Signal Feature Extraction Method Research Articles

Cardiac Multi-Frequency Vibration Signal Sensor Module and Feature Extraction Method Based on Vibration Modeling.

Cardiovascular diseases pose a long-term risk to human health. This study focuses on the rich-spectrum mechanical vibrations generated during cardiac activity. By combining Fourier series theory, we propose a multi-frequency vibration model for the heart, decomposing cardiac vibration into frequency bands and establishing a systematic interpretation for detecting multi-frequency cardiac vibrations. Based on this, we develop a small multi-frequency vibration sensor module based on flexible polyvinylidene fluoride (PVDF) films, which is capable of synchronously collecting ultra-low-frequency seismocardiography (ULF-SCG), seismocardiography (SCG), and phonocardiography (PCG) signals with high sensitivity. Comparative experiments validate the sensor's performance and we further develop an algorithm framework for feature extraction based on 1D-CNN models, achieving continuous recognition of multiple vibration features. Testing shows that the recognition coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) of the 8 features are 0.95, 2.18 ms, and 4.89 ms, respectively, with an average prediction speed of 60.18 us/point, meeting the re-quirements for online monitoring while ensuring accuracy in extracting multiple feature points. Finally, integrating the vibration model, sensor, and feature extraction algorithm, we propose a dynamic monitoring system for multi-frequency cardiac vibration, which can be applied to portable monitoring devices for daily dynamic cardiac monitoring, providing a new approach for the early diagnosis and prevention of cardiovascular diseases.

Status Recognition of Marine Centrifugal Pumps Based on a Stacked Sparse Auto-Encoder

Marine centrifugal pumps (MCPs) are widely used in ships, so it is important to identify their status accurately for their maintenance. Due to the influence of load, friction, and other non-linear factors, the vibration signal of an MCP shows non-linear and non-stationary characteristics, and it is difficult to extract the state characteristics contained in the vibration signal. To solve the difficulty of feature extraction of non-linear non-stationary vibration signals generated by MCPs, a novel MCP frequency domain signal feature extraction method based on a stacked sparse auto-encoder (SSAE) is proposed. The characteristic parameters of MCP frequency domain signals are extracted via the SSAE model for classification training, and different statuses of MCPs are identified. The vibration signals in different MCP statuses were collected for feature extraction and classification training, and the MCP status recognition accuracy based on the time domain feature and fuzzy entropy feature was compared. According to the test data, the accuracy of MCP status recognition based on the time domain feature is 71.2%, the accuracy of MCP status recognition based on the fuzzy entropy feature is 87.7%, and the accuracy of MCP status recognition based on the proposed method is 100%. These results show that the proposed method can accurately identify each status of an MCP under test conditions.

Compositional Development of Chinese Pipa Music in the New Media Era

Abstract In the new media era, the creation and development of Chinese pipa music faces new challenges and opportunities. This paper analyses the characteristics of pipa music and explores its innovative path and development trend in the contemporary environment through technical means. Music signal preprocessing, feature extraction and other methods are adopted to provide an in-depth characterization of pipa music. In the study, the pipa music was classified concerning pitch, loudness, timbre and rhythm, and the musical features were extracted using techniques such as MFCC and LPC. The experimental data contained 300 pieces of pipa music of different styles, and a series of specific numerical results were derived from the Characterization of this music. For example, in the MFCC feature retrieval experiment, the retrieval accuracy of a 30-second music clip was improved by 0.024 compared to a 10-second clip. In addition, by analyzing the timbre features of 12 pipas, and using MATLAB to process the data, a quantitative judgement of the goodness of the timbre was derived. The possibility of combining pipa music with electronic music was explored with new media technology. The artistic expression of pipa music can be effectively analyzed and enhanced through scientific methods, providing fresh ideas for the innovative development of pipa music.

Complementary Ensemble Empirical Mode Decomposition and Maximum Correlated Kurtosis Deconvolution for Wind Turbine Bearing Fault Feature Extraction

Wind turbine bearing fault signals exhibit characteristics such as nonlinearity, non-stationarity, and susceptibility to external noise interference, making it challenging to extract fault features and identify them accurately. In light of these issues, this paper proposes a fault signal feature extraction method that combines complementary ensemble empirical mode decomposition (CEEMD) and maximum correlated kurtosis deconvolution (MCKD). CEEMD is utilized to decompose the signals, reducing mode mixing and eliminating residual auxiliary noise in the decomposition process, thereby obtaining linear and stationary signals that enhance the significance of the features. MCKD is employed to increase the kurtosis value of the fault signals, facilitating the extraction of continuous transient impacts from weak fault signals. Finally, a Convolutional Neural Network (CNN) is employed to validate that the method enhances fault characteristics and improves the accuracy of fault signal recognition.

Signal Feature Extraction Method of Three-Band Infrared Flame Sensor

A three-band infrared flame sensor is designed and experimentally investigated for the analysis of infrared signals from three radiation sources (Flame, man and ambient). The Fourier transform and empirical mode decomposition (EMD) is applied to the analysis of the spectral characteristics, intrinsic modes and energy spectra. The results show that the received flame infrared signal exhibits significant low-frequency and energy characteristics. Near the zero frequency, the received infrared signal energy of the flame, man and ambient is 3.4 × 104 J, 2.8 × 104 J and 2.1 × 104 J, respectively. In this way the flame signal features can be accurately identified from different radiation sources. It can be believed that the results of the three-band infrared sensor presented here are useful guiding significance and practical application value for fire detection.

A Combined Virtual Electrode-Based ESA and CNN Method for MI-EEG Signal Feature Extraction and Classification.

A Brain-Computer Interface (BCI) is a medium for communication between the human brain and computers, which does not rely on other human neural tissues, but only decodes Electroencephalography (EEG) signals and converts them into commands to control external devices. Motor Imagery (MI) is an important BCI paradigm that generates a spontaneous EEG signal without external stimulation by imagining limb movements to strengthen the brain's compensatory function, and it has a promising future in the field of computer-aided diagnosis and rehabilitation technology for brain diseases. However, there are a series of technical difficulties in the research of motor imagery-based brain-computer interface (MI-BCI) systems, such as: large individual differences in subjects and poor performance of the cross-subject classification model; a low signal-to-noise ratio of EEG signals and poor classification accuracy; and the poor online performance of the MI-BCI system. To address the above problems, this paper proposed a combined virtual electrode-based EEG Source Analysis (ESA) and Convolutional Neural Network (CNN) method for MI-EEG signal feature extraction and classification. The outcomes reveal that the online MI-BCI system developed based on this method can improve the decoding ability of multi-task MI-EEG after training, it can learn generalized features from multiple subjects in cross-subject experiments and has some adaptability to the individual differences of new subjects, and it can decode the EEG intent online and realize the brain control function of the intelligent cart, which provides a new idea for the research of an online MI-BCI system.

Adaptive detection of impact signals with two-dimensional piecewise tri-stable stochastic resonance and its application in bearing fault diagnosis

In this paper, we introduce an innovative method known as Adaptive Two-Dimensional Piecewise Tri-Stable Stochastic Resonance (TDPTSR) and discuss its applicability in extracting weak fault features. Initially, addressing the saturation issue of the Standard Tri-Stable Stochastic Resonance (STSR) system, we construct a novel piecewise tri-stable potential function. This function independently adjusts the depth and width of potential wells. We explore the particle output characteristics of the Piecewise Tri-Stable Stochastic Resonance (PTSR) system using the fourth-order Runge-Kutta algorithm. Subsequently, we delve into the performance analysis of the TDPTSR system. Leveraging adiabatic approximation theory, we derive the equivalent potential function, steady-state probability density (SPD), mean first-pass time (MFPT), and output signal-to-noise ratio (SNR) of the TDPTSR. We also investigate the impact of changes in system parameters on these metrics. Additionally, drawing from kurtosis and cosine similarity properties, we introduce a Modified Kurtosis Index (MKC) as a measurement index for impact signal detection. We propose a method for impact signal feature extraction by integrating MKC with the TDPTSR system. Simulation analysis demonstrates the suitability of MKC as an index for quantifying Stochastic Resonance and the superior detection capability of the TDPTSR system. Finally, we apply TDPTSR to the fault diagnosis of two types of bearings, optimizing system parameters using a genetic algorithm (GA). Comparative analysis with STSR and PTSR confirms that our proposed method in this paper yields further improvements in Signal-to-Noise Ratio Improvement (SNRI), higher spectral peak magnitude (Amout) at characteristic frequencies, and increased recognition quantity (Δ).

Machine learning based hybrid anomaly detection technique for automatic diagnosis of cardiovascular diseases using cardiac sympathetic nerve activity and electrocardiogram.

Coronary artery diseases (CADs) are the leading cause of death worldwide and early diagnosis is crucial for timely treatment. To address this, our study presents a novel automated Artificial Intelligence (AI)-based Hybrid Anomaly Detection (AIHAD) technique that combines various signal processing, feature extraction, supervised, and unsupervised machine learning methods. By jointly and simultaneously analyzing 12-lead cardiac sympathetic nerve activity (CSNA) and electrocardiogram (ECG) data, the automated AIHAD technique performs fast, early, and accurate diagnosis of CADs. In order to develop and evaluate the proposed automated AIHAD technique, we utilized the fully labeled STAFF III and PTBD databases, which contain the 12-lead wideband raw recordings non-invasively acquired from 260 subjects. Using these wideband raw recordings, we developed a signal processing technique that simultaneously detects the 12-lead CSNA and ECG signals of all subjects. Using the pre-processed 12-lead CSNA and ECG signals, we developed a time-domain feature extraction technique that extracts the statistical CSNA and ECG features critical for the reliable diagnosis of CADs. Using the extracted discriminative features, we developed a supervised classification technique based on Artificial Neural Networks (ANNs) that simultaneously detects anomalies in the 12-lead CSNA and ECG data. Furthermore, we developed an unsupervised clustering technique based on Gaussian mixture models (GMMs) and Neyman-Pearson criterion, which robustly detects outliers corresponding to CADs. Using the automated AIHAD technique, we have, for the first time, demonstrated a significant association between the increase in CSNA signals and anomalies in ECG signals during CADs. The AIHAD technique achieved highly reliable detection of CADs with a sensitivity of 98.48 %, specificity of 97.73 %, accuracy of 98.11 %, positive predictive value of 97.74 %, negative predictive value of 98.47 %, and F1-score of 98.11 %. Hence, the automated AIHAD technique demonstrates superior performance compared to the gold standard diagnostic test ECG in the diagnosis of CADs. Additionally, it outperforms other techniques developed in this study that separately utilize either only CSNA data or only ECG data. Therefore, it significantly increases the detection performance of CADs by taking advantage of the diversity in different data types and leveraging their strengths. Furthermore, its performance is comparatively better than that of most previously proposed machine and deep learning methods that exclusively used ECG data to diagnose or classify CADs. Additionally, it has a very low implementation time, which is highly desirable for real-time detection of CADs. The proposed automated AIHAD technique may serve as an efficient decision-support system to increase physicians' success in fast, early, and accurate diagnosis of CADs. It may be highly beneficial and valuable, particularly for asymptomatic patients, for whom the diagnostic information provided by ECG alone is not sufficient to reliably diagnose the disease. Hence, it may significantly improve patient outcomes by enabling timely treatments and considerably reducing the mortality of cardiovascular diseases (CVDs).

Conditioning and monitoring of grinding wheels: A state-of-the-art review

<p>Grinding wheel condition monitoring is an important step towards the prediction of grinding wheel faulty conditions. It is beneficial to define techniques to minimize the wear of the grinding wheels and finally enhance the life of the grinding wheels. Grinding wheel condition monitoring is done by two techniques such as (i) direct and (ii) indirect. Direct monitoring employs optical sensors and computer vision techniques, and indirect monitoring is done by signal analysis such as acoustic emission (AE), vibration, cutting force, etc. Methods implemented for grinding wheel monitoring in the published research papers are reviewed. The review is compiled in five sections: (a) process parameters measurement, (b) data acquisition systems, (c) signal analysis techniques, (d) feature extraction, and (e) classification methods. In today’s era of Industry 4.0, a large amount of manufacturing data is generated in the industry. So, conventional machine learning techniques are insufficient to analyze real-time conditioning monitoring of the grinding wheels. However, deep learning techniques such as artificial neural network (ANN), convolutional neural network (CNN) have shown prediction accuracy above 99%.</p>

Abnormal event monitoring of underground pipelines using a distributed fiber-optic vibration sensing system

A distributed fiber-optic vibration sensing (DFOVS) system is developed for monitoring underground pipelines. This DFOVS has the advantages of simple structure, low cost, high sensitivity, wide frequency response range, and suitability for long-distance monitoring. Based on the DFOVS system, an abnormal pipeline event monitoring (APEM) method is proposed. This APEM method consists of three components: a signal feature extraction method that integrates signal feature calculations in time and frequency domains with a feature dimension reduction method, an abnormal event detection method based on multivariate statistical analysis, and an abnormal event localization method based on cross-correlation analysis. The DFOVS-based APEM method solves the problem of long-distance and real-time monitoring of various abnormal pipeline events, such as leakage, mechanical impact and excavation. Experiments results show that the proposed method has a detection rate higher than 96%, a false alarm rate lower than 2%, and an average localization error less than 14 m.

Biomedical signal processing methods for neuromarketing: A comparative study

Neuromarketing involves the integration of neuropsychology into marketing research, focusing on analyzing consumer sensory-motor actions, including cognitive and emotional responses to marketing stimuli, using advanced technologies. It represents one of the latest strategies in marketing research and has the potential to shape the future of the field. Numerous studies have already been conducted in this domain to enhance research outcomes. Nevertheless, the literature indicates that there are still opportunities for further advancements and improvements. A literature review was conducted in this article to explore the availability of an openly accessible dataset widely utilized by researchers in the field of neuromarketing. We examined the signal preprocessing, feature selection, feature extraction, and classification methods employed in studies utilizing the dataset created by Yadava et al.

Feature extraction method of HPLC communication signal based on genetic algorithm

AbstractCommunication quality is a key factor affecting the effectiveness of distribution network operation status management. Therefore, it is required that the communication performance of the distribution network management system be good, and the communication signal can directly reflect the communication quality. Therefore, a genetic algorithm (GA) based HPLC communication signal feature extraction method is proposed. The process of constructing reference samples, training recognition models, constructing noise samples, self‐coding denoising processing, and denoising sample recognition has completed the identification of power line channel transmission characteristics. The state feature quantity of the topological line is based on the positioning results of the diagnostic function, and the calculated value of the reflection coefficient calculation model at the head end of the topological line is compared with the measured value. By optimizing the feature quantities containing topological line states through GAs, abnormal feature quantities close to the true state of topological lines are obtained. Experimental analysis shows that this method can correctly identify the position of reactive power compensation in the transmission line, and can also correctly identify the extraction of topological anomaly features. Therefore, the research method is of great significance in improving the safe operation of power systems and extending the service life of topology lines.

The detection and classification method of sub-macroscopic defects inside steel with an ultrasonic testing and CatBoost-based stacking model

With the increasing demand for advanced steel, the internal cleanness of steel becomes an important evaluation indicator of material quality. Sub-macroscopic defects are randomly distributed inside steel materials, which have seriously affected material stability and fatigue life because they are not covered by existing testing standards. Besides, the existing detection methods generally have problems such as low efficiency and complexity. In this paper, we propose a non-destructive inclusion testing and classification framework based on ultrasonic testing experiments, signal feature extraction and machine-learning methods. Under the optimal experimental detection conditions we found through experiments, a large-scale sub-macroscopic inclusion signal data set is established to realize the classification of defects. Moreover, empirical mode decomposition (EMD) and other feature extraction algorithms are applied to further boost the model performance. We propose a CatBoost-based stacking fused model named Stacked-CBT, which obtains state-of-the-art experimental results with an accuracy rate of 86.65% and demonstrates that the proposed framework is feasible to classify the sub-macroscopic inclusion signals. To the best of our knowledge, there is no previous study in this field that has acquired such a large amount of experimental sub-macroscopic signal data while taking into consideration classification-specific designs.

Bearing failure diagnosis at time-varying speed based on adaptive clustered fractional Gabor transform

For bearing fault diagnosis at time-varying speed with tachometer-free and non-resampling, the crucial process is to obtain a high-resolution time-frequency representation and extract fault features. However, current multi-component non-stationary signal feature extraction methods based on time-frequency transform suffer from fixed parameter settings and insufficient resolution for low signal-to-noise ratio signals. To address these issues, a novel adaptive clustered fractional Gabor transform is proposed and applied to extract bearing fault features at time-varying speed. Firstly, the grey wolf optimization is utilized to adaptively search for the optimal fractional order and Gauss window length based on the maximum spectral kurtosis and the generalized time-bandwidth product to achieve the most adequate fractional Gabor spectrum (FrGS). Then, the Clustering by Fast Search and Find of Density Peaks algorithm reconstructs the sparse representation of the FrGS, remapping multi-component signals into single-component clusters. Bearing fault diagnosis is achieved by matching the relative order of each cluster with the bearing fault characteristic coefficients. Simulation signals validate the superiority of the feature extraction method, and experimental signals validate the feasibility of the bearing fault diagnosis method.

Feature Extraction of Motor Imagery EEG via Discrete Wavelet Transform and Generalized Maximum Fuzzy Membership Difference Entropy: A Comparative Study

Identifying motor imagery (MI) electroencephalogram (EEG) is an important way to achieve brain–computer interface (BCI), but its applicability is heavily dependent on the performance of feature extraction procedure. In this paper, a feature extraction method based on generalized maximum fuzzy membership difference entropy (GMFMDE) and discrete wavelet transform (DWT) was proposed for the feature extraction of EEG signals. The influence of different distance calculation methods, embedding dimensions and tolerances were studied to find the best configuration of GMFMDE for the feature extraction of MI–EEG. The gradient boosting decision tree (GBDT) classifier was used to classify the features extracted from GMFMDE and DWT. The average classification accuracy of 93.71% and the maximum classification accuracy of 96.96% were obtained, which proved the effectiveness of the proposed feature extraction method for EEG signal feature extraction.

Transformer Vibration Signal Feature Extraction Based on Compressed Sensing and Wavelet Packet

Transformers are important equipment in the power system. Aiming at the collection, transmission, storage, and processing of massive high-dimensional vibration data in the process of transformer vibration online monitoring, this paper proposes a transformer vibration signal feature extraction method based on compressed sensing and wavelet packets. Firstly, the PartHadamard measurement matrix is used to compress the transformer vibration signal, and then the characteristics of the transformer vibration signal are extracted based on the wavelet packet decomposition and normalized wavelet information entropy. The proposed method completes feature extraction on the measured vibration data of the transformer, and three classification algorithms are applied to carry out simulation experiments. The results show that the proposed feature extraction method can effectively extract the vibration characteristics reflecting the operating state of the transformer while greatly reducing the dimension of vibration data, which provides a reference for the online monitoring of transformer vibration.

Validation of target tracking performance through signal feature extraction method based on 1D convolutional neural network

In order to perform active vibration and noise control, a variety of adaptive algorithms have been investigated and developed. Especially, neural network-based signal tracking algorithms, such as the radial basis function neural network (RBFNN), are being widely utilized. For signal tracking problems, a reference signal has a significantly important role, while it is difficult to determine the reference signal when it has relatively complex spectrum. Thus, this study focuses on the signal feature extraction for manipulating appropriate reference signals easily and validation of signal tracking performance with it. In order to carried out the signal feature extraction, 1D convolutional neural network (1D CNN) is implemented and trained based on the CWRU bearing dataset. In addition, multi normalized least mean square (NLMS), neural network-based signal tracking algorithm, and diagonal recurrent neural network (DRNN) is employed to confirm the signal tracking performance. The proposed method shows that the 1D CNN-based signal feature extraction method could find the signal feature properly and the reference signal obtained from this methodology is employed to a target tracking problem, which shows a great performance.

Radar Moving Target Detection Method Based on SET2 and AlexNet

Aiming at the nonstationary characteristics of echo signal for a high-speed maneuvering target, a signal feature extraction method is proposed by combining the time-frequency analysis and convolution neural network, and then the automatic detection of radar moving target in a noisy environment is realized. Firstly, the echo signal is modelled as a more accurate Gaussian modulation-linear frequency modulation (GM-LFM) signal and converted into the time-frequency image by a second-order synchroextracting transform (SET2). Then, ridge extraction is applied to extract the maximum energy ridge from the time-frequency distribution, and the data set is constructed by the maximum energy ridge. Finally, the data set is input into AlexNet for training, and the deep-level features of echo signal are extracted to realize the automatic moving targets detection. Simulation results show that SET2 and RE can effectively enhance the time-frequency characteristics of echo signal under the noisy environment, and the detection accuracy and noise robustness of the proposed method are better than that of SET1 and smooth pseudo-Wigner–Ville distribution (SPWVD).

Automatic anomaly identification of wireless communication signal based on nonlinear analysis

In order to effectively reduce the risk of high-risk intrusion signals to wireless communication networks, an automatic anomaly identification of wireless communication signals based on nonlinear analysis is proposed. The fuzzy phase frequency feature of denoised abnormal signal is extracted and the wireless communication signal parameters are estimated by optimizing the fuzzy network abnormal signal feature extraction method of particle filter. The wireless communication signal is collected based on nonlinear analysis, the wireless communication signal network model is constructed, and the Back Propagation Neural Network—Dempster Shafer method is used to classify and fuse the results of feature extraction to realize the automatic identification of wireless communication signal anomalies. The experimental results show that the signal initial frequency estimation and signal slope estimation obtained by the research method have high accuracy, strong coverage performance, large total amount of automatic identification abnormal signals and high compatibility level.

Research on Weak Signal Feature Extraction Method of Rolling Bearing Based on Refined Composite Multi-Scale Weighted Entropy

Rolling bearing health status monitoring is essential for identifying early failures and avoiding un-planned downtime in industrial systems. To overcome the problems of existing entropy methods with multiple faults that are easily confounded at different scales, a weak signal feature extraction method based on refined composite multi-scale weighted entropy is proposed in this paper. The time–frequency domain features are constituted into a multi-dimensional original fault feature set, and the feature sensitivity is evaluated in terms of four feature evaluation criteria, in order to filter out a sensitive feature subset. Three types of refined composite multi-scale entropy are combined with sensitive feature parameters, in a weighted manner, through the use of the Hadamard product operation. The effects of different combinations of feature parameters on the refined composite multi-scale entropy are analyzed through experimental validation. According to the analysis of the experimental data from two test stations, the fault recognition rate reached 100% and 92.22%, respectively, based on the RCMWE method, starting from the first features. The results indicate that the proposed method can identify bearing fault types under different damage states at any scale, with the fault recognition rate being more stable than that of other methods. The proposed method can effectively distinguish rolling bearing health and fault states, providing higher classification accuracy for rolling bearing fault types and fault damage degrees. This puts forward a new idea for rolling bearing health state assessment, which has high engineering application value.