bearings. A simulated signal and two bearing fault signal from the train rolling element bearings show that the proposed method can well extract the bearing fault signal. Note that the proposed method can extract the periodic fault signal for bearing fault diagnosis. Thus, it should be helpful in the diagnosis of other rotating machinery, such as gears or blades.

Underdetermined blind source extraction of early vehicle bearing faults based on EMD and kernelized correlation maximization

The incipient bearing fault diagnosis is crucial to the industrial machinery maintenance. Developed based on the blind source separation, blind source extraction (BSE) has recently become the focus of intensive research work. However, owing to certain industrial restrictions, the number of sensors is usually less than that of the source signals, which is defined as an underdetermined BSE problem to identify the fault signals. The kernelized methods are found to be robust to the noise, especially in the presence of outliers, which makes it a suitable tool to extract fault signatures submerged in the strong environment noise. Thus, this paper proposes a new underdetermined BSE method based on the empirical mean decomposition and kernelized correlation. The experimental results indicate that the extracted fault signature presents more obvious periodicity. Two important parameters of this method, including the multi-shift number and the kernel size are investigated to improve the algorithm performance. Furthermore, performance comparisons with underdetermined BSE based on the second order correlation are made to emphasize the advantage of the presented method. The application of the proposed method is validated using the simulated signal and the rolling element bearing signal of the train vehicle axle.

A two-step blind source extraction method and its application in fault diagnosis of rolling element bearing

The vibration signals will take on cyclical characteristics when fault arises in the rolling element bearing, and a two-step blind source extraction (BSE) method for fault diagnosis of rolling element bearing is proposed in the paper using the above property. Firstly, calculate the theoretical basic cyclet of the target source fault signal, and the weighted separation matrix ŵ and desired source signal are obtained coarsely. Secondly, use ŵ as the initial weighted matrix and apply the fixed-point algorithm basing on high-order statistics on the observed signals, and much more perfect target source signal is got at last. The proposed method has the following advantages over other BSE method such as constrained independent component analysis (CICA) basing on the analyzed results of simulation and experiment: The fundamental period τ of the target source signal does not needed to be estimated accurately, and the reference signal also does not need to be constructed precisely. However, these two conditions are the necessary prerequisites of CICA. Besides, the proposed method also has the advantage of higher accuracy over the other recent BSE methods through comparison.

Blind source extraction of rolling bearings' multi-type faults based on self-learned sparse atomics

The feature of rolling element bearings' multi-type faults is very hard to extract using common feature extraction method such as envelope demodulation, and the main reason is that there exists mutual coupling effect when multi-type faults arise in rolling element bearing synchronously. Blind source extraction originating from blind source separation is an effective method for feature extraction of rolling bearings' multi-type faults. However, the extraction result would not be ideal if blind source extraction is used directly due to the above stated mutual coupling effect. Sparse representation is a relative new signal processing method, which could capture the latent fault feature components buried in the vibration signal. So, blind source extraction of rolling element bearings' multi-type faults based on sparse representation is proposed in the paper. Firstly, the self-learned sparse atomics originating from sparse representation is applied to the multi-type faults vibration signals directly and several learned atomics are obtained. Then, the multi-type faults vibration signals are reconstructed based on the obtained learned atomics and sparse multi-type faults vibration signals are obtained. Thirdly, the blind source extraction method is applied to the reconstructed sparse vibration signals. Lastly, envelope demodulation is applied to the blind source extraction results respectively and satisfactory fault feature extraction results are obtained. The feasibility and effectiveness of the proposed method are verified through simulation and experiment.

Second-Order Cyclostationary Statistics-Based Blind Source Extraction From Convolutional Mixtures

Blind source extraction (BSE) aims to extract the source of interest (SOI) from the outputs of a mixing system, which is a challenging problem. A property existing in many signals is cyclostationarity and this property has been widely exploited in BSE. While various cyclostationarity-based BSE methods have been reported in the literature, they usually require the mixing system to be instantaneous. In this paper, we address BSE in the context that the mixing system is convolutional. Specifically, a new BSE method is developed to extract cyclostationary source signal from the outputs of a multiple-input-multiple-output finite-impulse-response mixing system. It is shown that if the SOI has a unique cyclostationary frequency, it can be recovered from the measured data. The effectiveness of the proposed BSE method is demonstrated by simulation results.

A Framework of Target Detection in Hyperspectral Imagery Based on Blind Source Extraction

A framework based on the blind source extraction (BSE) algorithm is proposed to detect targets in remotely sensed hyperspectral images. The mean square cross prediction error (MSCPE)-based BSE method is used as the kernel algorithm where the autoregressive (AR) parameters of the targets' spectra are utilized as priors. Numerical simulations show that the proposed framework highlights the desired signal, suppresses the backgrounds, and is able to detect the distribution of the target. In the experiments, the data from the Rochester Institute of Technology (RIT) were used to evaluate the framework. The proposed method achieved a better performance in the tradeoff between the PD and the PFA with subpixel target detection compared with the constrained energy minimization (CEM), the adaptive cosine estimator (ACE), the matched filter (MF), the generalized likelihood ratio test (GLRT), the adaptive matched subspace detector (AMSD), and the orthogonal subspace projection (OSP).

Blind source extraction based on time-frequency characteristics for underwater object acoustic scattering

The physical mechanism and signal characteristics of acoustic scattering are the vital basis for target recognition. But underwater target acoustic scattering components are aliasing in time-frequency (TF) domain, for which the target elastic acoustic scattering characteristics are difficult to detect. Additionally, the existing blind source separation methods are effective only on condition that the number of array elements is equal to or greater than the number of the source signals. To address these problems, a novel TF domain blind source extraction method of separating target acoustic scattering components is proposed in this paper. The method only uses the TF energy characteristic differences among the target acoustic scattering components, and special limitations on target echo structures are unnecessary. Image morphology filter is used to remove the cross-term interference from time-frequency distribution (TFD) of the received array signals. Then, the single source which shows maximum energy concentration at the corresponding auto-term TF points is extracted through three operations: i) selecting the single source auto-term TF points from the auto-term ones; ii) constructing the spatial TFD matrix according to the selected single source auto-term TF points; iii) obtaining the single source by decomposing the eigenvalue of their spatial TFD matrix. Finally, the extracted single signal is excluded by the tightening process from the received array signals, and each single signal is separated successively by repeating the above steps. In addition, a signal processing model which can describe the physical characteristics of the target echoes is established based on the separated signal components. Simulations illustrate that the image morphological filter can remove the cross-term interference and improve the TF resolution of the Wigner-Ville distribution. Anechoic pool experimental results show that the TF domain blind source extraction algorithm can well separate each target acoustic scattering component, it can also achieve a higher output signal-to-noise ratio. Furthermore, the separated elastic acoustic scattering components are in good agreement with the results computed by the surface wave generating theory, so the method can provide the robust and reliable feature for underwater target recognition.

A Method for Extracting Biomedical Signals with Specific Temporal Structure

This paper proposes a blind source extraction (BSE) method on the basis of several time-delay autocorrelations from primary sources. This method is a second-order statistic learning algorithm, which can extract a desired biomedical signal while it shows a specific temporal structure. In contrast to conventional BSE methods, it is simple and do not need to choose any learning step size. Computer simulations on biomedical measurements demonstrate its validity and high performance in the process of revealing the underlying desired signal.

Noisy component extraction with reference

Blind source extraction (BSE) is particularly attractive to solve blind signal mixture problems where only a few source signals are desired. Many existing BSE methods do not take into account the existence of noise and can only work well in noise-free environments. In practice, the desired signal is often contaminated by additional noise. Therefore, we try to tackle the problem of noisy component extraction. The reference signal carries enough prior information to distinguish the desired signal from signal mixtures. According to the useful properties of Gaussian moments, we incorporate the reference signal into a negentropy objective function so as to guide the extraction process and develop an improved BSE method. Extensive computer simulations demonstrate its validity in the process of revealing the underlying desired signal.

A single stage approach to blind source extraction based on second order statistics

A single stage blind source extraction (BSE) approach to extract one desired source signal from an instantaneous mixture of several signals is studied in this paper. In contrast to methods based on blind source separation (BSS) followed by a classifier we extract directly the desired source signal. In recent papers it has been shown that prior information about the desired source's mixing column or autocorrelation function can be incorporated in a single stage BSE algorithm. The focus of this paper is to give insight in the BSE problem for scenarios where joint diagonalization of correlation matrices is applicable, i.e., to exploit the full potential of the SOS of observed signals.In this work we unify and expand these previously presented BSE methods, which lead to the unified instantaneous blind source extraction (UIBSE) algorithm. We show that UIBSE is even more flexible than previously presented since both types of prior information may be combined to obtain better results in selecting the desired source. Above that, the same algorithm may be used for different objective functions. Finally, a performance comparison with BSS algorithms that are followed by a classifier shows that UIBSE is more robust to noisy observations.

Harmonic retrieval by period blind source extraction method: Model and algorithm

A frequently encountered problem in signal processing field is harmonic retrieval in additive colored Gaussian or non-Gaussian noise, especially when the frequencies of the harmonic signals are closely spaced in frequency domain. The purpose of this paper is to develop novel harmonic retrieval algorithm based on blind source extraction (BSE) method from linear mixtures of harmonic signals using only one observed channel signal. First, we establish the blind source separation (BSS) based harmonic retrieval model in additive noise using the only one observed channel, at the same time, the fundamental principle of BSE based harmonics retrieval algorithm is analyzed in detail. Then, based on the established harmonic BSS model, we propose a BSE approach to the harmonic retrieval using the concept of period BSE method, as a result, the harmonic retrieval algorithm using only one channel mixed signals is derived. Simulation results show that the proposed algorithm is able to separate the harmonic source signals and yield ideal performance.

Robust Method for Separation of Noisy Biomedical Signals

Biomedical signals are a rich source of information about physiological processes, but they are often contaminated by noise. In order to separate biomedical signals from mixtures effectually, we propose a novel blind source extraction method via independent component analysis (ICA). The robustness with respect to noise of this method lies in two-fold: on the one hand, the method does not lead to biassed estimates and, on the other hand, it minimizes the amount of signal and noise interference on the estimated sources. Preliminary results tested with ECG signals have demonstrated that the proposed method may be promising for blindly separating biomedical signals in the presence of noise and further decompose the mixed signals into subcomponents.

A reference-based blind source extraction algorithm

By utilizing a priori information available as reference, constrained independent component analysis (cICA) or independent component analysis with reference (ICA-R) achieves some advantages over other methods. However, ICA-R is very time-consuming; moreover, it is very difficult to determine its threshold parameter, once the value is improperly chosen the algorithm will fail to converge. In order to overcome these drawbacks, a very simple blind source extraction method, whose optimization function is simply the closeness measure between the desired output and its corresponding reference in ICA-R, is proposed in this paper. Experiments with synthesized data and real-world electrocardiograph data confirm its validity and superiority.

Post-Nonlinear Blind Extraction in the Presence of Ill-Conditioned Mixing

An extension of blind source extraction (BSE) of one or a group of sources to the case of ill-conditioned and post-nonlinear (PNL) mixing is introduced. This is achieved by a ldquomixed objectiverdquo type of cost function which jointly maximizes the kurtosis of a recovered source and estimates a measure of nonlinearity within the mixing system. This helps to circumvent problems with existing BSE methods, which are limited to noiseless and linear mixing models. Simulations illustrate the performance of the proposed algorithm and its usefulness, especially in the presence of very ill-conditioned mixing systems.

MACBSE: Extracting signals with linear autocorrelations

This paper proposes blind source extraction methods based on several time-delay autocorrelations of primary sources, called MACBSE. The MACBSE approaches are batch fixed-point learning algorithms for extraction of source signals with linear autocorrelations. The fixed-point algorithms are very simple and do not need to choose any learning step sizes. Furthermore, the convergence properties of the algorithms are analyzed. Their efficiencies are demonstrated by extensive computer simulations.