Showing papers on "Time–frequency analysis published in 2013"
TL;DR: A systematic review of over 20 major time-frequency analysis methods reported in more than 100 representative articles published since 1990 can be found in this article, where their fundamental principles, advantages and disadvantages, and applications to fault diagnosis of machinery have been examined.
719 citations
TL;DR: The proposed fault diagnosis technique based on acoustic emission (AE) analysis with the Hilbert-Huang Transform (HHT) and data mining tool can increase reliability for the faults diagnosis of ball bearing.
Abstract: This paper presents a fault diagnosis technique based on acoustic emission (AE) analysis with the Hilbert-Huang Transform (HHT) and data mining tool HHT analyzes the AE signal using intrinsic mode functions (IMFs), which are extracted using the process of Empirical Mode Decomposition (EMD) Instead of time domain approach with Hilbert transform, FFT of IMFs from HHT process are utilized to represent the time frequency domain approach for efficient signal response from rolling element bearing Further, extracted statistical and acoustic features are used to select proper data mining based fault classifier with or without filter K-nearest neighbor algorithm is observed to be more efficient classifier with default setting parameters in WEKA APF-KNN approach, which is based on asymmetric proximity function with optimize feature selection shows better classification accuracy is used Experimental evaluation for time frequency approach is presented for five bearing conditions such as healthy bearing, bearing with outer race, inner race, ball and combined defect The experimental results show that the proposed method can increase reliability for the faults diagnosis of ball bearing
245 citations
TL;DR: Time-frequency analysis plays a significant role in seismic data processing and interpretation and demonstrates that this method promises higher spectral-spatial resolution than the short-time Fourier transform or wavelet transform.
Abstract: Time-frequency analysis plays a significant role in seismic data processing and interpretation. Complete ensemble empirical mode decomposition decomposes a seismic signal into a sum of oscillatory components, with guaranteed positive and smoothly varying instantaneous frequencies. Analysis on synthetic and real data demonstrates that this method promises higher spectral-spatial resolution than the short-time Fourier transform or wavelet transform. Application on field data thus offers the potential of highlighting subtle geologic structures that might otherwise escape unnoticed.
162 citations
TL;DR: This work focuses on sinusoidal desired signals with sparse frequency-domain representation but shows that the analysis can be straightforwardly generalized to nonsinusoidal signals with known structures.
Abstract: A compressive sensing (CS) approach for nonstationary signal separation is proposed. This approach is motivated by challenges in radar signal processing, including separations of micro-Doppler and main body signatures. We consider the case where the signal of interest assumes sparse representation over a given basis. Other signals present in the data overlap with the desired signal in the time and frequency domains, disallowing conventional windowing or filtering operations to be used for desired signal recovery. The proposed approach uses linear time-frequency representations to reveal the data local behavior. Using the L-statistics, only the time-frequency (TF) points that belong to the desired signal are retained, whereas the common points and others pertaining only to the undesired signals are deemed inappropriate and cast as missing samples. These samples amount to reduced frequency observations in the TF domain. The linear relationship between the measurement and sparse domains permits the application of CS techniques to recover the desired signal without significant distortion. We focus on sinusoidal desired signals with sparse frequency-domain representation but show that the analysis can be straightforwardly generalized to nonsinusoidal signals with known structures. Several examples are provided to demonstrate the effectiveness of the proposed approach.
151 citations
15 Oct 2013
TL;DR: Main features of the second major release of the LTFAT toolbox, which includes generalizations of the Gabor transform, the wavelets module, the frames framework and the real-time block processing framework, are introduced.
Abstract: The Large Time Frequency Analysis Toolbox (LTFAT) is a modern Octave/Matlab toolbox for time-frequency analysis, synthesis, coefficient manipulation and visualization. It’s purpose is to serve as a tool for achieving new scientific developments as well as an educational tool. The present paper introduces main features of the second major release of the toolbox which includes: generalizations of the Gabor transform, the wavelets module, the frames framework and the real-time block processing framework.
107 citations
06 May 2013
TL;DR: The analysis results show that the matching demodulation transform is powerful in the analysis of FM signals and is an effective tool for the feature extraction of rub-impact faults.
Abstract: Conventional time-frequency representation (TFR) methods have played an important role in characterizing the time-frequency (TF) pattern of the nonstationary signal. In this paper, a new TFR algorithm, called matching demodulation transform (MDT), is introduced to extract the feature of highly oscillating frequency modulation for rotor rub-impact fault. When the early rub-impact fault occurs in the rotor system, the vibration signals will present frequency modulation feature because of the periodic rub-impact between the stator and the rotor. Through an iterative demodulation procedure, the highly oscillating frequency modulation feature is represented with satisfactory energy concentration in TF plane. The validity of the technique is then demonstrated on a real rotor system of a gas turbine with rub-impact fault. The analysis result of this application shows that the MDT method is powerful in the analysis of frequency modulation signals and is an effective tool for the feature extraction of rub-impact faults.
88 citations
TL;DR: Comparison with several conventional TFD methods on both numerical multicomponent signal and bat echolocation signal validates the potential and the effectiveness of the proposed time–frequency fusion technique.
Abstract: Chirplet transform (CT) is effective in characterization of instantaneous frequency (IF) for monocomponent linear-frequency-modulated signal. However, the CT is not suitable to analyze multicomponent signal with nonlinear-frequency-modulated component. In this paper, a time–frequency fusion technique based on polynomial CT (PCT) (TFPCT) is proposed to characterize the time–frequency structure of such signals. The TFPCT relies on the fact that the PCT is able to concentrate the energy closely along the IF of the monocomponent signal in time–frequency distribution (TFD). For multicomponent signal, the TFPCT first estimates the proper coefficients with respect to individual component and, second, produces a series of the TFD using the PCT. Each TFD has better energy concentration along the IF of one component. Then, in order to reduce the interference of unwanted component and preserve the component of interest, each TFD is filtered and grouped as an image. At last, the TFPCT combines these TFDs to be an eventual fused TFD, which has the energy concentrating closely along the IF of all components. Comparison with several conventional TFD methods on both numerical multicomponent signal and bat echolocation signal validates the potential and the effectiveness of the proposed method.
88 citations
TL;DR: This letter presents a novel algorithm to compute the instantaneous frequency (IF) of a multicomponent nonstationary signal using a combination of fractional spectrograms (FS).
Abstract: This letter presents a novel algorithm to compute the instantaneous frequency (IF) of a multicomponent nonstationary signal using a combination of fractional spectrograms (FS). A high resolution time frequency distribution (TFD) is defined by combining FS computed using windows of varying lengths and chirp rates. The IF of individual signal components is then computed by applying a peak detection and component extraction procedure. The mean square error (MSE) of IF estimates computed with the AFS is lower than the MSE of IF estimates obtained from other TFDs for SNR varying from -5 dB to 16 dB.
74 citations
TL;DR: In this article, a nonlinear time-frequency feature based on a timefrequency manifold (TFM) technique is proposed, which combines non-stationary information and the nonlinear information of analyzed signals, and hence exhibits valuable properties.
70 citations
TL;DR: In this article, a simplified mathematical model was developed and a series of experiments were carried out on a roller rig for the detection of wheel flats and rail surface defects using three commonly used time-frequency analysis techniques: Short-Time Fourier Transform, Wigner-Ville transform and wavelet transform.
Abstract: Damage to the surface of railway wheels and rails commonly occurs in most railways. If not detected, it can result in the rapid deterioration and possible failure of rolling stock and infrastructure components causing higher maintenance costs. This paper presents an investigation into the modelling and simulation of wheel-flat and rail surface defects. A simplified mathematical model was developed and a series of experiments were carried out on a roller rig. The time–frequency analysis is a useful tool for identifying the content of a signal in the frequency domain without losing information about its time domain characteristics. Because of this, it is widely used for dynamic system analysis and condition monitoring and has been used in this paper for the detection of wheel flats and rail surface defects. Three commonly used time–frequency analysis techniques: Short-Time Fourier Transform, Wigner–Ville transform and wavelet transform were investigated in this work.
68 citations
TL;DR: In this study, a method is proposed for crackle detection using time-frequency and time-scale analysis from pulmonary signals and the results of individual feature sets and ensemble of sets, which are extracted using different window and wavelet types, for both pre- processed and non-pre-processed data with different machine learning algorithms are extensively evaluated and compared.
TL;DR: A TF mitigation technique is proposed to enable GNSS receivers to reduce the interference affecting the incoming signal, and is effective in terms of acquisition and tracking performance and outperforms other algorithms described in the literature.
Abstract: Undesired interfering signals are considered to be one of the main threats to the correct behavior of new Global Navigation Satellite System (GNSS) receivers. There is a huge variety of interference that can occur in the real world, such as narrow band, wide band, impulsive, stationary, or nonstationary CWs. Processing techniques based on time-frequency (TF) distributions allow one to detect and mitigate many different types of these undesired signals. A TF mitigation technique is proposed to enable GNSS receivers to reduce the interference affecting the incoming signal. The proposed algorithm uses a synthesis technique based on the orthogonal-like Gabor expansion, in order to obtain an estimate of the interference, that is then subtracted from the input signal. In the presence of a wide class of deterministic interfering signals, the proposed mitigation strategy is effective in terms of acquisition and tracking performance and outperforms other algorithms described in the literature.
TL;DR: In this article, the authors proposed a method for extracting second-order cyclostationary components from a vibration signal, which allows to estimate the amount of energy of each cyclic component of interest in the time-frequency domain.
09 Sep 2013
TL;DR: It is shown that the sparse signal reconstruction methods applied to the time-lag domain improve the TFR over the direct application of Fourier transform to the IAF and that the use of signal-adaptive kernels provides superior performance compared to data-independent kernels when missing data are present.
Abstract: In this paper, we examine the time-frequency representation (TFR) and sparse reconstruction of non-stationary signals in the presence of missing data samples. These samples lend themselves to missing entries in the instantaneous auto-correlation function (IAF) which, in turn, induce artifacts in the time-frequency distribution and ambiguity function. The artifacts are additive noise-like and, as such, can be mitigated by using proper time-frequency kernels. We show that the sparse signal reconstruction methods applied to the time-lag domain improve the TFR over the direct application of Fourier transform to the IAF. Additionally, the paper demonstrates that the use of signal-adaptive kernels provides superior performance compared to data-independent kernels when missing data are present.
TL;DR: A lower-bound for the uncertainty product of a signal in the two LCT domains is proposed that is sharper than those in the existing literature, and the conditions that give rise to the equal relation of the new uncertainty principle are deduced.
Abstract: This study devotes to uncertainty principles under the linear canonical transform (LCT) of a complex signal. A lower-bound for the uncertainty product of a signal in the two LCT domains is proposed that is sharper than those in the existing literature. We also deduce the conditions that give rise to the equal relation of the new uncertainty principle. The uncertainty principle for the fractional Fourier transform is a particular case of the general result for LCT. Examples, including simulations, are provided to show that the new uncertainty principle is truly sharper than the latest one in the literature, and illustrate when the new and old lower bounds are the same and when different.
TL;DR: In this article, a non-linear Lamb wave signal processing strategy aimed at extending the capability of activepassive networks of PZT transducers for defect detection is proposed, which allows to use chirp shaped pulses in actuation, instead of classically applied spiky pulses, requiring thus lower input voltages.
TL;DR: In this article, a time-frequency signal processing procedure aimed at extending pulse-echo defect detection methods based on guided waves to irregular waveguides is proposed, which returns the distance traveled by a guided wave that has propagated along a waveguide composed by segments with different dispersive properties by processing the detected echo signal.
Abstract: A time–frequency signal processing procedure aimed at extending pulse-echo defect detection methods based on guided waves to irregular waveguides is proposed. In particular, the procedure returns the distance traveled by a guided wave that has propagated along a waveguide composed by segments with different dispersive properties by processing the detected echo signal. To such aim, the acquired signal is processed by means of a two-step procedure. First, a warped frequency transform (WFT) is used to compensate the dispersion of the guided wave due to the traveled distance in a portion of the waveguide that is assumed as reference. Next, a further compensation is applied to remove from the warped signal the group delay introduced by the remaining irregular portion of the waveguide. Thanks to this processing strategy, the actual distance traveled by the wave in the regular portion of the irregular waveguide is revealed. Thus, the proposed procedure is suitable for automatically locate defect-induced reflections in irregular waveguides and can be easily implemented in real applications for structural health monitoring purposes. The potential of the procedure is demonstrated and validated numerically by simulating and processing Lamb waves propagating in waveguides made up of different uniform, tapered and curved segments.
26 May 2013
TL;DR: Simulations showed that perfect reconstruction can be achieved using fast iterative methods and preconditioning even using one filter per ERB and a very low redundancy, and comparison with a linear gammatone filterbank showed that the ERBlet approximates well the auditory time-frequency resolution.
Abstract: This paper describes a method for obtaining a perceptually motivated and perfectly invertible time-frequency representation of a sound signal. Based on frame theory and the recent non-stationary Gabor transform, a linear representation with resolution evolving across frequency is formulated and implemented as a non-uniform filterbank. To match the human auditory time-frequency resolution, the transform uses Gaussian windows equidistantly spaced on the psychoacoustic “ERB” frequency scale. Additionally, the transform features adaptable resolution and redundancy. Simulations showed that perfect reconstruction can be achieved using fast iterative methods and preconditioning even using one filter per ERB and a very low redundancy (1.08). Comparison with a linear gammatone filterbank showed that the ERBlet approximates well the auditory time-frequency resolution.
TL;DR: A method for detecting slow-moving weak targets in sea clutter based on time-frequency iteration decomposition, which demonstrates that it not only detects the slow- Moving weak target but also shows its instantaneous state.
Abstract: The echo scattered from a slow-moving weak target on sea surface is nonstationary due to the influence of waves. Time-frequency distributions are good tools to analyze it. In this paper, we propose a method for detecting slow-moving weak targets in sea clutter, which is based on time-frequency iteration decomposition. This method consists of three stages. First, we present a fast signal synthesis method (FSSM) based on eigenvalue decomposition. The FSSM can synthesize a signal faster and more accurately from the Wigner distribution (WD). Then, we present a signal iteration decomposition method (IDM) from the masked WD and the FSSM. By the IDM, the small component of a signal can be obtained, even when it is very close to a large component in the time-frequency plane. Finally, the proposed method results from the IDM and two criteria. Here the two criteria are defined to select the target signal. The proposed method is evaluated by X-band sea echo with a weak simulated target or a real target. The results demonstrate that it not only detects the slow-moving weak target but also shows its instantaneous state.
TL;DR: A systematic approach to the design, implementation and analysis of left-invariant evolution schemes acting on Gabor transform, primarily for applications in signal and image analysis and an explicit medical imaging application is introduced.
TL;DR: A comparison with short-time Fourier transform (STFT) and continuous wavelet transform (CWT) for both synthesized and real EEG data is performed and results show that the proposed algorithm can provide optimal time-frequency resolution as compared to STFT and CWT.
Abstract: Time-Frequency analysis of electroencephalogram (EEG) during different mental tasks received significant attention. As EEG is non-stationary, time-frequency analysis is essential to analyze brain states during different mental tasks. Further, the time-frequency information of EEG signal can be used as a feature for classification in brain-computer interface (BCI) applications. To accurately model the EEG, band-limited multiple Fourier linear combiner (BMFLC), a linear combination of truncated multiple Fourier series models is employed. A state-space model for BMFLC in combination with Kalman filter/smoother is developed to obtain accurate adaptive estimation. By virtue of construction, BMFLC with Kalman filter/smoother provides accurate time-frequency decomposition of the bandlimited signal. The proposed method is computationally fast and is suitable for real-time BCI applications. To evaluate the proposed algorithm, a comparison with short-time Fourier transform (STFT) and continuous wavelet transform (CWT) for both synthesized and real EEG data is performed in this paper. The proposed method is applied to BCI Competition data IV for ERD detection in comparison with existing methods. Results show that the proposed algorithm can provide optimal time-frequency resolution as compared to STFT and CWT. For ERD detection, BMFLC-KF outperforms STFT and BMFLC-KS in real-time applicability with low computational requirement.
TL;DR: An application example of a low-pass finite impulse response fractional order differentiating filter in the FrFT domain based on the definition of Caputo fractional differintegral method has been investigated taking into account amplitude-modulated signal corrupted with high-frequency chirp noise.
Abstract: This paper proposes a novel closed-form analytical expression of the fractional derivative of a signal in the Fourier transform (FT) and the fractional Fourier transform (FrFT) domain by utilizing the fundamental principles of the fractional order calculus. The generalization of the differentiation property in the FT and the FrFT domain to the fractional orders has been presented based on the Caputo's definition of the fractional differintegral, thereby achieving the flexibility of different rotation angles in the time-frequency plane with varying fractional order parameter. The closed-form analytical expression is derived in terms of the well-known higher transcendental function known as confluent hypergeometric function. The design examples are demonstrated to show the comparative analysis between the established and the proposed method for causal signals corrupted with high-frequency chirp noise and it is shown that the fractional order differentiating filter based on Caputo's definition is presenting good performance than the established results. An application example of a low-pass finite impulse response fractional order differentiating filter in the FrFT domain based on the definition of Caputo fractional differintegral method has also been investigated taking into account amplitude-modulated signal corrupted with high-frequency chirp noise.
TL;DR: The results establish a lower bound for the nonlinearity and complexity of the algorithms employed by the authors' brains in parsing transient sounds, rule out simple "linear filter" models of early auditory processing, and highlight timing acuity as a central feature in auditory object processing.
Abstract: The time-frequency uncertainty principle states that the product of the temporal and frequency extents of a signal cannot be smaller than 1/(4 π). We study human ability to simultaneously judge the frequency and the timing of a sound. Our subjects often exceeded the uncertainty limit, sometimes by more than tenfold, mostly through remarkable timing acuity. Our results establish a lower bound for the nonlinearity and complexity of the algorithms employed by our brains in parsing transient sounds, rule out simple "linear filter" models of early auditory processing, and highlight timing acuity as a central feature in auditory object processing.
TL;DR: In this article, the authors combine the concepts of time-frequency manifold (TFM) and image template matching, and propose a novel TFM correlation matching method to enhance identification of the periodic faults.
TL;DR: The threshold principle in the process of using the wavelet transform to de-noise the system is researches, and EMD (empirical mode decomposition) with wavelet threshold de- noising is combined to solve the problem.
Abstract: gear transmission system is a complex non-stationary and nonlinear time-varying coupling system. When faults occur on gear system, it is difficult to extract the fault feature. This paper researches the threshold principle in the process of using the wavelet transform to de-noise the system, and combines EMD (empirical mode decomposition) with wavelet threshold de-noising to solve the problem. The wavelet threshold de-noising is acts on the high-frequency IMF (Intrinsic Mode Function) component of the signal, and does overcome the defect by simply highlighting the fault feature. On this basis, the pre-processed signal is analyzed by the method of time-frequency analysis to extract the feature of the signal. The result shows that the SNR (signal-noise ratio) of the signal was largely improved, and the fault feature of the signal can therefore be effectively extracted. Combined with time-frequency analyses in the different running conditions (300 rpm, 900 rpm), various faults such as tooth root crack, tooth wear and multi-fault can be identified effectively. Based on this theory and combining the merits of MATLAB and VC
TL;DR: Various novel block based time–frequency domain adaptive filter structures for cardiac signal enhancement are presented and the performance of the block based algorithms is superior to the LLMS counterparts in terms of signal to noise ratio improvement (SNRI), excess mean square error (EMSE) and misadjustment (M).
TL;DR: Four TFD algorithms to minimise both the computation and memory loads are presented, optimised for a specific kernel category.
TL;DR: This paper describes one such approach, based upon ordinary least squares deconvolution of induced responses to input functions encoding the onset of different components within each trial, and considers optimal forms for convolution models ofinduced responses, in terms of impulse response basis function sets.
TL;DR: Preliminary results strongly support the vision of seeking the right form of sparsity for the right application to enable sparsity-cognizant estimation of robust parametric models for nonstationary signal analysis, through readily appreciated applications in frequency-hopping communications and speech compression.
Abstract: Recent research and experimental findings, as well as technological development and commercialization efforts, suggest that even a modest amount of data can deliver superior signal modeling and reconstruction performance if sparsity is present and accounted for. Early sparsity-aware signal processing techniques have mostly targeted stationary signal analysis using offline algorithms for signal and image reconstruction from Fourier samples. On the other hand, sparsity-aware time-frequency tools for nonstationary signal analysis have recently received growing attention. In this context, sparse regression has offered a new paradigm for instantaneous frequency estimation, over classical time-frequency representations. Standard techniques for estimating model parameters from time series yield erroneous fits when, e.g., abrupt changes or outliers cause model mismatches. Accordingly, the need arises for basic research in robust processing of nonstationary parametric models that leverage sparsity to accomplish tasks such as tracking of signal variations, outlier rejection, robust parameter estimation, and change detection. This article aims at delineating the analytical background of sparsity-aware time-series analysis and introducing sparsity-aware robust and nonstationary parametric models to the signal processing readership, through readily appreciated applications in frequency-hopping (FH) communications and speech compression. Preliminary results strongly support the vision of seeking the right form of sparsity for the right application to enable sparsity-cognizant estimation of robust parametric models for nonstationary signal analysis.
TL;DR: In this paper, a modified S-Transform (MST) is proposed as a simple means of controlling temporal and frequency resolution as a function of frequency, which is accomplished by replacing the frequency in the normalized Gaussian window of the S-transform with a linear frequency function.
Abstract: Although the S-Transform (ST) has better time resolution than the Continuous Wavelet Transform (CWT) at low frequencies, resolution may still not be adequate for certain seismic interpretation purposes. A modified S-Transform (MST), which reduces from the Generalized S-Transform (GST) is proposed as a simple means of controlling temporal and frequency resolution as a function of frequency. This is accomplished by replacing the frequency in the normalized Gaussian window of the S-transform with a linear frequency function. With appropriate choice of slope and intercept of this linear function, greatly improved temporal resolution can be accomplished at low frequencies, while retaining the temporal resolution of the ST at high frequencies. Synthetic examples show that low frequency ST spectra may suffer from interferences from nearby reflectors, and that this effect can be greatly reduced or eliminated using the MST. Application to real seismic data indicates that the MST can be useful in direct hydrocarbon indication.