Noise Components Research Articles

Optimized deep neural network models for blood pressure classification using Fourier analysis-based time-frequency spectrogram of photoplethysmography signal.

Appropriate blood pressure (BP) management through continuous monitoring and rapid diagnosis helps to take preventive care against cardiovascular diseases (CVD). As hypertension is one of the leading causes of CVDs, keeping hypertension under control by a timely screening of subjects becomes lifesaving. This work proposes estimating BP from motion artifact-affected photoplethysmography signals (PPG) by applying signal processing techniques in realtime. This paper proposes a deep neural network-based methodology to accurately classify PPG signals using a Fourier theory-based time-frequency (TF) spectrogram. This work uses the Fourier decomposition method (FDM) to transform a PPG signal into a TF spectrogram. In the proposed work, the last three layers of the pre-trained deep neural network, namely, GoogleNet, DenseNet, and AlexNet, are modified and then used to classify the PPG signal into normotension, pre-hypertension, and hypertension. The proposed framework is trained and tested using the MIMIC-III and PPG-BP databases using five-fold training and testing. Out of the three deep neural networks, the proposed framework with the DenseNet-201 network performs best, with a test accuracy of 96.5%. The proposed work uses FDM to compute the TF spectrogram to accurately separate the motion artifacts and noise components from a noise-corrupted PPG signal. Capturing more frequency components that contain more information from PPG signals makes the deep neural networks extract better and more meaningful features. Thus, training a deep neural network model with clean PPG signal features improves the generalized capability of a BP classification model when tested in realtime.

Weighted knowledge distillation of attention-LRCN for recognizing affective states from PPG signals

The recognition of affective states is important for regulating stress levels and maintaining mental health, and it is known that affective states can be inferred from physiological signals. However, in practice, the problem of affect recognition contains many challenges due to various types of external noise and different individual characteristics. This paper proposes a deep learning model called Attention-LRCN for recognizing affective states from photoplethysmography (PPG) signals. We construct a long-term recurrent convolutional network to extract temporal features from spectrograms, and a novel attention module is introduced to alleviate the effect of noise components in PPG signals. Moreover, to improve the recognition accuracy, we propose a weighted knowledge distillation technique, which is a teacher–student learning framework. We quantify the uncertainty of teacher’s predictions, and the predictive uncertainty is utilized to adaptively compute the weight of the distillation loss. To demonstrate the effectiveness of the proposed method, experiments were conducted on the WESAD dataset, which is a public dataset for stress and affect detection. We also collected our own dataset from 34 subjects to verify the accuracy of the proposed method. Experimental results demonstrate that the proposed method significantly outperforms previous algorithms on both the public and real-world datasets. The code is available at https://github.com/ziiho08/Attention-LRCN.

IT Diagnostics of Parkinson’s Disease Based on the Analysis of Voice Markers and Machine Learning

The results of studying the parameters of the spectra of speech signals by machine learning with the use of neural networks are presented. This study was carried out in order to confirm experimentally the possibility of performing an assessment of these parameters for the detection of Parkinson’s disease in the early stages (IT diagnostics). During the study, the public database was used, which systematized the spectra of vowel sounds uttered by patients with Parkinson’s disease. The applied method is binary data classification. In the course of the study, the speech data spectrum was first preprocessed, which consisted of filtering it in order to remove its noise components and eliminate bursts and gaps in it. Then the parameters of the processed spectrum of speech data were determined: average value, maximum and minimum, peak, wavelet coefficients, MFCC and TQWT. After that, the object was selected using the PCA algorithm. The model was trained using the Knn and Random Forest algorithms, as well as the Bayesian neural network. The Bayesian optimization algorithm and the GridSearch method were used to find the best model hyperparameters. It has been established that when using Knn, Random Forest and Bayesian neural network, it is possible to increase the accuracy of recognition of Parkinson’s disease by 94.7; 88.16 and 74.74 %, respectively. A similar study by other scientists showed that the recognition accuracy of data sets was only 86 %.

Bearing fault feature extraction method: improved weighted envelope spectrum

Many existing cyclic spectrum analysis methods are ineffective in solving bearing fault diagnosis in multi-information frequency bands. Based on this, an improved weighted envelope spectrum (IWES) method is proposed. Firstly, IWES uses the fault information intensity index to quantify bearing fault information and evaluate bearing fault information of spectral frequency bands in spectral coherence. Secondly, the threshold function is constructed to determine the threshold adaptively, so as to identify the spectrum frequency components with rich fault information in the spectral coherence. Meanwhile, a weight function is designed based on threshold function to eliminate the interference noise components and keep the fault information. Finally, the spectral coherence and weight function are used to generate IWES with multi-band information. The bearing experiments show that the IWES has excellent noise robustness and can accurately extract the bearing fault characteristic frequency.

Carrier Phase Residual Modeling and Fault Monitoring Using Short-Baseline Double Difference and Machine Learning

Global Navigation Satellite Systems (GNSS) are used to provide accurate position, navigation, and time (PNT) information to users in various sectors of our society including transportation. Augmentation systems such as differential GNSS (DGNSS), real-time kinematics (RTK), and Precise Point Positioning (PPP) improve the GNSS performance, and providing reliable measurements from its reference station is very crucial. To ensure safe and accurate PNT solutions, code and carrier measurements must be monitored for potential faults or a performance degrade. Although there exist numerous methods to model and monitor the measurements, research on the carrier phase measurements is not as extensive as the code measurements. This paper introduces a split of residuals into receiver noise and multipath components to customize their estimation according to their respective statistical properties. This study also proposes a method to use machine learning-based non-linear regression to effectively model and monitor potential faults in the GNSS measurements including the carrier phase. A training dataset is used to model the nominal quantities of GNSS measurement residuals, and inflation factors are applied to over-bound the fault-free residuals. These inflated residuals are coupled with uncertainty factors to compute thresholds for monitoring carrier phase residuals, and the effectiveness of the thresholds is validated with a test dataset by achieving the false alarm rate of 6.61×10−6, slightly lower than the desired level of 10−5.

Computational Aeroacoustics Study of Propellers with Vibrational Motion

In this work, we conduct a numerical investigation of aerodynamics and aeroacoustics of propeller blades with vibrations, which could occur in practical unmanned aerial vehicles as the blades are often made of lightweight materials. The simulations resolve the sound generation from the unsteady turbulent flows using an acoustic-wave preserved artificial compressibility method. Then, the sound projection to the far-field observers is made using an integral solution of the Ffowcs-Williams and Hawkings equation. The study shows that periodic blade vibrations with small amplitudes can lead to aerodynamic thrust fluctuations. The blade vibration also affects the generation of tip vortices and the near-blade flow structures due to the periodic change of the effective angle of attack. Consequently, significant tonal noise at the harmonics of rotational frequency is produced, and the noise can propagate to both upstream and downstream directions of the rotor disc plane. A noise source analysis is performed to identify the contribution of different noise components. Results show that the extra tonal noise is mainly caused by the Doppler effect due to the blade axial motion and the influence of the thrust fluctuations. Moreover, the study also suggests that the high-frequency broadband noise seems to be insensitive to the blade vibration.

Robust Expectation Propagation Detector in Impulsive Noise Channel

In this work, we propose a robust expectation propagation (REP) detector for a system contaminated with impulsive noise. The core idea of the REP is to apply a disperse transformation to the received signal and an associated channel sparsification operation. The proposed disperse transformation distributes the impulsive noise component to multiple physical resource elements. The transformed impulsive noise exhibits greater Gaussianity and better matches the Gaussian assumption of the EP algorithm. The associated channel sparsification can alleviate the adverse effects of deleterious message propagation caused by impulsive noise during the iterative process. Simulation results demonstrate that the proposed REP detector exhibits greater resilience to impulsive noise than the classical EP detector, especially for systems with a high probability of impulsive noise occurrence and/or high-order modulation.

FORECASTING THE NUMBER OF PASSENGER AT JENDERAL AHMAD YANI SEMARANG INTERNATIONAL AIRPORT USING HYBRID SINGULAR SPECTRUM ANALYSIS-NEURAL NETWORK (SSA-NN) METHOD

Transportation was an important sector of supporting the economic growth of a country. The impact of the Covid-19 2020 pandemic at Achmad Yani International Airport in Semarang resulted in the movement of the number of passengers decreasing quite drastically, but in mid-2020 the movement of the number of passengers had slowly increased. Forecasting was done to determine the flow of movement of the number of passengers in the future using the Hybrid Singular Spectrum Analysis (SSA)-Neural Network (NN) method. The SSA method was expected to be able to decompose various patterns in the data into trend, seasonality and noise. Furthermore, the NN method was used to analyze nonlinear patterns in the data. The results showed that the best method was a combination of the SSA method with a window length of 40 and the NN method with a 6-8-1 network architecture (6 input neurons, 8 hidden neurons and 1 output neuron) for the trend component, 11-15-1 (11 neurons input, 15 hidden neurons and 1 output neuron) for the seasonal component, and 10-15-1 (10 input neurons, 15 hidden neurons and 1 output neuron) for the noise component. The model produces a prediction error based on a MAPE value of 0.54% or an accuracy rate of 99.46%.

EEG Analysis for Images Encode and Retrieval under White Noise Stimulation

Utilization of an EEG machine aids researchers in detecting the brain responses and states during provided stimulation. The EEG comes with multiple electrode channels for recording the electrical activity of the brain. This work aims to discover the brain responses under stimulation of white noise during images encoding and retrieving using the EEG machine. The motivation of this research is because limited previous works determine the association between memory encoding and retrieving process based on EEG analysis. Twenty college students participated in this experimental work. They were required to remember the images in two different conditions: control/silent and exposure to white noise with two task difficulties: easy and difficult phases. The EEG was recorded during the remembering and recalling period for data acquisition. The obtained raw EEG dataset was imported to .txt format for further processing. The Butterworth bandpass filter with a frequency range of 4 to 45 Hz was applied to filter the low and high noise components in the EEG signal. Next, the mean of de-noise EEG voltage was obtained to select the most affected channels. It was found that the Fp1, Fp2, F7, and F3 channels achieved the highest mean voltage that higher than 5 μV. Therefore, these channels were selected for extraction of relative rhythm power to investigate the effect of white noise on memory performance. The findings showed that the alpha, gamma, beta, and theta activities were highly found during the encoding and retrieving process when participants were exposed to white noise. These brain rhythms were related to participants' attentional level, information processing and calm state during the assessment. The behavioral data showed that the participants successfully remembered the images more when exposed to white noise compared to the control condition with percentage differences of 21% and 33% for easy and difficult phases due to the stochastic resonance effect and activation of brain rhythms.

Prescribed-Time Mean-Nonovershooting Control under Finite-Time Vanishing Noise

We develop a prescribed-time mean-nonovershooting stabilizing feedback law for stochastic nonlinear systems with noise that vanishes in finite time. The prescribed time of stabilization must be strictly after the noise vanishes, but it may occur as early as the user desires after the noise vanishes. In fact, more generally, the feedback is stabilizing and mean-overshoot free for any noise that, given the prescribed time of stabilization, satisfies certain decay rate properties which require, in particular, that no noise component vanish slower than linearly in the “time to go” until the prescribed time. In contrast to the existing stochastic prescribed-time designs where only multiplicative noise is allowed, our design can deal with multiplicative and additive noise simultaneously. A new controller is designed to guarantee that the mean of the system output prescribed time tracks a given trajectory without overshooting, that the fourth moment of the tracking error between states and derivatives of the reference trajectory converges to zero in prescribed time, and that the controller and all of the states are mean-square bounded. Finally, a simulation example is given to illustrate the prescribed-time mean-nonovershooting design.

Vehicular visible light communications noise analysis and modeling.

Vehicular visible light communications (VVLC) is promising intelligent transportation systems technology with the utilization of light-emitting diodes. The main degrading factor for the performance of VVLC systems is noise. Traditional VVLC systems noise modeling is based on the additive white Gaussian noise assumption in the form of shot and thermal noise. In this paper, to investigate both time correlated and white noise components of the VVLC channel noise, we propose a noise analysis based on Allan variance, which provides a time-series analysis method to identify noise from the data. The results show that white noise and random walk are observed in the VVLC systems. We also propose a motion detection algorithm based on the adaptive Gaussian mixture (GM) model to generate a double Gaussian model of VVLC channel noise. We further present a study on the error performance of a VVLC system considering channel noise to be a mixture of Gaussian components. We derive the analytical expressions of probability of error for binary phase-shift keying and quadrature phase-shift keying constellations. It has been observed that, in the presence of GM noise, the system performance degrades significantly from the usual one expected in a Gaussian noise environment and becomes a function of the mixing coefficients of the GM distribution.

Sparse Bayesian Learning with hierarchical priors for duct mode identification of tonal noise

Fan noise has become an important noise component of civil aviation engines with the increasing bypass ratio. The mode identification method is the key to in-duct fan noise testing. The number of modes that can propagate through the duct increases dramatically for high-frequency in-duct fan noise, which requires a large number of microphones to measure the sound field in the duct. Since the number of modes is usually larger than the number of microphones, it is difficult to realize mode identification. A Sparse Bayesian Learning algorithm for mode identification is proposed to solve this problem in this paper. The in-duct sound field is described by a statistical model, and the inverse problem of mode identification is expressed in a Bayesian framework. The non-convex optimization problem of the cost function is considered, iteratively achieving the sparsest solution via a block coordinate descent algorithm in a majorization-minimization framework. The mode identification method based on Sparse Bayesian Learning has high sparsity and can achieve parameter adaptation. The effectiveness of the Sparse Bayesian Learning method for duct mode identification is verified by numerical simulations and experimental tests. The results show that the mode identification method of Sparse Bayesian Learning can accurately identify the target mode using fewer microphones.

A denoising method for ultrasonic testing of rubber composites based on improved symplectic geometric mode decomposition

Ultrasonic nondestructive testing is widely used in industry due to its wide detection range and high detection efficiency. However, for multi-layer stacked rubber materials, ultrasound is easily affected by noise interference, it is a challenge work to reduce or separate the noise via signal processing. Therefore, a new ultrasonic echo denoising method based on symplectic geometry mode decomposition is proposed. It can fully decompose the useful components and noise components in the signal, and has a good screening effect on the useful components and noise components. In order to verify the effectiveness and robustness of the method, we conduct experimental research on the model with simulated and measured signals. The analysis results show that our method outperforms other noise reduction methods.

Minimum entropy deconvolution enhanced by KLOF and phase editing for fault diagnosis of rotating machinery

Repetitive impulses are critical characteristics in vibration signals of fault rotating machinery and minimum entropy deconvolution (MED) is an effective tool to extract these characteristics for fault diagnosis. However, early fault-induced impulses are usually weak and submerged in heavy background noise when rotating machinery operate in severe environment. Under this circumstance, MED generally fails to isolate useful impulses from contaminated signals. To overcome the shortcoming, this paper proposed a kernel local outlier factor MED enhanced by phase editing method (KPMED) for fault diagnosis of rotating machinery. First, non-Gaussian noise components in fault signals are removed by kernel-based local outlier factor (KLOF) in data cleaning way. Second, the residual Gaussian noise components are eliminated by phase editing (PE) in complex domain with phase perspective. Finally, the fault-induced repetitive periodic impulses are extracted by selecting MED filtered signal with large value of kurtosis for diagnosis. The effectiveness of KPMED is validated by a simulated bearing fault signal and two vibration signals collected from a pinion of a wind turbine and a rolling element bearing, respectively. Compared with original MED and its corresponding variants, KPMED is more accurate and suitable for fault diagnosis of rotating machinery under severe working conditions.

An adaptive autoregressive pre-whitener for speech and acoustic signals based on parametric NMF

A common assumption in many speech and acoustic processing methods is that the noise is white and Gaussian (WGN). Although making this assumption results in simple and computationally attractive methods, the assumption is often too simple and crude in many applications. In this paper, we introduce a general purpose and online pre-whitener which can be used as a pre-processor with methods based on the WGN assumption, improving their reliability and performance in applications with colored noise. The pre-whitener is a time-varying filter whose coefficients are found using a parametric non-negative matrix factorization (NMF), based on autoregressive (AR) mixture modeling of both the noise component and the signal component constituting the noisy signal. Compared to other types of pre-whiteners, we show that the proposed pre-whitener has the best performance, especially in applications with non-stationary noise. We also perform a large number of experiments to quantify the benefits of using a pre-whitener as a pre-processor for methods based on the WGN-assumption. The applications of interest were pitch estimation and time-of-arrival (TOA) estimation, where the WGN assumption is very popular.

A time-varying instantaneous frequency fault features extraction method of rolling bearing under variable speed

Under variable speed, the time-frequency analysis (TFA) method for processing rolling bearing vibration signal is susceptible to noise, which cannot extract accurate time-frequency features. Plus, the special meaning of each time-frequency feature remains unclear. Based on the dynamic characteristics of rolling bearing and combined with the improved multisynchrosqueezing transform (IMSST), empirical Fourier decomposition (EFD), and generalized demodulation (GD), a time-varying instantaneous frequency fault features extraction method of rolling bearing under variable speed is proposed in this paper. Firstly, the instantaneous frequency (IF) trajectory of the rotating shaft is accurately represented by the better time-frequency resolution of IMSST, and the demodulation phase function for GD is constructed by combining the fault characteristic coefficient (FCC) of the rolling bearing. Then, the components in the signal that match the demodulation phase function are mapped into constant IF components. Through adaptive EFD and inverse demodulation, the time-varying IF components satisfying the conditions are separated. Finally, based on the Hilbert transform (HT) and adaptive time-frequency spectrum (ATFS), a high-quality time-frequency representation (TFR) is reconstructed to extract the time-varying IF fault features and identify the defect type of rolling bearing. The simulation and measured signal analysis results show that the proposed method can accurately extract fault features of rolling bearings with different defects under variable speed and identify fault types, thus maximally avoiding the influence of noise and interference components.

Timing analysis of Swift J0243.6+6124 with NICER and Fermi/GBM during the decay phase of the 2017–2018 outburst

ABSTRACT We present a timing and noise analysis of the Be/X-ray binary system Swift J0243.6+6124 during its 2017–2018 super-Eddington outburst using NICER/XTI observations. We apply a synthetic pulse timing analysis to enrich the Fermi/GBM spin frequency history of the source with the new measurements from NICER/XTI. We show that the pulse profiles switch from double-peaked to single-peaked when the X-ray luminosity drops below ∼7 × 1036 erg s−1. We suggest that this transitional luminosity is associated with the transition from a pencil beam pattern to a hybrid beam pattern when the Coulomb interactions become ineffective to decelerate the accretion flow, which implies a dipolar magnetic field strength of ∼5 × 1012 G. We also obtained the power density spectra (PDS) of the spin frequency derivative fluctuations. The red noise component of the PDS is found to be steeper (ω−3.36) than the other transient accreting sources. We find significantly high noise strength estimates above the super-Eddington luminosity levels, which may arise from the torque fluctuations due to interactions with the quadrupole fields at such levels.

Signal-to-Noise Ratio in Heat-Assisted-Recording Media: A Comparison between Simulations and Experiments

We develop a code to extract the signal-to-noise ratio (SNR) arising from the magnetic film in a recording medium. The approach allows us to separate the remanence and transition contributions from the global spatial noise. The results are in excellent agreement with the analysis performed on the same data sets by means of Seagate proprietary software based on ensemble wave-form analysis. We then apply this analytical approach to the results of heat-assisted magnetic recording (HAMR) dynamics simulations by means of the open-source multi-time-scale micromagnetic code mars and compare these with experimental spin-stand measurements of analogous systems. The proposed model could be used as the standard tool to understand the underlying physics of the noise components affecting HAMR operations and how to decrease the noise arising from the medium to improve the writing performance of HAMR.

Sensorless Control of SPMSM Based on High-Frequency Positive- and Negative-Sequence Current Dual-Demodulation

In surface-mounted permanent magnet synchronous motor (SPMSM) low-speed sensorless control, the high-frequency sinusoidal voltage injection method for rotor position estimation is affected by the low saturation salient rate of SPMSM and the system delay, resulting in estimation error. This paper proposes a dual-demodulation method utilizing the high-frequency positive- and negative-sequence stator current of the motor. This method is based on the principle of maximizing the effect of the salient inductance component and minimizing the significant component of noise, which improves the utilization of the current containing angle information and increases the signal-to-noise ratio of the position angle error fed to the rotor position observer. The DC bias of the current caused by system delay is also eliminated so that the accuracy of position estimation is improved. The proposed method is verified by experiment, and the results show that the accuracy of the estimated rotor position and dynamic performance of the estimator are greatly improved compared to existing methods.

A hybrid method of frequency-weighted energy operator and power spectrum fusion to detect bearing faults

The difficulties in early fault diagnosis of bearings mainly include two aspects: first, the initial damage size of the bearing is small, and the abnormal vibration caused by slight damage to the bearing is very weak. Second, vibration signals collected in actual industrial environments always contain strong noise interference. Therefore, traditional diagnostic procedures are not satisfactory. To address these challenges, this work provides a hybrid model combining frequency-weighted energy operator (FWEO) with power spectrum fusion (PSF) to identify weak fault features of bearings and detect different fault types. Different from traditional time-domain signal filtering, PSF is first used to reduce the interference of noise components in the power spectrum, which will not weaken the fault signal components during denoising. Second, the filtered signal is transformed into the time domain and FWEO is employed to further enhance the cyclic fault signal caused by the weak defect of the bearing. Finally, the existence of a fault is identified by observing the squared envelope spectrum of the signal. The effectiveness of the proposed hybrid model is demonstrated through two simulated fault signals and three different experimental fault signals. The results show that the proposed model has high anti-noise performance and robustness and can extract the fault frequency well.