Signals In Frequency Domain Research Articles

A unified deep learning framework for water quality prediction based on time-frequency feature extraction and data feature enhancement

Deep learning methods exhibited significant advantages in mapping highly nonlinear relationships with acceptable computational speed, and have been widely used to predict water quality. However, various model selection and construction methods resulted in differences in prediction accuracy and performance. Hence, a unified deep learning framework for water quality prediction was established in the paper, including data processing module, feature enhancement module, and data prediction module. In the established model, the data processing module based on wavelet transform method was applied to decomposing complex nonlinear meteorology, hydrology, and water quality data into multiple frequency domain signals for extracting self characteristics of data cyclic and fluctuations. The feature enhancement module based on Informer Encoder was used to enhance feature encoding of time series data in different frequency domains to discover global time dependent features of variables. Finally, the data prediction module based on the stacked bidirectional long and short term memory network (SBiLSTM) method was employed to strengthen the local correlation of feature sequences and predict the water quality. The established model framework was applied in Lijiang River in Guilin, China. The maximum relative errors between the predicted and observed values for dissolved oxygen (DO), chemical oxygen demand (CODMn) were 12.4% and 20.7%, suggesting a satisfactory prediction performance of the established model. The validation results showed that the established model was superior to all other models in terms of prediction accuracy with RMSE values 0.329, 0.121, MAE values 0.217, 0.057, SMAPE values 0.022, 0.063 for DO and CODMn, respectively. Ablation tests confirmed the necessity and rationality of each module for the established model framework. The established method provided a unified deep learning framework for water quality prediction.

Robust prediction of remaining useful lifetime of bearings using deep learning

Predicting the remaining useful lifetime (RUL) of bearings in electric motors is crucial to reduce repair costs in industrial maintenance. With the technological advances of Industry 4.0, physical models for prognostics and RUL prediction have been replaced by data-driven models that require no expert feature extraction. Instead, the model itself learns which features are important. However, these models are normally trained and tested on the same dataset, i.e. under the same operating conditions. This limits the application of a model to other operating conditions unless the model is fine-tuned with data corresponding to those conditions. This paper proposes a novel robust health prognostics technique that detects inner-race bearing failures and predicts the RUL of electric motor bearings under various motor conditions without model retraining or fine-tuning. The model combines time and frequency-domain vibration signal analyses to extract features, a stacked variational denoising autoencoder (SVDAE) to fuse these features and build a Health Indicator and a bidirectional long short-term memory (BiLSTM) neural network to predict the remaining useful lifetime of the bearings. The proposed model is trained with a dataset, validated with another dataset and finally tested with seven additional datasets corresponding to vibrations gathered from different motors and operating conditions. The results are more robust and accurate than those of the literature, the robustness of the prediction with different motor and operating conditions is proven, and there is no need to retrain or fine-tune the model, making the proposed model suitable for recently installed equipment.

The Early Diagnosis of Rolling Bearings’ Faults Using Fractional Fourier Transform Information Fusion and a Lightweight Neural Network

In response to challenges associated with feature extraction and diagnostic models’ complexity in the early diagnosis of bearings’ faults, this paper presents an innovative approach for the early fault diagnosis of rolling bearings. This method combined concepts from frequency domain signal analysis with lightweight neural networks. To begin, vibration signals from rolling bearings were collected using vibration sensors, and the mean square value was utilized as an indicator for accurate early fault signal extraction. Subsequently, employing the fractional Fourier transform, the time domain signal was converted into a frequency domain signal, which provided more detailed frequency feature information. The fusion process combined amplitude frequency and phase frequency information, and was visualized as a Gram angle field map. The lightweight neural network Xception was selected as the primary fault diagnosis tool. Xception, a convolutional neural network (CNN) variant, was chosen for its lightweight design, which maintains excellent performance while significantly reducing model parameters. The experimental results demonstrated that the Xception model excelled in rolling bearing fault diagnosis, particularly when utilizing fused information datasets. This outcome underscores the advantages of combining information fusion and the Xception model to enhance the accuracy of early rolling bearing fault diagnosis, and offers a viable solution for health monitoring and fault diagnosis in industrial settings.

Using Computer Technique for Developing Method for Vibration Damage Estimation Under Combined Random and Deterministic Loading

Abstract This paper is focusses on developing a novel method for vibration damage estimation for military helicopters, fighter aircrafts and any other aircraft exposed to combined stochastic and deterministic loading. The first stage of the research focused on frequency domain damage prediction, which is the legacy method proposed by Bishop and developed by Sweitzer, Schlesinger, Woodward, Kerr, Murthy, Datta and, Atkins. The mentioned frequency domain-based method is used in commercial software, e.g., MSC CAE Fatigue. Frequency domain damage prediction is based on superposition of spectral moments and Dirlik method of Rainflow Cycle Counting algorithm in frequency domain. The first phase of the research showed the legacy algorithm based on transfer function developed using FEM (Finite Element Method) method in Abaqus environment and is very conservative. The second stage of the research aims to develop a novel method which allowing for more robust and accurate damage estimation. For this purpose, the Monte Carlo method for retrieving random signal in the time domain from signal in frequency domain was used. To obtain the system transfer function, – the 1 g load harmonic system response was obtained using FEM analysis. It was subsequently scaled linearly by the PSD input curve for random loading and sine wave, or sine sweep function for deterministic loading to calculate the cumulative system response of the linear system. The research allows the development of a novel method to precisely estimate vibration damage using combined time and frequency domains approach, based on effective frequency domain FEM analysis of the linear system. The new proposed method can be also used for precise replication of test conditions via considering signal clipping and frequency resolution used for real testing.

Study on the Three-component Response Characteristics of Ground-airborne Electromagnetic Field

At any point in space, an electromagnetic field consists of three components: Hx, Hy, and Hz. The ground airborne electromagnetic method (GAEM) does not usually retrieve the Hx and Hy components of the electromagnetic field, less data may result in less accurate goaf interpretations. In this study, we numerically simulate the characteristics of the frequency-domain ground-airborne electromagnetic signal sources, quantify the induction signal distribution laws in the x-, y-, and z- directions, and identify the various factors that influence ground-airborne electromagnetic signals. The results of this study provide theoretical support for the future development of a three-component receiver for ground airborne electromagnetic field surveys. Furthermore, this data also provides information that will prove valuable in the design of field work construction.

A Fine-Grained Approach for EEG-Based Emotion Recognition Using Clustering and Hybrid Deep Neural Networks

Emotion recognition, as an important part of human-computer interaction, is of great research significance and has already played a role in the fields of artificial intelligence, healthcare, and distance education. In recent times, there has been a growing trend in using deep learning techniques for EEG emotion recognition. These methods have shown higher accuracy in recognizing emotions when compared with traditional machine learning methods. However, most of the current EEG emotion recognition performs multi-category single-label prediction, and is a binary classification problem based on the dimensional model. This simplifies the fact that human emotions are mixed and complex. In order to adapt to real-world applications, fine-grained emotion recognition is necessary. We propose a new method for building emotion classification labels using linguistic resource and density-based spatial clustering of applications with noise (DBSCAN). Additionally, we integrate the frequency domain and spatial features of emotional EEG signals and feed these features into a serial network that combines a convolutional neural network (CNN) and a long short-term memory (LSTM) recurrent neural network (RNN) for EEG emotion feature learning and classification. We conduct emotion classification experiments on the DEAP dataset, and the results show that our method has an average emotion classification accuracy of 92.98% per subject, validating the effectiveness of the improvements we have made to our emotion classification method. Our method for emotion classification holds potential for future use in the domain of affective computing, such as mental health care, education, social media, and so on. By constructing an automatic emotion analysis system using our method to enable the machine to understand the emotional implications conveyed by the subjects’ EEG signals, it can provide healthcare professionals with valuable information for effective treatment outcomes.

Localization of mechanical and electrical defects in dry-type transformers using an optimized acoustic imaging approach.

This paper presents an acoustic imaging localization system designed to pinpoint common defects in dry-type transformers by analyzing the unique sounds they produce during operation. The system includes an optimized microphone array and an improved multiple signal classification algorithm. Sound signal characteristics of typical defects, such as foreign object intrusion, screw loosening, and partial discharge, are investigated. A 64-element, 8-arm spiral microphone array is designed using a particle swarm optimization algorithm. The multiple signal classification algorithm enhances acoustic imaging quality in field environments by transforming the input from time-domain to preprocessed frequency-domain signals. The power spectra of subarray and main array are combined, forming the optimization algorithm's output. Experimental results demonstrate the system's effectiveness and accuracy.

Remaining useful life prediction of rolling bearings based on TCN-MSA

As a pivotal element within the drive system of mechanical equipment, the remaining useful life (RUL) of rolling bearings not only dictates the lifespan of the equipment’s drive system but also the overall machine. An inaccurate prediction of the RUL of rolling bearings could hinder the formulation of maintenance strategies and lead to a chain of failures stemming from bearing malfunction, culminating in potentially catastrophic accidents. This paper designs a novel temporal convolutional network-multi-head self-attention (TCN-MSA) model for predicting the RUL of rolling bearings. This model considers the intricate non-linearity and complexity of mechanical equipment systems. It captures long-term dependencies using the causally inflated convolutional structure within the temporal convolutional network (TCN) and simultaneously extracts features from the frequency domain signal. Subsequently, by employing the multi-head self-attention (MSA) mechanism, the model discerns the significance of different features throughout the degradation process of rolling bearings by analyzing global information. The final prediction for rolling bearings’ RUL has been successfully attained. To underline the excellence of the method presented in this paper, a comparative analysis was performed with existing methods, such as convolutional neural network, gate recurrent unit, and TCN. The results highlight that the model designed in this paper surpasses other existing methods in predicting the RUL of rolling bearings, demonstrating superior prediction accuracy and robust generalization capability.

Patient Confidential Data Hiding and Transmission System Using Amplitude Quantization in the Frequency Domain of ECG Signals.

The transform domain provides a useful tool in the field of confidential data hiding and protection. In order to protect and transmit patients' information and competence, this study develops an amplitude quantization system in a transform domain by hiding patients' information in an electrocardiogram (ECG). In this system, we first consider a non-linear model with a hiding state switch to enhance the quality of the hidden ECG signals. Next, we utilize particle swarm optimization (PSO) to solve the non-linear model so as to have a good signal-to-noise ratio (SNR), root mean square error (RMSE), and relative root mean square error (rRMSE). Accordingly, the distortion of the shape in each ECG signal is tiny, while the hidden information can fulfill the needs of physiological diagnostics. The extraction of hidden information is reversely similar to a hiding procedure without primary ECG signals. Preliminary outcomes confirm the effectiveness of our proposed method, especially an Amplitude Similarity of almost 1, an Interval RMSE of almost 0, and SNRs all above 30.

CNSC-17. WHOLE-BRAIN SPECTRAL CHANGE IN GLIOBLASTOMA PATIENTS: ASSOCIATIONS WITH CLINICAL COMORBIDITIES AND OVERALL SURVIVAL

Abstract Glioblastoma, a highly aggressive form of brain tumor, is a brain-wide disease. We evaluated the impact of tumor burden on whole brain resting-state functional magnetic resonance imaging (rs-fMRI) activity. Specifically, we analyzed rs-fMRI signals in the temporal frequency domain in terms of the power-law exponent and fractional amplitude of low-frequency fluctuations (fALFF). We contrasted 189 patients with newly-diagnosed glioblastoma versus 189 age-matched reference participants. The patient and reference datasets were matched for age and head motion. The principal finding was markedly flatter spectra and reduced grey matter fALFF in the patients as compared to the reference dataset. We posit that the whole-brain spectral change is attributable to global dysregulation of excitatory and inhibitory balance and metabolic demand in the tumor-bearing brain. Additionally, we observed that clinical comorbidities, in particular, seizures, and MGMT promoter methylation, were associated with flatter spectra. Notably, the degree of change in spectra was predictive of overall survival. Our findings suggest that frequency domain analysis of rs-fMRI activity provides prognostic information in glioblastoma patients and offers a means of noninvasively studying the effects of glioblastoma on the whole brain.

Automatic Classification of Spectra with IEF-SCNN

Telescopes such as the Large Sky Area Multi-Object Spectroscopic Telescope and the Sloan Digital Sky Survey have produced an extensive collection of spectra, challenging the feasibility of manual classification in terms of accuracy and efficiency. To overcome these limitations, machine learning techniques are increasingly being utilized for automated spectral classification. However, these approaches primarily treat spectra as frequency domain signals, and lack robustness in low signal-to-noise ratio (S/N) scenarios and for small datasets of rare celestial objects. Moreover, they frequently neglect nuanced expert astronomical understanding. In this study, we draw inspiration from the human spectral discrimination process and propose a new model called the Image-EFficientNetV2-Spectrum Convolutional Neural Network (IEF-SCNN). IEF-SCNN combines spectral images using EfficientNetV2 with one-dimensional (1D) spectra through a 1DCNN. This integration effectively incorporates astronomical expertise into the classification process. Specifically, we plot the spectrum as an image and then classify it in a way that incorporates an attention mechanism. This attention mechanism mimics human observation of images for classification, selectively emphasizing relevant information while ignoring irrelevant details. Experimental data demonstrate that IEF-SCNN outperforms existing models in terms of the F1-score and accuracy metrics, particularly for low S/N (<6) data. Using progressive learning and an attention mechanism, the model trained on 12,000 M-class stars with an S/N below 6 achieved an accuracy of 87.38% on a 4000-sample test set. This surpasses traditional models (support vector machine with 83.15% accuracy, random forest with 65.40%, and artificial neural network with 84.40%) and the 1D stellar spectral CNN (85.65% accuracy). This research offers a foundation for the development of innovative methods for the automated identification of specific celestial objects, and can promote the creation of user-friendly software for astronomers who may not have computational expertise.

Clutter Rejection for FM Ranging Airborne Radar

In airborne radar, it is important to reject clutter signals to detect targets of interest. High pulse repetition frequency (HPRF) waveform provides robust detection performance in a clutter environment by separating clutter signals from the target echo signals in a Doppler frequency domain, but it suffers from range ambiguity. The range ambiguity in HPRF can be resolved by frequency modulation (FM) ranging. However, the Doppler frequencies of both the clutter and target echo signals change linearly with the range due to changes in the carrier frequency in FM ranging. In such a case, the target echo signal can be rejected by a conventional clutter rejection frequency even if it is not masked by a clutter signal. This paper proposes an optimum clutter rejection frequency for FM ranging airborne radar by considering the Doppler frequency spread in FM ranging. The optimum clutter rejection frequency is derived by calculating the maximum Doppler frequency of the spread clutter signal caused by FM ranging. The simulation and flight test results verify that the proposed clutter rejection frequency is the optimum value that can improve the detection capability of FM ranging airborne radar.

A wide-beam NCS algorithm for multi-receiver SAS based on azimuth spectrum superposition

The existing multi-receiver synthetic aperture sonar (SAS) imaging algorithms are suitable for narrow-beam width, which will lead to a decrease in imaging quality under wide-beam condition and are not in line with the development needs of SAS. We propose a non-linear chirp scaling algorithm (NCSA) for wide beam multi-receiver SAS. Firstly, the point target reference spectrum (PTRS) of each receiver is obtained by the Lagrange inversion theorem (LIT), and then the under-sampled signal in the azimuth frequency domain is obtained through azimuth spectrum extension; Then, considering the cubic term of range frequency in the PTRS and the linear variation of equivalent frequency modulation slope with range, each receiver is imaged using the NCSA, and coherent superposition is performed in the azimuth frequency domain to eliminate spectrum aliasing caused by azimuth spectrum extension; Finally, the azimuth inverse transform is performed on the superimposed signal to obtain the focusing imaging. Computer simulation experiments and field data verify that this method is superior to the existing SAS imaging algorithm, improving the quality of wide-beam imaging, avoiding the interpolation operation of the traditional range-Doppler algorithm, and saving computation cost.

Elevator Fault Diagnosis Method Based on IAO-XGBoost under Unbalanced Samples

Elevators are essential tools in daily life; timely and accurate fault diagnosis plays a crucial role in ensuring their safe operation. However, the existing elevator fault diagnosis methods often neglect the imbalance between the actual collected normal samples and the fault samples, resulting in low diagnostic accuracy. In this study, we propose an improved Aquila optimizer (IAO) extreme gradient boosting tree (XGBoost)-based elevator fault diagnosis method under unbalanced samples. The proposed method includes three main components: multi-domain feature extraction, sample balancing, and fault diagnosis. In the feature extraction phase, the time domain, frequency domain and entropy features of the vibration signal are extracted. In the sample balance phase, aiming at the problem of unbalanced fault samples, after feature selection using recursive feature elimination (RFE), the minority class samples are oversampled by applying SMOTE-Tomek. In the fault diagnosis phase, IAO is used to optimize the hyperparameters in the XGBoost, and the optimized hyperparameters are brought into XGBoost for fault diagnosis. The fault diagnosis accuracy of the method proposed in this study can reach 99.06%, and the method can accurately identify the fault state of the elevator.

Wearable Sleep Monitoring System Based on Machine Learning Using Snoring Sound Signal

Abstract According to the obstructive sleep apnea Syndrome (OSAS), a wearable sleep monitoring system is designed based on machine learning using snoring sound signal. The system picks up snoring signal via bone conduction sensor, and calculates the apnea-hypopnea index (AHI). By analyzing the snoring signal in frequency domain, spectral entropy and other frequency-domain features are selected. Finally, the neural network classifier model is established. In the model, the input variables are eight frequency-domain features, and the output response is related to AHI value. Trained by machine learning, the result shows that the average accuracy in identifying the severity of the four kinds of OSAS categories is 59%. The system uses the measured data of snoring to analyze the symptoms of OSAS, so as to realize the preliminary forecast based on the snoring data. The system proposed in this paper has a good application development prospect in intelligent monitoring and medical instruments.

Frequency analysis of passive sonar signals using multiple-measurement physics-informed neural network

We propose a deep learning based frequency analysis for passive sonar signals by considering a linear system based on a relation between time and frequency domain signal representations. An adaptive learned iterative shrinkage thresholding algorithm (Ada-LISTA) is utilized as the deep learning architecture to solve the linear system effectively and to ensure generalization performance. To obtain more reliable solutions, we involve the linear system in loss function during training. Futhermore, in order to reduce noise effectively, we employ multiple measurements from time and space domains, which have common frequency components of passive sonar signals. Thus, the loss function is modified for solutions from the multiple measurements to have shared frequency components. We conduct experiments using synthetic and underwater in-situ data to examine performance of frequency analysis considering the linear system in designing the architecture and finding the optimal weights. The frequency analysis demonstrates superior performance in detecting frequency components and reducing noise, compared to fast Fourier transform and sparse Bayesian learning with low computational burden during its application to the test measurements.

Closed-form multiple source direction-of-arrival estimator under reverberant environments.

Accurate direction-of-arrival (DOA) estimation of multiple sources, simultaneously active in a reverberant environment, remains a challenge, as the multi-path acoustic reflections and overlapped periods dramatically distort the direct-path wave propagation. This article proposes a prominent solution localizing multiple sources in a reverberant environment using closed-form estimates, circumventing any exhaustive search over the two-dimensional directional space. Apart from a low complexity cost, the algorithm has robustness to reverberant, inactive, and overlapped periods and an ease of operation in practice, achieving sufficient accuracy compared to state-of-the-art approaches. Specifically, this algorithm localizes an unknown number of sources through four steps: (i) decomposing the frequency domain signals on a spherical array to the spherical harmonics domain; (ii) extracting the first-order relative harmonic coefficients as the input features; (iii) achieving direct-path dominance detection and localization using closed-form estimation; and (iv) estimating the number of sources and their DOAs based on those pass the direct-path detection. Experimental results, using extensive simulated and real-life recordings, confirm the algorithm with a significantly reduced computational complexity, while preserving competitive localization accuracy as compared to the baseline approaches. Additional tests confirm this low-complexity algorithm even with a potential capacity for online DOA tracking of multiple moving sources.

Research on charge induction law and application of coal samples with different fissures

Indoor testing are performed to explore the charge induction law during the uniaxial compression fracture process of coal samples, and the charge time and frequency domain signals of coal samples with different primary fissures are analyzed in the paper. On-site monitoring of charge in different fissures distribution areas of underground coal tunnels, and the charge signals of different drillingdepths in coal seams are analyzed. The results show that the uniaxial compressive strength and elastic modulus of multi-fissured coal samples are less than those of less fissured coal samples, and the Poisson’s ratio is greater than those of less fissured coal samples. The charge induction signal intensity during the fracture process of multi-fissured coal samples is relatively low, but it is concentrated at the low frequency of 0–50 Hz in the compacting elasticity stage. The charge signal intensity during the fracture process of coal samples with less fissure is relatively high, and the charge frequency during the reinforcement damage stage is concentrated at a low frequency of 0–50 Hz. Therefore, the sudden appearance of low-frequency charge signals is more suitable as effective precursor information for the instability and failure of less fissured coal bodies. The average charge intensity is small in the multi-fissured area with a drilling depth of 1–4 m in the coal seam, and the average charge intensity of the coal body with less fissures is larger in the 5–12 m region. The on-site charge monitoring results have good consistency with the indoor test results. This study has guiding significance in setting up a charge monitoring warning index of instability failure in different coal body fissures regions.

Dual-Pass Hollow-Core Fiber Gas Spectroscopy Using a Reflective Configuration With Heterodyne-Based Signal Detection

Hollow-core optical fibers have been used for laser gas sensing for almost two decades. However, to date, the vast majority of sensing systems have used hollow-core fiber only as a replacement for bulky gas cells (single- or multi-pass). Here, we investigate a reflective, dual-pass configuration in which light enters and exits hollow-core fiber from the same side. The advantage of this approach is the fact that the hollow-core fiber can be used not only as a cell but also as an endoscope-like probe, which does not need any additional tubing for sample delivery. <p xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">In this paper, we identify unwanted reflections from various fiber-to-fiber connections in the setup as the main challenge limiting the performance of the sensor in reflective configuration. We propose and experimentally demonstrate that the problem of reflections can be addressed using heterodyne-based detection, which allows for identifying and extracting useful signals in the radio frequency domain. Absorption and dispersion spectroscopy of methane are performed as a proof-of-concept with a detection limit of methane concentration below 10 ppm. Discussion insights into further development are provided, including strategies for miniaturization and detection limit improvement.

Layered hybrid PAM-DMT for IM/DD OWC systems.

Traditional pulse-amplitude-modulated discrete multitone modulation (PAM-DMT) suffers from poor overall performance of spectral and power efficiencies in optical wireless communication (OWC) systems. We propose layered hybrid PAM-DMT (LHPAM-DMT) to utilize more subcarriers to improve the performance. The real part of frequency domain signal is divided into several layers and symmetry biases are added in time domain to generate real-valued and nonnegative signals for intensity modulation with direct detection (IM/DD) OWC systems. By appropriately designing the orthogonality between the signals in lower layers and signals & added biases in higher layers, we further propose an iterative receiver to recover the transmitted information. Theoretical derivation proves that LHPAM-DMT has higher spectral efficiency than PAM-DMT and lower complexity than layered PAM-DMT. Numerical results suggest that LHPAM-DMT is more power efficient than PAM-DMT as well as direct-current (DC) biased optical OFDM (DCO-OFDM), one of the most popular schemes.