Second-order Blind Identification Research Articles

Study on a Novel Bearing Fault Diagnosis Method from Frequency and Energy Perspective

AbstractEarly identification of faults in rolling element bearings is a challenging task; especially extracting transient characteristics from a noisy signal and identifying bearings fault become critical steps. In this paper, a novel method for real time fault detection in rolling element bearings is proposed to deal with non-stationary fault signals from frequency and energy perspective. Second-order blind identification (SOBI) and wavelet packet decomposition are organically integrated to diagnose the early bearing faults, the fault vibration signals are processed by SOBI algorithm, and feature information is extracted; meanwhile, fault vibration signals are decomposed by the wavelet packet, the energy of terminal nodes(at the bottom layer of wavelet packet decomposition) are analyzed because the energy of terminal nodes has different sensitive to different component faults. Therefore, the bearing faults can be diagnosed by organic combination of fault characteristic frequency analysis and energy of the terminal nodes, and the effectiveness, feasibility and robustness of the proposed method have been verified by experimental data.

Identification of Dominant Low Frequency Oscillation Modes Based on Blind Source Separation

On-line monitoring and analysis of low frequency oscillation (LFO) are important for stability and security of a power system. This paper proposes a blind source separation (BSS) based method for LFO modal analysis of one-channel measured signal, which consists of a second-order blind identification (SOBI) algorithm and the Hilbert transform (HT) technique. This is the first time when a BSS technique is applied as a modal decomposition tool in this field. The proposed SOBI-HT based dominant mode identification method combines an iteration procedure with an identification procedure to determine the model order, and then identify dominant modes. The performance of the proposed method is evaluated on numerical simulation signals and a record of real data, verifying its effectiveness and accuracy.

Condition assessment of structure with tuned mass damper using empirical wavelet transform

Tuned mass damper (TMD) has been one of the most commonly used passive vibration control devices over the past few decades. While an optimally designed TMD can significantly suppress the structural vibration, detuning often occurs due to various reasons such as change in operating conditions or variation in primary structure properties, resulting in degradation of TMD’s performance. In order to restore its performance, it is necessary to estimate the modal properties of the primary structure, and perform the re-tuning process. Such an exercise requires powerful signal processing methods to successfully extract the structural modes in the presence of closely-spaced modes. This study focuses on the identification of modal frequencies and damping of the structure installed with a TMD. In view of the advantages and limitations of existing modal identification methods, this paper provides a new technique that combines the second-order blind identification (SOBI) method with the empirical wavelet transform (EWT) to delineate closely-spaced frequencies. While the SOBI method does not guarantee the separation of closely-spaced modes and suffers from the limitation of generating mode-mixed modal responses, the EWT operates on the modal responses estimated by the SOBI and yields the closely-spaced natural frequencies. The proposed method is illustrated using a six-story simulation model with a wide range of detuning cases. An experiment on a three-story bench-scale model equipped with a TMD is also conducted to validate the applicability of the proposed method.

On using the Hilbert transform for blind identification of complex modes: A practical approach

The modal identification of dynamical systems under operational conditions, when subjected to wide-band unmeasured excitations, is today a viable alternative to more traditional modal identification approaches based on processing sets of measured FRFs or impulse responses. Among current techniques for performing operational modal identification, the so-called blind identification methods are the subject of considerable investigation. In particular, the SOBI (Second-Order Blind Identification) method was found to be quite efficient. SOBI was originally developed for systems with normal modes. To address systems with complex modes, various extension approaches have been proposed, in particular: (a) Using a first-order state-space formulation for the system dynamics; (b) Building complex analytic signals from the measured responses using the Hilbert transform. In this paper we further explore the latter option, which is conceptually interesting while preserving the model order and size. Focus is on applicability of the SOBI technique for extracting the modal responses from analytic signals built from a set of vibratory responses. The novelty of this work is to propose a straightforward computational procedure for obtaining the complex cross-correlation response matrix to be used for the modal identification procedure. After clarifying subtle aspects of the general theoretical framework, we demonstrate that the correlation matrix of the analytic responses can be computed through a Hilbert transform of the real correlation matrix, so that the actual time-domain responses are no longer required for modal identification purposes. The numerical validation of the proposed technique is presented based on time-domain simulations of a conceptual physical multi-modal system, designed to display modes ranging from normal to highly complex, while keeping modal damping low and nearly independent of the modal complexity, and which can prove very interesting in test bench applications. Numerical results for complex modal identifications are presented, and the quality of the identified modal matrix and modal responses, extracted using the complex SOBI technique and implementing the proposed formulation, is assessed.

On the effects of multimodal information integration in multitasking

There have recently been considerable advances in our understanding of the neuronal mechanisms underlying multitasking, but the role of multimodal integration for this faculty has remained rather unclear. We examined this issue by comparing different modality combinations in a multitasking (stop-change) paradigm. In-depth neurophysiological analyses of event-related potentials (ERPs) were conducted to complement the obtained behavioral data. Specifically, we applied signal decomposition using second order blind identification (SOBI) to the multi-subject ERP data and source localization. We found that both general multimodal information integration and modality-specific aspects (potentially related to task difficulty) modulate behavioral performance and associated neurophysiological correlates. Simultaneous multimodal input generally increased early attentional processing of visual stimuli (i.e. P1 and N1 amplitudes) as well as measures of cognitive effort and conflict (i.e. central P3 amplitudes). Yet, tactile-visual input caused larger impairments in multitasking than audio-visual input. General aspects of multimodal information integration modulated the activity in the premotor cortex (BA 6) as well as different visual association areas concerned with the integration of visual information with input from other modalities (BA 19, BA 21, BA 37). On top of this, differences in the specific combination of modalities also affected performance and measures of conflict/effort originating in prefrontal regions (BA 6).

Operational modal analysis of three-dimensional structures by second-order blind identification and least square generalized inverse

Most complex engineering structures are three-dimensional in practice. The process of one-dimensional extending to three-dimensional is a challenge that must be conquered by Operational Modal Analysis (OMA) methods when these methods are applied to complex engineering applications supported by scientific researches. This study put forward a new three-dimensional structure OMA method based on Second-Order Blind Identification (SOBI) and general reversion of least square. Firstly, modal coordinates decomposition of one-dimensional structural vibration response signal with SOBI. Secondly, the reasons that modal parameters identified by SOBI including energy uncertainty, order uncertainty and modal missing are explained in theory. Thirdly, the SOBI algorithm is used to decompose the response signals of displacement of a direction whose vibration response is the largest, then the other two directions are calculated by using the least square generalized inverse algorithm, and the modal parameters of three-dimensional structures are identified by the matrix assembly method. Numerical simulation results in a cylindrical shell demonstrated that this novel method is practical and effective by applied to practice in OMA of three-dimensional structures, and robustness to Gauss measurement noise disturbances.

Non-contact, synchronous dynamic measurement of respiratory rate and heart rate based on dual sensitive regions

BackgroundCurrently, many imaging photoplethysmography (IPPG) researches have reported non-contact measurements of physiological parameters, such as heart rate (HR), respiratory rate (RR), etc. However, it is accepted that only HR measurement has been mature for applications, and other estimations are relatively incapable for reliable applications. Thus, it is worth keeping on persistent studies. Besides, there are some issues commonly involved in these approaches need to be explored further. For example, motion artifact attenuation, an intractable problem, which is being attempted to be resolved by sophisticated video tracking and detection algorithms.MethodsThis paper proposed a blind source separation-based method that could synchronously measure RR and HR in non-contact way. A dual region of interest on facial video image was selected to yield 6-channels Red/Green/Blue signals. By applying Second-Order Blind Identification algorithm to those signals generated above, we obtained 6-channels outputs that contain blood volume pulse (BVP) and respiratory motion artifact. We defined this motion artifact as respiratory signal (RS). For the automatic selections of the RS and BVP among these outputs, we devised a kurtosis-based identification strategy, which guarantees the dynamic RR and HR monitoring available.ResultsThe experimental results indicated that, the estimation by the proposed method has an impressive performance compared with the measurement of the commercial medical sensors.ConclusionsThe proposed method achieved dynamic measurement of RR and HR, and the extension and revision of it may have the potentials for more physiological signs detection, such as heart rate variability, eye blinking, nose wrinkling, yawn, as well as other muscular movements. Thus, it might provide a promising approach for IPPG-based applications such as emotion computation and fatigue detection, etc.

The dynamic finite element model calibration method of concrete dams based on strong-motion records and multivariate relevant vector machines

In this work, a new finite element (FE) model calibration method of concrete dams based on strong-motion records and multivariate relevant vector machines (MRVM) is proposed. The modal features of a dam are extracted using second order blind identification (SOBI) based method at first. For some selected combinations of uncertain parameters of the FE model using the Latin hypercube design, the corresponding structural modal features are calculated using the finite element method (FEM). With these data, a procedure to calibrate the uncertain parameters of a dam’s dynamic FE model is developed. By taking the uncertain parameters as inputs and the calculated structural modal features using FEM as outputs, the MRVM model is trained to record the complex relationship between them. Then, the genetic algorithm (GA) is adopted to solve the optimization problem corresponding to the dynamic FE model calibration problem, and the trained MRVM model, instead of FEM, is used to obtain the modal parameters of a dam for different feasible solutions during the optimization search process to improve the computational efficiency. Using the simulated seismic response records of a numerical example the accuracy, robustness and computation efficiency of the proposed dynamic FE model calibration method is verified. The analysis result using the strong-motion records of a realistic concrete dam indicates that the proposed dynamic FE model calibration method has good performance.

Single-Trial Evoked Potential Estimation Using Iterative Principal Component Analysis

In this paper, we have presented an iterative principal component analysis (PCA) method to obtain single trial evoked potential. The performance of the iterative PCA has been compared with the performance of other iterative component analysis methods, such as independent component analysis, canonical correlation analysis (CCA), and the second-order blind identification, using simulated data at different signal-to-noise ratios as well as actual recordings of visual evoked potentials. In both the simulated and real cases, iterative PCA and CCA perform better than the other methods to estimate the amplitudes. In the estimated trials of the proposed method, the latency of the evoked potentials lies between ±4 ms range of true latency in about 90% of the case; but for other methods, this percentage is around 60%.

Increasing the efficiency and efficacy of second-order blind identification (SOBI) methods

This paper proposes a new technique to increase the efficiency and effectiveness of second-order blind identification (SOBI) methods by reducing the number of time-lagged covariance matrices required to produce highly accurate mixing matrix estimates. The technique is based on randomly selecting the time-lagged covariance matrices as opposed to choosing them sequentially, which takes advantage of a property of independence with regard to the selection of time-lagged covariance matrices, while simultaneously reducing the correlation between selected covariance matrices. The proposed randomized approach is first applied to undamped and damped sinusoids to demonstrate its effectiveness. The randomized method is then applied to a three-degree-of-freedom linear structure subjected to random excitation. Finally, the proposed method is incorporated into a modified version of the SOBI method and applied to perform identification of the UCLA Factor building from recorded earthquake responses. In each case, the performance of the randomized approach is compared with the traditional sequential selection of time-lagged covariance matrices, and the randomized approach consistently demonstrates superior performance both in terms of accuracy and efficiency. Additionally, the proposed randomized SOBI approach was easily inserted into the modified SOBI method demonstrating its malleability for SOBI variants and other blind source separation methods. Copyright © 2016 John Wiley & Sons, Ltd.

Detection of sudden structural damage using blind source separation and time–frequency approaches

Seismic signal processing is one of the most reliable methods of detecting the structural damage during earthquakes. In this paper, the use of the hybrid method of blind source separation (BSS) and time–frequency analysis (TFA) is explored to detect the changes in the structural response data. The combination of the BSS and TFA is applied to the seismic signals due to the non-stationary nature of them. Firstly, the second-order blind identification technique is used to decompose the response signal of structural vibration into modal coordinate signals which will be mono-components for TFA. Then each mono-component signal is analyzed to extract instantaneous frequency of structure. Numerical simulations and a real-world seismic-excited structure with time-varying frequencies show the accuracy and robustness of the developed algorithm. TFA of extracted sources shows that used method can be successfully applied to structural damage detection. The results also demonstrate that the combined method can be used to identify the time instant of structural damage occurrence more sharply and effectively than by the use of TFA alone.

Wearable Sensor Localization Considering Mixed Distributed Sources in Health Monitoring Systems.

In health monitoring systems, the base station (BS) and the wearable sensors communicate with each other to construct a virtual multiple input and multiple output (VMIMO) system. In real applications, the signal that the BS received is a distributed source because of the scattering, reflection, diffraction and refraction in the propagation path. In this paper, a 2D direction-of-arrival (DOA) estimation algorithm for incoherently-distributed (ID) and coherently-distributed (CD) sources is proposed based on multiple VMIMO systems. ID and CD sources are separated through the second-order blind identification (SOBI) algorithm. The traditional estimating signal parameters via the rotational invariance technique (ESPRIT)-based algorithm is valid only for one-dimensional (1D) DOA estimation for the ID source. By constructing the signal subspace, two rotational invariant relationships are constructed. Then, we extend the ESPRIT to estimate 2D DOAs for ID sources. For DOA estimation of CD sources, two rational invariance relationships are constructed based on the application of generalized steering vectors (GSVs). Then, the ESPRIT-based algorithm is used for estimating the eigenvalues of two rational invariance matrices, which contain the angular parameters. The expressions of azimuth and elevation for ID and CD sources have closed forms, which means that the spectrum peak searching is avoided. Therefore, compared to the traditional 2D DOA estimation algorithms, the proposed algorithm imposes significantly low computational complexity. The intersecting point of two rays, which come from two different directions measured by two uniform rectangle arrays (URA), can be regarded as the location of the biosensor (wearable sensor). Three BSs adopting the smart antenna (SA) technique cooperate with each other to locate the wearable sensors using the angulation positioning method. Simulation results demonstrate the effectiveness of the proposed algorithm.

System identification through nonstationary data using Time–Frequency Blind Source Separation

Classical output-only system identification (SI) methods are based on the assumption of stationarity of the system response. However, measured response of buildings and bridges is usually non-stationary due to strong winds (e.g. typhoon, and thunder storm etc.), earthquakes and time-varying vehicle motions. Accordingly, the response data may have time-varying frequency contents and/or overlapping of modal frequencies due to non-stationary colored excitation. This renders traditional methods problematic for modal separation and identification. To address these challenges, a new SI technique based on Time–Frequency Blind Source Separation (TFBSS) is proposed. By selectively utilizing “effective” information in local regions of the time–frequency plane, where only one mode contributes to energy, the proposed technique can successfully identify mode shapes and recover modal responses from the non-stationary response where the traditional SI methods often encounter difficulties. This technique can also handle response with closely spaced modes which is a well-known challenge for the identification of large-scale structures. Based on the separated modal responses, frequency and damping can be easily identified using SI methods based on a single degree of freedom (SDOF) system. In addition to the exclusive advantage of handling non-stationary data and closely spaced modes, the proposed technique also benefits from the absence of the end effects and low sensitivity to noise in modal separation. The efficacy of the proposed technique is demonstrated using several simulation based studies, and compared to the popular Second-Order Blind Identification (SOBI) scheme. It is also noted that even some non-stationary response data can be analyzed by the stationary method SOBI. This paper also delineates non-stationary cases where SOBI and the proposed scheme perform comparably and highlights cases where the proposed approach is more advantageous. Finally, the performance of the proposed method is evaluated using a full-scale non-stationary response of a tall building during an earthquake and found it to perform satisfactorily.

Application of Blind Source Separation Algorithms and Ambient Vibration Testing to the Health Monitoring of Concrete Dams

In this work, using AVT data, a health monitoring method for concrete dams based on two different blind source separation (BSS) methods, that is, second-order blind identification (SOBI) and independent component analysis (ICA), is proposed. A modal identification procedure, which integrates the SOBI algorithm and modal contribution, is first adopted to extract structural modal features using AVT data. The method to calculate the modal contribution index for SOBI-based modal identification methods is studied, and the calculated modal contribution index is used to determine the system order. The selected modes are then used to calculate modal features and are analysed using ICA to extract some independent components (ICs). The square prediction error (SPE) index and its control limits are then calculated to construct a control chart for the structural dynamic features. For new AVT data of a dam in an unknown health state, the newly calculated SPE is compared with the control limits to judge whether the dam is normal. With the simulated AVT data of the numerical model for a concrete gravity dam and the measured AVT data of a practical engineering project, the performance of the dam health monitoring method proposed in this paper is validated.

Characterization of Electrophysiological Propagation by Multichannel Sensors.

The propagation of electrophysiological activity measured by multichannel devices could have significant clinical implications. Gastric slow waves normally propagate along longitudinal paths that are evident in recordings of serosal potentials and transcutaneous magnetic fields. We employed a realistic model of gastric slow wave activity to simulate the transabdominal magnetogastrogram (MGG) recorded in a multichannel biomagnetometer and to determine characteristics of electrophysiological propagation from MGG measurements. Using MGG simulations of slow wave sources in a realistic abdomen (both superficial and deep sources) and in a horizontally-layered volume conductor, we compared two analytic methods (second-order blind identification, SOBI and surface current density, SCD) that allow quantitative characterization of slow wave propagation. We also evaluated the performance of the methods with simulated experimental noise. The methods were also validated in an experimental animal model. Mean square errors in position estimates were within 2cm of the correct position, and average propagation velocities within 2mm/s of the actual velocities. SOBI propagation analysis outperformed the SCD method for dipoles in the superficial and horizontal layer models with and without additive noise. The SCD method gave better estimates for deep sources, but did not handle additive noise as well as SOBI. SOBI-MGG and SCD-MGG were used to quantify slow wave propagation in a realistic abdomen model of gastric electrical activity. These methods could be generalized to any propagating electrophysiological activity detected by multichannel sensor arrays.

Multi‐mode clutter suppression of multiple‐input–multiple‐output over‐the‐horizon radar based on blind source separation

Spread-Doppler clutter caused by multi-mode propagation seriously affects the over-the-horizon radar detection performance for slow ships. To solve this problem, the authors propose to suppress multi-mode clutter using spatial information in the multiple-input–multiple-output over-the-horizon radar system. Traditional processing algorithms need prior information of the target direction. However, it is often not readily available. In this study, they propose a novel multi-mode clutter suppression algorithm based on the second-order blind identification (SOBI) technique. The proposed algorithm does not need prior information of the target direction and can effectively suppress the multi-mode clutter. Considering that the SOBI algorithm affected by the similarity in sources spectra, they improve the proposed algorithm using the spatial smoothing technique. Simulation results show that, in the single target scene, the improved algorithm can effectively eliminate the spectrum burrs caused by incomplete separation, making separated signals smoother. Signal-to-clutter-noise ratio is greatly improved. The good multi-mode clutter suppression performance is obtained. In the multi-mode scenario that double targets with the same radial velocity, that is, sources spectra have similarities, the algorithm based on SOBI cannot separate all targets. Although the improved algorithm can make all targets visible and effectively suppresses the multi-mode clutter.

Separation of Uncorrelated Stationary time series using Autocovariance Matrices

In blind source separation, one assumes that the observed p time series are linear combinations of p latent uncorrelated weakly stationary time series. To estimate the unmixing matrix, which transforms the observed time series back to uncorrelated latent time series, second‐order blind identification (SOBI) uses joint diagonalization of the covariance matrix and autocovariance matrices with several lags. In this article, we find the limiting distribution of the well‐known symmetric SOBI estimator under general conditions and compare its asymptotical efficiencies to those of the recently introduced deflation‐based SOBI estimator. The theory is illustrated by some finite‐sample simulation studies.

Techniques for estimating the unknown functions of incomplete experimental spectral and correlation response matrices

In this paper, we propose analytical and numerical straightforward approximate methods to estimate the unknown terms of incomplete spectral or correlation matrices, when the cross-spectra or cross-correlations available from multiple measurements do not cover all pairs of transducer locations. The proposed techniques may be applied whenever the available data includes the auto-spectra at all measurement locations, as well as selected cross-spectra which implicates all measurement locations. The suggested methods can also be used for checking the consistency between the spectral or correlation functions pertaining to measurement matrices, in cases of suspicious data. After presenting the proposed spectral estimation formulations, we discuss their merits and limitations. Then we illustrate their use on a realistic simulation of a multi-supported tube subjected to turbulence excitation from cross-flow. Finally, we show the effectiveness of the proposed techniques by extracting the modal responses of the simulated flow-excited tube, using the SOBI (Second Order Blind Identification) method, from an incomplete response matrix 11This paper is an extended version of the work presented at the International Conference on Condition Machinery in Non-Stationary Operations (CMMNO13), May 8–10, 2013, Ferrara, Italy..

Blind Source Separation of Instantaneous Mixture of Delayed Sources Using High-Order Taylor Approximation

This paper deals with the problem of blind source separation (BSS), where observed signals are a mixture of delayed sources. In reference to a previous work, when the delay time is small such that the first-order Taylor approximation holds, delayed observations are transformed into an instantaneous mixture of original sources and their derivatives, for which an extended second-order blind identification (SOBI) approach is used to recover sources. Inspired by the results of this previous work, we propose to generalize its first-order Taylor approximation to suit higher-order approximations in the case of a large delay time based on a similar version of its extended SOBI. Compared to SOBI and its extended version for a first-order Taylor approximation, our method is more efficient in terms of separation quality when the delay time is large. Simulation results verify the performance of our approach under different time delays and signal-to-noise ratio conditions, respectively.

Evaluation of performance to detect default mode network among some algorithms applied to resting-state fMRI data.

Significant correlation exists in the blood-oxygen-level-dependent (BOLD) signals of resting-state fMRI across different regions in the brain. These regions form the default mode network (DMN), salience network (SN), sensory networks, and others. Among these, the DMN is widely investigated in relation to various mental diseases. Several analytic methods are available for obtaining the DMN activity from individuals' fMRI time-series signals, but a fully effective method has not yet been established. In the present study, we examined a functional connectivity analysis and three algorithms of blind source separation including independent component analysis, second-order blind identification, and non-negative matrix factorization using a set of resting-state fMRI data measured for twelve young participants. Results showed that the second-order blind identification yielded superior performance for the DMN detection, indicating significant activation in all DMN regions based on statistical parametric maps.