Showing papers on "Noise (signal processing) published in 2004"

Stable recovery of sparse overcomplete representations in the presence of noise

Abstract: Overcomplete representations are attracting interest in signal processing theory, particularly due to their potential to generate sparse representations of signals. However, in general, the problem of finding sparse representations must be unstable in the presence of noise. This paper establishes the possibility of stable recovery under a combination of sufficient sparsity and favorable structure of the overcomplete system. Considering an ideal underlying signal that has a sufficiently sparse representation, it is assumed that only a noisy version of it can be observed. Assuming further that the overcomplete system is incoherent, it is shown that the optimally sparse approximation to the noisy data differs from the optimally sparse decomposition of the ideal noiseless signal by at most a constant multiple of the noise level. As this optimal-sparsity method requires heavy (combinatorial) computational effort, approximation algorithms are considered. It is shown that similar stability is also available using the basis and the matching pursuit algorithms. Furthermore, it is shown that these methods result in sparse approximation of the noisy data that contains only terms also appearing in the unique sparsest representation of the ideal noiseless sparse signal.

Signal processing apparatus and method

Abstract: A method and an apparatus to analyze two measured signals that are modeled as containing desired and undesired portions such as noise, FM and AM modulation. Coefficients relate the two signals according to a model defined in accordance with the present invention. In one embodiment, a transformation is used to evaluate a ratio of the two measured signals in order to find appropriate coefficients. The measured signals are then fed into a signal scrubber which uses the coefficients to remove the unwanted portions. The signal scrubbing is performed in either the time domain or in the frequency domain. The method and apparatus are particularly advantageous to blood oximetry and pulserate measurements. In another embodiment, an estimate of the pulserate is obtained by applying a set of rules to a spectral transform of the scrubbed signal. In another embodiment, an estimate of the pulserate is obtained by transforming the scrubbed signal from a first spectral domain into a second spectral domain. The pulserate is found by identifying the largest spectral peak in the second spectral domain.

Estimation of number of spectrally distinct signal sources in hyperspectral imagery

Abstract: With very high spectral resolution, hyperspectral sensors can now uncover many unknown signal sources which cannot be identified by visual inspection or a priori. In order to account for such unknown signal sources, we introduce a new definition, referred to as virtual dimensionality (VD) in this paper. It is defined as the minimum number of spectrally distinct signal sources that characterize the hyperspectral data from the perspective view of target detection and classification. It is different from the commonly used intrinsic dimensionality (ID) in the sense that the signal sources are determined by the proposed VD based only on their distinct spectral properties. These signal sources may include unknown interfering sources, which cannot be identified by prior knowledge. With this new definition, three Neyman-Pearson detection theory-based thresholding methods are developed to determine the VD of hyperspectral imagery, where eigenvalues are used to measure signal energies in a detection model. In order to evaluate the performance of the proposed methods, two information criteria, an information criterion (AIC) and minimum description length (MDL), and the factor analysis-based method proposed by Malinowski, are considered for comparative analysis. As demonstrated in computer simulations, all the methods and criteria studied in this paper may work effectively when noise is independent identically distributed. This is, unfortunately, not true when some of them are applied to real image data. Experiments show that all the three eigenthresholding based methods (i.e., the Harsanyi-Farrand-Chang (HFC), the noise-whitened HFC (NWHFC), and the noise subspace projection (NSP) methods) produce more reliable estimates of VD compared to the AIC, MDL, and Malinowski's empirical indicator function, which generally overestimate VD significantly. In summary, three contributions are made in this paper, 1) an introduction of the new definition of VD, 2) three Neyman-Pearson detection theory-based thresholding methods, HFC, NWHFC, and NSP derived for VD estimation, and 3) experiments that show the AIC and MDL commonly used in passive array processing and the second-order statistic-based Malinowski's method are not effective measures in VD estimation.

OFDM systems in the presence of phase noise: consequences and solutions

Abstract: We provide an exact analysis of orthogonal frequency-division multiplexing (OFDM) performance in the presence of phase noise Unlike most methods which assume small phase noise, we examine the general case for any phase noise levels After deriving a closed-form expression for the signal-to-noise-plus-interference ratio (SINR), we exhibit the effects of phase noise by precisely expressing the OFDM system performance as a function of its critical parameters This helps in understanding the meaning of small phase noise and how it reflects on the proper parameters selection of a specific OFDM system In order to combat phase noise, we also provide in this paper a general phase-noise suppression scheme, which, by analytical and numerical results, proves to be quite effective in practice

Survey of Image Denoising Techniques

Abstract: Removing noise from the original signal is still a challenging problem for researchers. There have been several published algorithms and each approach has its assumptions, advantages, and limitations. This paper presents a review of some significant work in the area of image denoising. After a brief introduction, some popular approaches are classified into different groups and an overview of various algorithms and analysis is provided. Insights and potential future trends in the area of denoising are also discussed.

Maximum Likelihood Estimation of Signal Amplitude and Noise Variance From MR Data

Abstract: In MRI, the raw data, which are acquired in spatial frequency space, are intrinsically complex valued and corrupted by Gaussian-distributed noise. After applying an inverse Fourier transform, the data remain complex valued and Gaussian distributed. If the signal amplitude is to be estimated, one has two options. It can be estimated directly from the complex valued data set, or one can first perform a magnitude operation on this data set, which changes the distribution of the data from Gaussian to Rician, and estimate the signal amplitude from the obtained magnitude image. Similarly, the noise variance can be estimated from both the complex and magnitude data sets. This article addresses the question whether it is better to use complex valued data or magnitude data for the estimation of these parameters using the maximum likelihood method. As a performance criterion, the mean-squared error (MSE) is used.

Natural excitation technique and eigensystem realization algorithm for phase i of the iasc-asce benchmark problem: simulated data

Abstract: A benchmark study in structural health monitoring based on simulated structural response data was developed by the joint IASC–ASCE Task Group on Structural Health Monitoring. This benchmark study was created to facilitate a comparison of various methods employed for the health monitoring of structures. The focus of the problem is simulated acceleration response data from an analytical model of an existing physical structure. Noise in the sensors is simulated in the benchmark problem by adding a stationary, broadband signal to the responses. A structural health monitoring method for determining the location and severity of damage is developed and implemented herein. The method uses the natural excitation technique in conjunction with the eigensystem realization algorithm for identification of modal parameters, and a least squares optimization to estimate the stiffness parameters. Applying this method to both undamaged and damaged response data, a comparison of results gives indication of the location and extent of damage. This method is then applied using the structural response data generated with two different models, different excitations, and various damage patterns. The proposed method is shown to be effective for damage identification. Additionally the method is found to be relatively insensitive to the simulated sensor noise.

Time-delay estimation via linear interpolation and cross correlation

Abstract: Time-delay estimation (TDE), which aims at measuring the relative time difference of arrival (TDOA) between different channels is a fundamental approach for identifying, localizing, and tracking radiating sources Recently, there has been a growing interest in the use of TDE based locator for applications such as automatic camera steering in a room conferencing environment where microphone sensors receive not only the direct-path signal, but also attenuated and delayed replicas of the source signal due to reflections from boundaries and objects in the room This multipath propagation effect introduces echoes and spectral distortions into the observation signal, termed as reverberation, which severely deteriorates a TDE algorithm in its performance This paper deals with the TDE problem with emphasis on combating reverberation using multiple microphone sensors The multichannel cross correlation coefficient (MCCC) is rederived here, in a new way, to connect it to the well-known linear interpolation technique Some interesting properties and bounds of the MCCC are discussed and a recursive algorithm is introduced so that the MCCC can be estimated and updated efficiently when new data snapshots are available We then apply the MCCC to the TDE problem The resulting new algorithm can be treated as a natural generalization of the generalized cross correlation (GCC) TDE method to the multichannel case It is shown that this new algorithm can take advantage of the redundancy provided by multiple microphone sensors to improve TDE against both reverberation and noise Experiments confirm that the relative time-delay estimation accuracy increases with the number of sensors

A magnitude/phase-locked loop system based on estimation of frequency and in-phase/quadrature-phase amplitudes

Abstract: This paper introduces a new phase-locked loop (PLL) system. The proposed system provides the dominant frequency component of the input signal and estimates its frequency. The mechanism of the proposed PLL is based on estimating in-phase and quadrature-phase amplitudes of the desired signal and, hence, has application advantages for communication systems which employ quadrature modulation techniques. The studies demonstrate that the proposed PLL also provides a superior performance for power system applications. Derivation of the mathematical model and theoretical stability analysis of the proposed PLL are carried out using dynamical systems theory. Advantages of the proposed PLL over the conventional PLLs are its capability of providing the fundamental component of the input signal which is not only locked in phase but also in amplitude to the actual signal while providing an estimate of its frequency. Computer simulation is used to evaluate its performance. Evaluations confirm structural robustness of the proposed PLL with respect to noise and distortions.

Biosignal and Medical Image Processing

Abstract: Introduction Biosignals Biosignal Measurement Systems Transducers Amplifier/Detector Analog Signal Processing and Filters ADC Conversion Data Banks Summary Problems Biosignal Measurements, Noise, and Analysis Biosignals Noise Signal Analysis: Data Functions and Transforms Summary Problems Spectral Analysis: Classical Methods Introduction Fourier Series Analysis Power Spectrum Spectral Averaging: Welch's Method Summary Problems Noise Reduction and Digital Filters Noise Reduction Noise Reduction through Ensemble Averaging Z-Transform Finite Impulse Response Filters Infinite Impulse Response Filters Summary Problems Modern Spectral Analysis: The Search for Narrowband Signals Parametric Methods Nonparametric Analysis: Eigenanalysis Frequency Estimation Problems TimeFrequency Analysis Basic Approaches The Short-Term Fourier Transform: The Spectrogram The WignerVille Distribution: A Special Case of Cohen's Class Cohen's Class Distributions Summary Problems Wavelet Analysis Introduction Continuous Wavelet Transform Discrete Wavelet Transform Feature Detection: Wavelet Packets Summary Problems Optimal and Adaptive Filters Optimal Signal Processing: Wiener Filters 8.2 Adaptive Signal Processing 8.3 Phase-Sensitive Detection 8.4 Summary Problems Multivariate Analyses: Principal Component Analysis and Independent Component Analysis Introduction: Linear Transformations Principal Component Analysis Independent Component Analysis Summary Problems Chaos and Nonlinear Dynamics Nonlinear Systems Phase Space Estimating the Embedding Parameters Quantifying Trajectories in Phase Space: The Lyapunov Exponent Nonlinear Analysis: The Correlation Dimension Tests for Nonlinearity: Surrogate Data Analysis Summary Exercises Nonlinearity Detection: Information-Based Methods Information and Regularity Mutual Information Function Spectral Entropy Phase-Space-Based Entropy Methods Detrended Fluctuation Analysis Summary Problems Fundamentals of Image Processing: The MATLAB Image Processing Toolbox Image-Processing Basics: MATLAB Image Formats Image Display Image Storage and Retrieval Basic Arithmetic Operations Block-Processing Operations Summary Problems Image Processing: Filters, Transformations, and Registration Two-Dimensional Fourier Transform Linear Filtering Spatial Transformations Image Registration Summary Problems Image Segmentation Introduction Pixel-Based Methods Continuity-Based Methods Multithresholding Morphological Operations Edge-Based Segmentation Summary Problems Image Acquisition and Reconstruction Imaging Modalities CT, PET, and SPECT Magnetic Resonance Imaging Functional MRI Summary Problems Classification I: Linear Discriminant Analysis and Support Vector Machines Introduction Linear Discriminators Evaluating Classifier Performance Higher Dimensions: Kernel Machines Support Vector Machines Machine Capacity: Overfitting or "Less Is More" Extending the Number of Variables and Classes Cluster Analysis Summary Problems Classification II: Adaptive Neural Nets Introduction Training the McCulloughPitts Neuron The Gradient Decent Method or Delta Rule Two-Layer Nets: Back Projection Three-Layer Nets Training Strategies Multiple Classifications Multiple Input Variables Summary Problems Appendix A: Numerical Integration in MATLAB Appendix B: Useful MATLAB Functions Bibliography Index

Bootstrap techniques for signal processing

Abstract: The statistical bootstrap is one of the methods that can be used to calculate estimates of a certain number of unknown parameters of a random process or a signal observed in noise, based on a random sample. Such situations are common in signal processing and the bootstrap is especially useful when only a small sample is available or an analytical analysis is too cumbersome or even impossible. This book covers the foundations of the bootstrap, its properties, its strengths and its limitations. The authors focus on bootstrap signal detection in Gaussian and non-Gaussian interference as well as bootstrap model selection. The theory developed in the book is supported by a number of useful practical examples written in MATLAB. The book is aimed at graduate students and engineers, and includes applications to real-world problems in areas such as radar and sonar, biomedical engineering and automotive engineering.

Challenges and Recent Developments in Hearing Aids: Part I. Speech Understanding in Noise, Microphone Technologies and Noise Reduction Algorithms

Abstract: This review discusses the challenges in hearing aid design and fitting and the recent developments in advanced signal processing technologies to meet these challenges. The first part of the review discusses the basic concepts and the building blocks of digital signal processing algorithms, namely, the signal detection and analysis unit, the decision rules, and the time constants involved in the execution of the decision. In addition, mechanisms and the differences in the implementation of various strategies used to reduce the negative effects of noise are discussed. These technologies include the microphone technologies that take advantage of the spatial differences between speech and noise and the noise reduction algorithms that take advantage of the spectral difference and temporal separation between speech and noise. The specific technologies discussed in this paper include first-order directional microphones, adaptive directional microphones, second-order directional microphones, microphone matching algorithms, array microphones, multichannel adaptive noise reduction algorithms, and synchrony detection noise reduction algorithms. Verification data for these technologies, if available, are also summarized.

Noise-dependent postfiltering

Abstract: The paper introduces a modification of the commonly used postfilter that improves performance when acoustic background noise is present. The modification consists of replacing the nonadaptive postfilter parameters that govern the degree of spectral emphasis (commonly denoted as /spl gamma//sub 1/ and /spl gamma//sub 2/) with parameters that adapt to the noise statistics. We describe an effective mapping from the noise statistics to the emphasis parameters and provide a low complexity noise estimation algorithm that is sufficient for this application. The resulting noise-adaptive postfilter successfully attenuates the background noise and naturally converges to the conventional postfilter at high SNR conditions. Thus, the speech enhancement problem is solved with minimal modification of legacy codecs, since the existing structure of the speech codec is used. Test results indicate that the presented algorithm significantly outperforms the standard postfilter with non-adaptive parameters.

Speech enhancement based on the general transfer function GSC and postfiltering

Abstract: In speech enhancement applications microphone array postfiltering allows additional reduction of noise components at a beamformer output. Among microphone array structures the recently proposed general transfer function generalized sidelobe canceller (TF-GSC) has shown impressive noise reduction abilities in a directional noise field, while still maintaining low speech distortion. However, in a diffused noise field less significant noise reduction is obtainable. The performance is even further degraded when the noise signal is nonstationary. In this contribution we propose three postfiltering methods for improving the performance of microphone arrays. Two of which are based on single-channel speech enhancers and making use of recently proposed algorithms concatenated to the beamformer output. The third is a multichannel speech enhancer which exploits noise-only components constructed within the TF-GSC structure. This work concentrates on the assessment of the proposed postfiltering structures. An extensive experimental study, which consists of both objective and subjective evaluation in various noise fields, demonstrates the advantage of the multichannel postfiltering compared to the single-channel techniques.

Transmitter identification using embedded pseudo random sequences

Abstract: A transmitter identification system for DTV distributed transmission network using embedded pseudo random sequences is investigated. Different orthogonal pseudo random sequences and their suitability for transmitter identification are discussed. Code generators are developed to study the auto-correlation and cross-correlation properties of the Kasami sequences. To speed up the identification process, the embedded pseudo random sequence is preferred to be time-synchronized with the DTV frame structure. Therefore, the length of the identification code has to be truncated before it is fitted into each field of the ATSC DTV signal. The impact of truncation noise and in-band DTV interference on transmitter identification is also investigated. It is shown that the auto-correlation and cross-correlation properties are only slightly affected by truncation. It is also found that the dominant interference to the transmitter identification is the in-band DTV signal. The signal to truncation noise ratio and signal to DTV interference ratio in the correlation output are derived, and verified via simulation. It is further recognized that in-band DTV interference can only be mitigated by increasing the code length or by time-domain averaging technique to smoothen out the in-band interference.

Noise characteristics of fiber optical parametric amplifiers

Abstract: We derive an analytical theory for the noise figure of an undepleted and lossless fiber optical parametric amplifier (FOPA). Both the signal and the wavelength converted idler are investigated. Our theory is applicable for both an ideal pump power source, as well as a noisy one. We find that a noisy pump source can severely degrade the performance at high gain due to the stochastic gain-variations the signal and idler will experience. The theory is compared with Monte Carlo simulations of the FOPA and an excellent agreement is obtained. Simulations in the gain-depleted region show the possibility to reach below quantum-limited, phase-insensitive amplification for single channel transmission.

Phase-based dual-microphone robust speech enhancement

Abstract: A dual-microphone speech-signal enhancement algorithm, utilizing phase-error based filters that depend only on the phase of the signals, is proposed. This algorithm involves obtaining time-varying, or alternatively, time-frequency (TF), phase-error filters based on prior knowledge regarding the time difference of arrival (TDOA) of the speech source of interest and the phases of the signals recorded by the microphones. It is shown that by masking the TF representation of the speech signals, the noise components are distorted beyond recognition while the speech source of interest maintains its perceptual quality. This is supported by digit recognition experiments which show a substantial recognition accuracy rate improvement over prior multimicrophone speech enhancement algorithms. For example, for a case with two speakers with a 0.1 s reverberation time, the phase-error based technique results in a 28.9% recognition rate gain over the single channel noisy signal, a gain of 22.0% over superdirective beamforming, and a gain of 8.5% over postfiltering.

Benign interference suppression for received signal quality estimation

Abstract: A receiver circuit suppresses effects of “benign” impairment from the calculation of received signal quality estimates, such that the estimate depends primarily on the effects of non-benign impairment. For example, a received signal may be subject to same-cell and other-cell interference plus noise, which is generally modeled using a Gaussian distribution, and also may be due to certain forms of self-interference, such as quadrature phase interference arising from imperfect derotation of the pilot samples used to generate channel estimates for the received signal. Such interference generally takes on a distribution defined by the pilot signal modulation, e.g., a binomial distribution for binary phase shift keying modulation. Interference arising from such sources is relatively “benign” as compared to Gaussian interference and thus should be suppressed or otherwise discounted in signal quality calculations. Suppression may be based on subtracting benign impairment correlation estimates from total impairment correlation estimates, or on filtering the benign impairment in channel estimation.

Flow Rate Measurements Using Flow-Induced Pipe Vibration

Abstract: This paper focuses on the possibility of a non-intrusive, low cost, flow rate measurement technique. The technique is based on signal noise from an accelerometer attached to the surface of the pipe. The signal noise is defined as the standard deviation of the frequency averaged time series signal. Experimental results are presented that indicate a nearly quadratic relationship between the signal noise and mass flow rate in the pipe. It is also shown that the signal noise - flow rate relationship is dependant on the pipe material and diameter.

Quantum noise properties of parametric amplifiers driven by two pump waves.

Abstract: In a parametric amplifier (PA) driven by two pump waves the signal sideband is coupled to three idler sidebands, all of which are frequency-converted (FC) images of the signal, and two of which are phase-conjugated (PC) images of the signal. If such a device is to be useful, the signal must be amplified, and the PC and FC idlers must be produced, with minimal noise. In this paper the quantum noise properties of two-sideband (TS) parametric devices are reviewed and the properties of many-sideband devices are determined. These results are applied to the study of two-pump PAs, which are based on the aforementioned four-sideband (FS) interaction. As a general guideline, the more sidebands that interact, the higher are the noise levels. However, if the pump frequencies are tuned to maximize the frequency bandwidth of the FS interaction, the signal and idler noise-figures are only slightly higher than the noise figures associated with the limiting TS interactions.

Signal enhancement by time-frequency peak filtering

Abstract: Time-frequency peak filtering (TFPF) allows the reconstruction of signals from observations corrupted by additive noise by encoding the noisy signal as the instantaneous frequency (IF) of a frequency modulated (FM) analytic signal. IF estimation is then performed on the analytic signal using the peak of a time-frequency distribution (TFD) to recover the filtered signal. This method is biased when the peak of the Wigner-Ville distribution (WVD) is used to estimate the encoded signal's instantaneous frequency. We characterize a class of signals for which the method implemented using the pseudo WVD is approximately unbiased. This class contains deterministic bandlimited nonstationary multicomponent signals in additive white Gaussian noise (WGN). We then derive the pseudo WVD window length that gives a reduced bias when TFPF is used for signals from this class. Testing of the method on both synthetic and real life newborn electroencephalogram (EEG) signals shows clean recovery of the signals in noise level down to a signal-to-noise ratio (SNR) of -9 dB.

Method and device for continuous monitoring of the concentration of an analyte

Abstract: A method for continuous monitoring of the concentration of an analyte by determining its change over time in the living body of a human or animal, in which at sequential points in time, measurement values of a measurement variable correlating with the desired concentration are measured as the measurement signal (z t) and the change over time of the concentration is determined from the measurement signal as the useful signal (y t) using a calibration, the determination of the useful signal (y t) from the measurement signal (z t) including a filter algorithm in the time domain, by which errors of the useful signal, which result from noise contained in the measurement signal, are reduced, and the filter algorithm including an operation in which the influence of an actual measurement value on the useful signal is weighted using a weighting factor (V). During the continuous monitoring, a signal variation parameter (.sigma.t) is determined on the basis of signal variations detected in close chronological relationship with the measurement of the actual measurement value. The weighting factor is dynamically adapted as a function of the signal variation parameter determined for the point in time of the actual measurement.

A sampling algorithm for digitally controlled boost PFC converters

Abstract: Digital control of a boost power factor correction (PFC) converter requires sampling of the input current. As the input current contains a considerable amount of switching ripple and high frequency switching noise, the choice of the sampling instant is very important. To avoid aliasing without employing a (very) high sampling frequency, the sampling is synchronized with the pulse width modulation (PWM). Sampling algorithms employing this technique successfully reject the input current ripple but are not immune to the high frequency switching noise present on all sampled signals. Therefore, a new sampling algorithm, called alternating-edge-sampling and intended for center-based or symmetric PWM, is deduced with as most important features: switching noise immunity, straightforwardness, accurate measurement of the averaged input current and the need for only few processor cycles. The operating principle, design issues and a theoretical study of the input current error induced by the sampling algorithm due to sampling instant timing errors are derived. All theoretical results are validated experimentally for a digitally controlled boost PFC converter switching at 50 kHz.

Lead selection for SSVEP-based brain-computer interface

Abstract: SSVEP-based brain-computer interface (BCI) has potential advantage of high information transfer rate However, individual difference greatly affects its practical applications This paper presents a method of lead selection to improve the applicability of SSVEP-based BCI system Independent component analysis (ICA) is employed to decompose EEGs over visual cortex into SSVEP signal and background noise Optimal bipolar lead is selected by comparing signal correlation and noise correlation between different channels The system with one optimal bipolar lead has reached an average transfer rate about 42bits/min for normal subjects It has also been successfully applied to an environmental controller for the motion-disabled

An information theoretic approach to source enumeration in array signal processing

Abstract: In this paper, a new information theoretic algorithm is proposed for signal enumeration in array processing. The approach is based on predictive description length (PDL) that is defined as the length of a predictive code for the set of observations. We assume that several models, with each model representing a certain number of sources, will compete. The PDL criterion is computed for the candidate models and is minimized over all models to select the best model and to determine the number of signals. In the proposed method, the correlation matrix is decomposed into two orthogonal components in the signal and noise subspaces. The maximum likelihood (ML) estimates of the angles-of-arrival are used to find the projection of the sample correlation matrix onto the signal and noise subspaces. The summation of the ML estimates of these matrices is the ML estimate of the correlation matrix. This method can detect both coherent and noncoherent signals. The proposed method can be used online and can be applied to time-varying systems and target tracking.

Statistical Wavelet-Based Method for Structural Health Monitoring

Abstract: A statistical pattern classification method based on wavelet packet transform (WPT) is developed in this study for structural health monitoring. The core of this method is the WPT with the ability of extracting minute abnormality from vibration signals. The vibration signals of a structure excited by a pulse load are first decomposed into wavelet packet components. Signal energies of these wavelet packet components are then calculated and sorted according to their magnitudes. Those components that are small in signal energy are discarded since they are easily contaminated by measurement noise. The remaining dominant component energies are defined as a novel condition index, the wavelet packet signature (WPS). Two damage indicators are then formulated to lump the discriminate information from the extracted WPS. Thresholds for damage alarming are established using the statistical properties and the 1-sided confidence limit of the damage indicators from successive measurements. To demonstrate, an experimental study on the health monitoring of a steel cantilever I beam is performed. Four damage cases involving line cuts of different severities in the flanges at 1 cross section are introduced. Vibration signals are obtained from an accelerometer installed at the free end of the beam. Results show that the health condition of the beam can be accurately monitored by the proposed method even when the signals are highly contaminated with noise. The proposed method does not require any prior knowledge of the structure being monitored and is suitable for on-line continuous monitoring of structural health condition.