Data In Short Time Research Articles

Human Gait phases recognition based on multi-source data fusion and BILSTM attention neural network

A human gait recognition algorithm with deep learning based on multi-source data fusion is proposed to assist exoskeleton realizing complex human-exoskeleton cooperative motion. A lightweight gait acquisition device simultaneously acquires three different types of sensor signals, i.e., thigh surface electromyography (sEMG), ground reaction force (GRF) of human feet and two joints motion information. The sample data is obtained by many multi-scene gait experiments involving stand still, walk up/down stairs, and walk up/down slope, etc. The proposed algorithm recognizes multiple gait phases (27 classes) in different motion patterns by using short-time gait data, which has two typical advantages: (1) Based on multi-source sensor distributed in human body, this wearable device is constructed for ease of use to address several Subjects (different weights and heights) with low cost and light weight. (2) The critical features of gait data are identified by Bi-directional Long Short-Term Memory (BILSTM) attention neural network with higher recognition accuracy than other state-of-art recognition methods, especially in arbitrary gait switching durations.

An integrated gear tooth crack analysis of coupled electromechanical model: A complexity-based approach

This paper presents a novel approach by conducting a complexity-based study of an electromechanical (EM) gearbox system with tooth root cracks, utilizing an analytical model and motor current signals for the first time, in contrast to existing studies that primarily use experimental vibration data for complexity-based analyses. The coupled EM model is developed using the modified Lagrangian approach and by incorporating improved time-varying mesh stiffness and gear crack models. Refined composite multiscale dispersion entropy (RCMDE) is employed to evaluate system complexity at various crack depths. The analysis focuses on residual current signals and uncovers an inverse correlation between RCMDE values and the severity of cracks. To demonstrate the efficacy of the proposed approach, a comparison study is conducted between RCMDE and multiscale dispersion entropy (MDE) on both residual current and vibration signals. The results show that RCMDE provides superior stability and effectiveness compared to MDE. Additionally, it has been found that residual current signals offer more robust insights into system complexity than residual vibration signals. This study contributes valuable insights into the behavior of gearbox systems experiencing crack-induced faults, especially concerning short-time series data, highlighting the advantages of using motor current signals and RCMDE for complexity-based analyses.

Extracting winter wheat based on multi-feature optimization of short-time series synthetic aperture radar data with dual polarizations

Extracting winter wheat based on multi-feature optimization of short-time series synthetic aperture radar data with dual polarizations

High-resolution frequency domain decomposition for modal analysis of bridges using train-induced free-vibrations

Modal parameters are structural inherent characteristics that can be applied for revealing performance of railway bridges. Free vibration signals generated by a passage of train are commonly utilized to estimate the modal parameters of railway bridges due to their higher signal-to-noise ratios compared to random vibrations caused by ambient loads. However, since free vibration signals rapidly decay over time, the available free-vibration data is typically short-time. When using the fast Fourier transform-based spectral estimation method for modal identification from short-time vibration data, a phenomenon known as spectral leakage occurs, leading to miss-identification of some structural modes. In this study, the classical frequency domain decomposition (FDD) is improved for modal identification of railway bridges, in which the higher resolution auto-power spectral density (PSD) and cross-PSD functions are calculated through the autoregressive (AR) model-based method. The AR model-based method improves both the smoothness and resolution of the PSD functions compared to the fast Fourier transform technique. These AR model-based PSD functions are then employed in the FDD process to facilitate frequency and mode shape identification while avoiding spurious noise modes. The proposed eigenvalue fitting technique is subsequently utilized to estimate damping ratios. Numerical simulation data as well as vibration data from an actual bridge are analyzed to validate the proposed method, with a comparison made to the Welch’s PSD-based method. The results demonstrate that the modified FDD approach enables more effective identification of structural modes, even in the presence of closely-spaced modes.

Efficient formulation of multitime generalized quantum master equations: Taming the cost of simulating 2D spectra.

Modern 4-wave mixing spectroscopies are expensive to obtain experimentally and computationally. In certain cases, the unfavorable scaling of quantum dynamics problems can be improved using a generalized quantum master equation (GQME) approach. However, the inclusion of multiple (light-matter) interactions complicates the equation of motion and leads to seemingly unavoidable cubic scaling in time. In this paper, we present a formulation that greatly simplifies and reduces the computational cost of previous work that extended the GQME framework to treat arbitrary numbers of quantum measurements. Specifically, we remove the time derivatives of quantum correlation functions from the modified Mori-Nakajima-Zwanzig framework by switching to a discrete-convolution implementation inspired by the transfer tensor approach. We then demonstrate the method's capabilities by simulating 2D electronic spectra for the excitation-energy-transfer dimer model. In our method, the resolution of data can be arbitrarily coarsened, especially along the t2 axis, which mirrors how the data are obtained experimentally. Even in a modest case, this demands O(103) fewer data points. We are further able to decompose the spectra into one-, two-, and three-time correlations, showing how and when the system enters a Markovian regime where further measurements are unnecessary to predict future spectra and the scaling becomes quadratic. This offers the ability to generate long-time spectra using only short-time data, enabling access to timescales previously beyond the reach of standard methodologies.

Rapid health estimation of in-service battery packs based on limited labels and domain adaptation

For large-scale in-service electric vehicles (EVs) that undergo potential maintenance, second-hand transactions, and retirement, it is crucial to rapidly evaluate the health status of their battery packs. However, existing methods often rely on lengthy battery charging/discharging data or extensive training samples, which hinders their implementation in practical scenarios. To address this issue, a rapid health estimation method based on short-time charging data and limited labels for in-service battery packs is proposed in this paper. First, a digital twin of battery pack is established to emulate its dynamic behavior across various aging levels and inconsistency degrees. Then, increment capacity sequences (△Q) within a short voltage span are extracted from charging process to indicate battery health. Furthermore, data-driven models based on deep convolutional neural network (DCNN) are constructed to estimate battery state of health (SOH), where the synthetic data is employed to pre-train the models, and transfer learning strategies by using fine-tuning and domain adaptation are utilized to enhance the model adaptability. Finally, field data of 10 EVs exhibiting different SOHs are used to verify the proposed methods. By using the △Q with 100 mV voltage change, the SOH of battery packs can be accurately estimated with an error around 3.2%.

Mechanistic kinetic modelling of lipid oxidation in vegetable oils to estimate shelf-life

Estimating the shelf-life of vegetable oils is important to develop solutions to reduce spoilage by lipid oxidation. Typically, the shelf-life is predicted by detecting secondary oxidation markers in accelerated shelf-life tests, which are time-consuming. Existing numerical approaches using early primary oxidation products as predictive markers do not account for variations in fatty acid types, antioxidants, or storage conditions. A mechanistic kinetic model was developed incorporating these factors as a step towards shelf-life prediction for vegetable oils. Specific kinetic constants for the reactions of each unsaturated fatty acid type account for variations in fatty acid composition, and oxygen mass transfer accounts for variations in oxygen conditions. A second acceleration of lipid oxidation observed in long-term storage experiments was described by a multiplication factor for the kinetic constants related to oxidation products. Our model accurately extrapolates short-time experimental data to estimate long term formation of oxidation products under the same conditions.

Real-time drive-by bridge damage detection using deep auto-encoder

Structural health condition monitoring of bridge structures has been a concern in the last decades due to their aging and deterioration, in which the core task is damage detection. Recently, the drive-by method has gained much attention as it only needs several sensors installed on the passing vehicle. In this paper, we proposed an automatic damage detection method, which can be exploited in real time when the vehicle is passing the bridge. There are three steps in the proposed method: (1) The vehicle’s framed short-time vibrations instead of full-length data are utilized for training a deep auto-encoder model; at this stage, not commonly used time-domain accelerations of the passing vehicle, but its selected frequency-domain responses are employed to circumvent the influence of noises, (2) For the bridge with unknown health conditions, damage indicators can be extracted from its passing vehicle’s short-time vibration data using the trained model, and (3) The bridge’s health states are determined by real-time extracted damage indicators. To verify the proposed idea, a U-shaped continuous beam and a model truck are used to simulate the vehicle bridge interaction system in engineering. Results showed that the proposed method could identify the bridge’s damage with an accuracy of 86.2% when different severity was considered. In addition, it was observed that higher damage severity could not be revealed by greater values of damage indicators in the laboratory test. Instead, a novel index called identified damage ratios was employed as a reference for assessing the severity of the bridge’s damage. It was shown that with the increase in damage severity, the index would increase and gradually approach 100%.

Tikhonov regularization stabilizes multi-parameter estimation of geothermal heat exchangers

Parameter estimation from thermal response tests (TRTs) becomes unreliable when testing time reduces or the number of estimated parameters increases because of low identifiability and ill-posed mathematical feature. To overcome this challenge, this paper reports an inversion algorithm integrating a short-time temperature response model and the zero-order Tikhonov regularization strategy. We applied the algorithm to a reference sandbox dataset and examined four scenarios: simultaneous estimation of four, five, six, or seven parameters of U-shaped geothermal heat exchangers. The preliminary results indicate that the Tikhonov regularization can improve the accuracy and precision of the nonlinear multi-parameter estimation of ground heat exchangers for both long (>48 h) and short (<48 h) tests. The improved performance is contributed to the short-time model, which enables the short-time high-sensitivity data to be useable, and the regularization, which stabilizes the iterative optimization-solving procedure.

Machine Learning Methods in Predicting Patients with Suspected Myocardial Infarction Based on Short-Time HRV Data.

Diagnosis of cardiovascular diseases is an urgent task because they are the main cause of death for 32% of the world’s population. Particularly relevant are automated diagnostics using machine learning methods in the digitalization of healthcare and introduction of personalized medicine in healthcare institutions, including at the individual level when designing smart houses. Therefore, this study aims to analyze short 10-s electrocardiogram measurements taken from 12 leads. In addition, the task is to classify patients with suspected myocardial infarction using machine learning methods. We have developed four models based on the k-nearest neighbor classifier, radial basis function, decision tree, and random forest to do this. An analysis of time parameters showed that the most significant parameters for diagnosing myocardial infraction are SDNN, BPM, and IBI. An experimental investigation was conducted on the data of the open PTB-XL dataset for patients with suspected myocardial infarction. The results showed that, according to the parameters of the short ECG, it is possible to classify patients with a suspected myocardial infraction as sick and healthy with high accuracy. The optimized Random Forest model showed the best performance with an accuracy of 99.63%, and a root mean absolute error is less than 0.004. The proposed novel approach can be used for patients who do not have other indicators of heart attacks.

A high-accuracy initial alignment method based on backtracking process for strapdown inertial navigation system

When the alignment time is short, the traditional initial alignment method of strapdown inertial navigation system has very low accuracy, and even the attitude angle error is difficult to converge. We propose a high-accuracy initial alignment method for strapdown inertial navigation system based on the backtracking process, which reuses short-time measurement data to improve the accuracy and speed of initial alignment. We introduce the backtracking process into coarse alignment and fine alignment of strapdown inertial navigation system. In the coarse alignment, the instantaneous acceleration error caused by the backtracking process is estimated and compensated. In the fine alignment, the optimal initialization method of the parameters in the Kalman filter is given. The initial alignment experimental results of strapdown inertial navigation system prove the effectiveness of the proposed method. Compared with the previous initial alignment method, the accuracy and speed of the proposed method are improved.

A Novel Method for Diagnosis of Broken Bars Based on Optimum Resolution of Prescient Direction Algorithm

Fourier transformation (FT) and Multiple Signal Classification (MUSIC) method suffer the insufficient ability in diagnosing broken rotor bar (BRB) fault using short-time data. Theoretical and simulation analyses show that the Optimum Resolution of Prescient Direction (ORPD) algorithm has the best frequency resolution performance due to a priori knowledge. The main objective of this paper is to detect BRB faults in induction machines using a condition monitoring architecture based on ORPD algorithm. In the proposed application, the ORPD algorithm with the best frequency resolution performance is used to estimate the fault-sensitive frequencies in the stator current signature. The prior information of BRB fault characteristic distribution is used to construct a weighting matrix in the ORPD algorithm, for acquiring the lowest signal-to-noise ratio resolution threshold. Once frequencies are estimated, their corresponding amplitudes are obtained by using the least squares estimator. The proposed methods were tested using experimental induction motors with different fault severity under the effect of several load levels or supply frequencies. Two types of power supply modes are considered: main and inverter. The results show that compared with traditional method which uses FT and MUSIC algorithm for fault diagnosing, the method based on ORPD algorithm has a higher frequency resolution and identification ability with short-time data, and still has good diagnostic performance even under light loading and lower supply frequencies.

Characterization of the annual regime of surface solar irradiance over Argentine Pampean Region using GL1.2 satellite-based data

Detailed description of the solar resource (spatial distribution and temporal variations) is highly desirable in the context of climate monitoring, agricultural planning and energy technologies; however, one shortcoming is the low density of solarimetric networks in many extended regions. Satellite-based estimations have become a potential key tool, provided their data show adequate temporal frequency, high resolution and acceptable accuracy. GL1.2 Model currently runs at Brazilian National Institute for Space Research (CPTEC/INPE), based on GOES VIS channel imagery. It was tested for characterizing solar radiation regime over Pampa Region in Argentina, which extends over more than 600,000 km2. Comparison with time series of nine stations of the Pampa network in the period 2011–2018 showed close agreement for daily mean irradiance using 10-days as well as monthly values (mean bias error generally |MBE| < 6 W m-2 standard deviation STD ≈ 10-15 W m-2 and resulting root mean square error RMSE ≈ 10–16 W m−2). Completeness of GL time series and high spatial definition (0.04°) allowed further time and space analysis of solar radiation regime. It confirmed that Pampean region is rather homogeneous with annual mean GL increasing from 185 W m−2 on coastal region up to 220 W m−2 on the western and northern limits. In addition, GL data detects sharper variations in a scale of few kilometers, such as transition between ground and water in oceanic coast or large rivers. Regional interannual variability is low: COV ≈ 3–4%. All stations in Pampa network show a well definite annual cycle, closely followed by GL estimates. Fourier analysis for GL monthly series shows a dominant deterministic one year signal (H8 harmonic), accounting for 93–95% of variance for all stations and throughout the Pampa region, in phase with the end of Southern Hemisphere spring. The H8 amplitude varies from 100 W m−2 along inner continental boundary up to 120 in southwestern and coastal bound. It is concluded that GL 1.2 model data can be used to accurately describe time series of daily mean irradiance, for 10-days and monthly scales and 4 km resolution, providing a "Test Reference Year" (TRY). This is precious in order to describe spatial-temporal patterns of regional solar climate. An otherwise "random signal" of 20 W m−2 is associated to transient meteorological phenomena with durations shorter than 10-days, such as cold fronts or convective activity. Their deterministic and/or statistical structure could be improved by the combined analysis of short-time scales Pampa network and GL data (daily, hourly or even minutely).

A novel real-time modal analysis method for operational time-varying structural systems based on short-time extension of multivariate VMD

Engineering structures often exhibit time-varying characteristics due to environmental erosion or working conditions during their service life. Therefore, time-varying modal identification plays a crucial role in assessing the health condition of them. However, only the response is measurable, making the identification challenging. Recently, variational mode decomposition (VMD) based methods have been reported on output-only modal identification (OMI) of time-varying structures. Nevertheless, these methods cannot obtain time-varying mode shapes and identify overlapped or wideband modes. This paper develops a short-time method based on multivariate VMD (MVMD) to address these problems. Firstly, long-time non-stationary vibration signals are divided into small chunks, where modes show narrowband characteristics. Then, MVMD is employed on these short-time data chunks to extract multivariate modes. Finally, the instantaneous frequencies are recovered from the center frequencies, and the mode shapes are solved from the extracted multivariate modes. A series of numerical and experimental examples demonstrate that the short-time MVMD (STMVMD) has good noise-robustness, overlapped mode decomposition ability, and mode-tracking ability of time-varying structures.

MD-GAN with multi-particle input: the machine learning of long-time molecular behavior from short-time MD data.

Molecular dynamics simulation is a method of investigating the behavior of molecules, which is useful for analyzing a variety of structural and dynamic properties and mechanisms of phenomena. However, the huge computational cost of large-scale and long-time simulations is an enduring problem that must be addressed. MD-GAN is a machine learning-based method that can evolve part of the system at any time step, accelerating the generation of molecular dynamics data [Endo et al., Proceedings of the AAAI Conference on Artificial Intelligence, 2018, 32]. For the accurate prediction of MD-GAN, sufficient information on the dynamics of a part of the system should be included with the training data. Therefore, the selection of the part of the system is important for efficient learning. In a previous study, only one particle (or vector) of each molecule was extracted as part of the system. The effectiveness of adding information from other particles to the learning process is investigated in this study. When the dynamics of three particles of each molecule were used in the polyethylene experiment, the diffusion was successfully predicted using the training data with a time length of approximately 40%, compared to the single-particle input. Surprisingly, the unobserved transition of diffusion in the training data was also predicted using this method. The reduced cost for the generation of training MD data achieved in this study is useful for accelerating MD-GAN.

Bearing Fault Diagnosis With Frequency Sparsity Learning

Extracting fault frequencies from noisy vibration signal is a challenging task for bearing fault diagnosis. The state-of-the-art sparse representation (SR) based methods usually consist of two steps: (i) fault impulse recovery in the time-domain and (ii) frequency transformation of the estimated signal envelope. However, any inaccurate time-domain signal recovery can cause an error accumulation problem for the following frequency transformation, and the frequency transformation itself encounters a low-resolution shortcoming especially for short-time sampling data. To handle these shortcomings, in this paper, we propose a novel sparse Bayesian learning (SBL) framework to evade the time-domain signal recovery and extract the fault frequencies directly from the frequency domain. We first present a new formulation for the sparse frequency recovery problem by utilizing the sparsity structure of the envelope spectrum, and then introduce a truncated off-grid model into the SBL framework to speed up the proposed method. Moreover, an improved grid refinement is developed to jointly combat the off-grid frequency mismatch and exploit the arithmetic sparsity structure of fault frequencies. Both simulation and experiment results indicate the effectiveness of our proposed method.

Quantitative analysis of non-equilibrium systems from short-time experimental data

Estimating entropy production directly from experimental trajectories is of great current interest but often requires a large amount of data or knowledge of the underlying dynamics. In this paper, we propose a minimal strategy using the short-time Thermodynamic Uncertainty Relation (TUR) by means of which we can simultaneously and quantitatively infer the thermodynamic force field acting on the system and the (potentially exact) rate of entropy production from experimental short-time trajectory data. We benchmark this scheme first for an experimental study of a colloidal particle system where exact analytical results are known, prior to studying the case of a colloidal particle in a hydrodynamical flow field, where neither analytical nor numerical results are available. In the latter case, we build an effective model of the system based on our results. In both cases, we also demonstrate that our results match with those obtained from another recently introduced scheme.

Genome-wide identification and expression analysis of LBD transcription factor genes in Moso bamboo (Phyllostachys edulis)

BackgroundMoso bamboo, the fastest growing plant on earth, is an important source for income in large areas of Asia, mainly cultivated in China. Lateral organ boundaries domain (LBD) proteins, a family of transcription factors unique to plants, are involved in multiple transcriptional regulatory pathways and play important roles in lateral organ development, pathogen response, secondary growth, and hormone response. The LBD gene family has not previously been characterized in moso bamboo (Phyllostachys edulis).ResultsIn this study, we identified 55 members of the LBD gene family from moso bamboo and found that they were distributed non-uniformly across its 18 chromosomes. Phylogenetic analysis showed that the moso bamboo LBD genes could be divided into two classes. LBDs from the same class share relatively conserved gene structures and sequences encoding similar amino acids. A large number of hormone response–associated cis-regulatory elements were identified in the LBD upstream promoter sequences. Synteny analysis indicated that LBDs in the moso bamboo genome showed greater collinearity with those of O. sativa (rice) and Zea mays (maize) than with those of Arabidopsis and Capsicum annuum (pepper). Numerous segmental duplicates were found in the moso bamboo LBD gene family. Gene expression profiles in four tissues showed that the LBD genes had different spatial expression patterns. qRT–PCR assays with the Short Time-series Expression Miner (STEM) temporal expression analysis demonstrated that six genes (PeLBD20, PeLBD29, PeLBD46, PeLBD10, PeLBD38, and PeLBD06) were consistently up-regulated during the rapid growth and development of bamboo shoots. In addition, 248 candidate target genes that function in a variety of pathways were identified based on consensus LBD binding motifs.ConclusionsIn the current study, we identified 55 members of the moso bamboo transcription factor LBD and characterized for the first time. Based on the short-time sequence expression software and RNA-seq data, the PeLBD gene expression was analyzed. We also investigated the functional annotation of all PeLBDs, including PPI network, GO, and KEGG enrichment based on String database. These results provide a theoretical basis and candidate genes for studying the molecular breeding mechanism of rapid growth of moso bamboo.

Short-Term Forecasting of Daily Confirmed COVID-19 Cases in Malaysia Using RF-SSA Model.

Novel coronavirus (COVID-19) was discovered in Wuhan, China in December 2019, and has affected millions of lives worldwide. On 29th April 2020, Malaysia reported more than 5,000 COVID-19 cases; the second highest in the Southeast Asian region after Singapore. Recently, a forecasting model was developed to measure and predict COVID-19 cases in Malaysia on daily basis for the next 10 days using previously-confirmed cases. A Recurrent Forecasting-Singular Spectrum Analysis (RF-SSA) is proposed by establishing L and ET parameters via several tests. The advantage of using this forecasting model is it would discriminate noise in a time series trend and produce significant forecasting results. The RF-SSA model assessment was based on the official COVID-19 data released by the World Health Organization (WHO) to predict daily confirmed cases between 30th April and 31st May, 2020. These results revealed that parameter L = 5 (T/20) for the RF-SSA model was indeed suitable for short-time series outbreak data, while the appropriate number of eigentriples was integral as it influenced the forecasting results. Evidently, the RF-SSA had over-forecasted the cases by 0.36%. This signifies the competence of RF-SSA in predicting the impending number of COVID-19 cases. Nonetheless, an enhanced RF-SSA algorithm should be developed for higher effectivity of capturing any extreme data changes.

Dynamic Heterogeneity of Filler-Associated Interphases in Polymer Nanocomposites.

Dynamically inhomogeneous polymer systems exhibit interphases with mobility gradients. These are believed to play key roles in the material's performance. A prominent example is particle-filled rubber, a special case of a crosslinked polymer nanocomposite, where favorable rubber-filler interactions may give rise to a nanoscale immobilized layer around the filler, including regions of intermediate mobility. Such intermediate domains may either form a separate shell-like layer or be a manifestation of dynamic heterogeneities, in which case the intermediately mobile material would be dispersed in the form of nanometer-sized subdomains. In this contribution, bidirectional proton NMR spin diffusion (SD) experiments applied to silica-filled acrylate rubber are combined with numerical simulations to provide microscopic insights into this question. While model calculations for different scenarios fit the given data similarly well for longer SD mixing time, the short-time data do support the presence of dynamic heterogeneities.