Search

GoogLeNet is a pre-trained Convolutional Neural Network (CNN) that allows transfer learning and has achieved high recognition rates in image classification tasks. A Recurrence Plot (RP) is an imaging method that depicts the recurrence of the state space system using coloured points and lines in 2D images. This work contributes to facilitating time series feature extraction by proposing a method that applies the GoogLeNet to time series images obtained with RP. The developed method is tested using simulated time series and selected time series from the M3 competition dataset. The results shows that the transfer learning approach allowed the extraction of business time series features by means of a GoogLeNet fine-tuned using 100 simulated time series. The combination of GoogLeNet and RPs outperforms the alternative and easier combination of GoogLeNet and plots of the time series and support the convenience of the RP transformation step. This application of deep learning techniques to business time series imaging offers opportunity for further developments. [ABSTRACT FROM AUTHOR]

Parkinson's disease (PD) is a progressive, debilitating neurological movement disorder that affects the person's muscle control, movement, speech, cognition and dexterity. For diagnosing PD in a clinical setting, in addition to the neurological examinations, clinicians use the unified Parkinson disease rating scale (UPDRS) to assess the motor and non-motor impairments. Such a clinical assessment highly depends on the experience and expertise of the clinicians, and it may result in biased evaluation. Hence, to assist the clinicians, we put forward a gait analysis-based deep convolutional neural network (DCNN) framework which leverages the potentials of variational mode decomposition (VMD) technique with the recurrence plots (RP) to enhance the PD severity classification performance. Specifically, transforming the VMD modes of vertical ground reaction force (VGRF) time series data into two-dimensional texture images to capture the temporal dependency, this work trains the DCNN classifier through recurrence images for its ability to extract the discriminative features among the PD severity levels. For evaluation, this study utilises the VGRF dataset of 93 PD subjects and 73 healthy controls from Physiobank for three different walking tests. Consequently, utilising VMD, RP and DCNN in a unified framework, this investigation shows that the PD severity rating can be significantly enhanced through DCNN model that is trained using RP of dominant intrinsic mode functions (IMFs). The novelty of the proposed framework lies in identifying the prominent gait biomarkers through dominant IMFs from power spectral analysis for reducing the computational burden of DCNN. Moreover, to handle the data over-fitting issue in the classifier, L2 regularisation technique, which penalises the weight parameters of the nodes, is used in combination with the dropout layer. Experimental results underscore that the proposed VMD-RP-DCNN architecture can address the spectral overlapping issue in VGRF decomposition and achieve an average PD severity prediction accuracy of 98.45%. [ABSTRACT FROM AUTHOR]

Copyright of Bearing is the property of Bearing Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Purpose: This paper proposes the application of Recurrence Plot (RP) to analyze dry friction on mechanical rotating systems. Method: The construction of Recurrence Plots is based on the evaluation of the system dynamics represented in a phase plane. The phase plane represents the relationship between displacement and velocity; but in this case, the vibration signals are measured with accelerometers; thus the displacement and velocity function must be obtained integrating the field data. In this paper, a comparison of two integration methods is included, one of the methods is based on the Simpson's integration, and the other is based on the empirical model decomposition (EMD) and the shifting principle of harmonic functions. The experiment consisted of a flexible rotor with an unbalance disk that rub a fixed plate. The tests included different operating speeds, keeping the unbalance constant. The vibrations were recorded with an accelerometer mounted on the rotor's support and a laser vibrometer (LV) that measured the lateral displacements of the disk. Results: The results indicated that the RP patterns show significant differences when the friction load is applied, but only some quantification parameters reflect these differences. The RP were evaluated using quantitative parameters that highlighted the nonlinear effect caused by friction. [ABSTRACT FROM AUTHOR]

Early and highly precise detection is essential for delaying the progression of coronary artery disease (CAD). Previous methods primarily based on single-modal data inherently lack sufficient information that compromises detection precision. This paper proposes a novel multi-modal learning method aimed to enhance CAD detection by integrating ECG, PCG, and coupling signals. A novel coupling signal is initially generated by operating the deconvolution of ECG and PCG. Then, various entropy features are extracted from ECG, PCG, and its coupling signals, as well as recurrence deep features also encoded by integrating recurrence plots and a parallel-input 2-D CNN. After feature reduction and selection, final classification is performed by combining optimal multi-modal features and support vector machine. This method was validated on simultaneously recorded standard lead-II ECG and PCG signals from 199 subjects. The experimental results demonstrate that the proposed multi-modal method by integrating all signals achieved a notable enhancement in detection performance with best accuracy of 95.96%, notably outperforming results of single-modal and joint analysis with accuracies of 80.41%, 86.51%, 91.44%, and 90.42% using ECG, PCG, coupling signal, and joint ECG and PCG, respectively. This indicates that our multi-modal method provides more sufficient information for CAD detection, with the coupling information playing an important role in classification. [ABSTRACT FROM AUTHOR]

Coronary artery disease (CAD) is an irreversible and fatal disease. It necessitates timely and precise diagnosis to slow CAD progression. Electrocardiogram (ECG) and phonocardiogram (PCG), conveying abundant disease-related information, are prevalent clinical techniques for early CAD diagnosis. Nevertheless, most previous methods have relied on single-modal data, restricting their diagnosis precision due to suffering from information shortages. To address this issue and capture adequate information, the development of a multi-modal method becomes imperative. In this study, a novel multi-modal learning method is proposed to integrate both ECG and PCG for CAD detection. Along with deconvolution operation, a novel ECG-PCG coupling signal is evaluated initially to enrich the diagnosis information. After constructing a modified recurrence plot, we build a parallel CNN network to encode multi-modal information, involving ECG, PCG and ECG-PCG coupling deep-coding features. To remove irrelevant information while preserving discriminative features, we add an autoencoder network to compress feature dimension. Final CAD classification is conducted by combining support vector machine and optimal multi-modal features. The experiment is validated on 199 simultaneously recorded ECG and PCG signals from non-CAD and CAD subjects, and achieves high performance with accuracy, sensitivity, specificity and f1-score of 98.49%, 98.57%,98.57% and 98.89%, respectively. The result demonstrates the superiority of the proposed multi-modal method in overcoming information shortages of single-modal signals and outperforming existing models in CAD detection. This study highlights the potential of multi-modal deep-coding information, and offers a wider insight to enhance CAD diagnosis. [ABSTRACT FROM AUTHOR]

The objective of this paper is to showcase the capability of the conventional circuit structure known as the Lumpkin oscillator, widely employed in practical applications, to operate in robust chaotic or hyperchaotic steady states. Through numerical analysis, we demonstrate that the generated signals exhibit a significant level of unpredictability and randomness, as evidenced by positive Lyapunov exponents, approximate entropy, recurrence plots, and other indicators of complex dynamics. We establish the structural stability of strange attractors through design and practical construction of a flow-equivalent fourth-order chaotic oscillator, followed by experimental measurements. The oscilloscope screenshots captured align well with the plane projections of the approximate solutions derived from the underlying mathematical models. [ABSTRACT FROM AUTHOR]

Atrial fibrillation (AF) is the most prevalent form of arrhythmia, with a rising incidence and prevalence worldwide, posing significant implications for public health. In this paper, we introduce an approach that combines the Recurrence Plot (RP) technique and the ResNet architecture to predict AF. Our method involves three main steps: using wavelet filtering to remove noise interference; generating RPs through phase space reconstruction; and employing a multi-level chained residual network for AF prediction. To validate our approach, we established a comprehensive database consisting of electrocardiogram (ECG) recordings from 1008 AF patients and 48,292 Non-AF patients, with a total of 2067 and 93,129 ECGs, respectively. The experimental results demonstrated high levels of prediction precision (90.5%), recall (89.1%), F1 score (89.8%), accuracy (93.4%), and AUC (96%) on our dataset. Moreover, when tested on a publicly available AF dataset (AFPDB), our method achieved even higher prediction precision (94.8%), recall (99.4%), F1 score (97.0%), accuracy (97.0%), and AUC (99.7%). These findings suggest that our proposed method can effectively extract subtle information from ECG signals, leading to highly accurate AF predictions. [ABSTRACT FROM AUTHOR]

The development of a bridge damage detection method relies on comprehensive dynamic responses pertaining to damage. The numerical model of a bridge can conveniently considers various damage scenarios and acquire pertinent data, while the entity of a bridge or its physical model proves challenging. Traditional methods for identifying bridge damage often struggle to effectively utilize data acquired from diverse domains, presenting a significant hurdle in addressing cross-domain issues. This study proposes a novel cross-domain damage identification method for suspension bridges using recurrence plots and convolutional neural networks. By employing parameter identification-based modal modification of numerical model, the gap between numerical model and physical models eliminated. Un-threshold multivariate recurrence plots are used for accurately characterizing dynamic responses and extracting deeper damage features. Due to the scarcity of experimental data, which limits the training of robust neural networks, a transfer learning tailored for convolutional neural networks is implemented. This strategy not only addresses the issue of small sample sizes but also significantly enhances the network's ability to identify structural damage across diverse bridge domains. The proposed damage identification method is validated using a combination of numerical simulations and physical experiments on a specific single-span suspension bridge. Results demonstrate that un-threshold multivariate recurrence plots reveal detailed internal structure and damage information. Furthermore, the utilization of improved convolutional neural networks effectively facilitates cross-domain structural damage identification, marking a significant advancement in the field of structural health monitoring. [ABSTRACT FROM AUTHOR]

The non-linear and non-stationary characteristics of vibration signals in rolling bearings make it difficult to accurately extract fault features. In addition, traditional fault diagnosis methods cannot fully explore the correlation characteristics between time-series of fault signals. To address the aforementioned issues, this paper introduces a recurrence plot (RP) coding technique into the field of fault diagnosis and proposes a bearing fault diagnosis method based on RP and the improved EfficientNen/2-S. Firstly, the method uses the RP coding technique to convert one-dimensional vibration signals into two-dimensional time-frequency images as inputs to the neural network. Then, the number of layers in the EfficientNet*S network is optimised by a non-linear attenuation strategy to reduce network parameters and improve the recognition speed. Finally, the attention mechanism is modified and the variable load dataset is constructed for training to improve the feature extraction ability and generalisation performance of the model. To verify the effectiveness of the proposed method, experiments are conducted based on the bearing datasets provided by Case Western Reserve University (CWRU). The experimental results show that the bearing fault diagnosis method based on RP and the improved EfficientNet\/2-S cannot only realise accurate identification of bearing faults but also accurately identify the degree of bearing fault with an accuracy of 99.85%. [ABSTRACT FROM AUTHOR]

The main goal of this work is summed up in a univariate chaotic analysis of the runoff series using a topological approach. As such, nineteen series of daily runoff from eight large watersheds in northern Algeria were analyzed. Firstly, preliminary analyses of two traditional dynamic and metric approaches have been tested. In the dynamic approach, three algorithms have shown that the largest Lyapunov exponent for the series is positive, which supports the hypothesis of the existence of chaos. In the metric approach, the Grassberger and Procaccia algorithm clearly shows the saturation of the correlation dimension, which indicates the existence of deterministic dynamics for all studied stations. Secondly, the application of the topological approach in this study constitutes in itself a new contribution to the demonstration of chaos in hydrology by using the recurrence plot (RP) and the recurrence quantification analysis (RQA). As such, the RP structures of the runoff series seem to be more comparable to chaotic systems. In addition, RQA parameters give high values of determinism and laminarity, which supports the hypothesis of the existence of deterministic chaos. The presence of chaos in the runoff series can be identified by the existence of a strong probability of recurrence, indicating a fairly low level of complexity and fairly high predictability. Ultimately, the comparison of these approaches together made it possible to confirm the hypothesis according to which the process generating the runoff series is deterministic and suggests low-dimensional chaotic dynamics. [ABSTRACT FROM AUTHOR]

In various fields such as healthcare and transportation, accurately classifying time series data can provide important support for decision-making. To further improve the accuracy of time series classification, we propose a Feature-fused Residual Network based on Multi-scale Signed Recurrence Plot (MSRP-FFRN). This method transforms one-dimensional time series into two-dimensional images, representing the temporal correlation of time series in a two-dimensional space and revealing hidden details within the data. To enhance these details further, we extract multi-scale features by setting receptive fields of different sizes and using adaptive network depths, which improves image quality. To evaluate the performance of this method, we conducted experiments on 43 UCR datasets and compared it with nine state-of-the-art baseline methods. The experimental results show that MSRP-FFRN ranks first on critical difference diagram, achieving the highest accuracy on 18 datasets with an average accuracy of 89.9%, making it the best-performing method overall. Additionally, the effectiveness of this method is further validated through metrics such as Precision, Recall, and F1 score. Results from ablation experiments also highlight the efficacy of the improvements made by MSRP-FFRN.

A fault diagnosis method based on fused RP (recurrence plot) and improved CNN (convolutional neural networks) is proposed for the traditional fault diagnosis method of linear vibrating screen manually designed and optimized features with feature quality uncertainty. A deep convolutional neural network combination model (MDCNN) with high-level feature fusion is designed. The collected multi-source vibration signals were converted into black and white recurrence plot, and fusion into three-channel RGB recurrence plot. Fault diagnosis was classified by the combination model of deep convolution neural network with high level feature fusion. The results show that the recurrence plot of homologous multi-sensor signals contains more information about the characteristics of the vibrating screen than the original colorful signal map or the Gram Point Field, which contributes to the feature learning and classification of MDCNN, with an average recognition accuracy of 97.59% under strong vibrating sift-5db noise. Compared with manifold classical neural network models, the MDCNN model is proved to be effective, stable and small space occupying. [ABSTRACT FROM AUTHOR]

Functional gastric disorders entail chronic or recurrent symptoms, high prevalence and a significant financial burden. These disorders do not always involve structural abnormalities and since they cannot be diagnosed by routine procedures, electrogastrography (EGG) has been proposed as a diagnostic alternative. However, the method still has not been transferred to clinical practice due to the difficulty of identifying gastric activity because of the low-frequency interference caused by skin–electrode contact potential in obtaining spatiotemporal information by simple procedures. This work attempted to robustly identify the gastric slow wave (SW) main components by applying multivariate variational mode decomposition (MVMD) to the multichannel EGG. Another aim was to obtain the 2D SW vectorgastrogram VGGSW from 4 electrodes perpendicularly arranged in a T-shape and analyse its dynamic trajectory and recurrence quantification (RQA) to assess slow wave vector movement in healthy subjects. The results revealed that MVMD can reliably identify the gastric SW, with detection rates over 91% in fasting postprandial subjects and a frequency instability of less than 5.3%, statistically increasing its amplitude and frequency after ingestion. The VGGSW dynamic trajectory showed a statistically higher predominance of vertical displacement after ingestion. RQA metrics (recurrence ratio, average length, entropy, and trapping time) showed a postprandial statistical increase, suggesting that gastric SW became more intense and coordinated with a less complex VGGSW and higher periodicity. The results support the VGGSW as a simple technique that can provide relevant information on the "global" spatial pattern of gastric slow wave propagation that could help diagnose gastric pathologies. [ABSTRACT FROM AUTHOR]

We investigate the response characteristics of a generalized FitzHugh-Nagumo model under an external sinusoidal current and the synchronization of two neurons coupled with a gap junction. In the autonomous case, we find analytically by the Lindsted's method that the system can admit tristable activities in silence, subthreshold, and nerve pulse; depending on the conductance parameters and the state of ionic conductance. In the presence of an external sinusoidal current, we find by numerical simulations that neurons can exhibit a coexistence between different spiking patterns and periodic waves, which are well observed in the structure of the recurrence plot. We further study the synchronization between coupled neurons each admitting bistable activities, such as a coexistence between chaotic (active) and silence (inactive) regimes. We apply recurrence analysis tool to reveal the range of the coupling parameter where synchronization occurs, as well as the dynamical transitions between the synchronous coexisting states (hysteresis phenomenon). The coupling strength is an indicator of the phenomenon of synchronization that can also bring the system to any of the desired synchronous attractors. These phenomena of synchronization and the control between synchronous states can be improved by the presence of an external electrical field. The switching of the coupled neurons to bursting patterns or to periodic waves explains the well-known properties of excitatory (switching on) or inhibitory (switching off) synaptic coupling, respectively; while the unstable signal separating the two stable synchronous signals can be taken as the synaptic threshold. Rather, this study adds to our theoretical understanding of the topic and poses new challenges for investigation. Experimental investigations are required to validate these conclusions in real-world settings, and biological implications must be evaluated within the particular framework of the modeling that was done. [ABSTRACT FROM AUTHOR]

The paper presents a system for analyzing cardiac activity with the possibility of continuous and remote monitoring. The created sensor mobile device monitors heart activity by means of the convenient and imperceptible registration of cardiac signals. At the same time, the behavior of the human body is also monitored through the accelerometer and gyroscope built into the device, thanks to which it is possible to signal in the event of loss of consciousness or fall (in patients with syncope). Conducting real-time cardio monitoring and the analysis of recordings using various mathematical methods (linear, non-linear, and graphical) enables the research, accurate diagnosis, timely assistance, and correct treatment of cardiovascular diseases. The paper examines the recordings of patients diagnosed with arrhythmia and syncope recorded by electrocardiography (ECG) sensors in real conditions. The obtained results are subjected to statistical analysis to determine the accuracy and significance of the obtained results. The studies show significant deviations in the patients with arrhythmia and syncope regarding the obtained values of the studied parameters of heart rate variability (HRV) from the accepted normal values (for example, the root mean square of successive differences between normal heartbeats (RMSSD) in healthy individuals is 24.02 ms, while, in patients with arrhythmia (6.09 ms) and syncope (5.21 ms), it is much lower). The obtained quantitative and graphic results identify some possible abnormalities and demonstrate disorders regarding the activity of the autonomic nervous system, which is directly related to the work of the heart. [ABSTRACT FROM AUTHOR]

The goal of the research is to describe the aggregation process inside the mucilage produced by plant seeds using molecular dynamics (MD) combined with time series algorithmic analysis based on the recurrence plots. The studied biological molecules model is seed mucilage composed of three main polysaccharides, i.e. pectins, hemicellulose, and cellulose. The modeling of biological molecules is based on the assumption that a classical–quantum passage underlies the aggregation process in the mucilage, resulting from non-covalent interactions, as they affect the macroscopic properties of the system. The applied recurrence plot approach is an important tool for time series analysis and data mining dedicated to analyzing time series data originating from complex, chaotic systems. In the current research, we demonstrated that advanced algorithmic analysis of seed mucilage data can reveal some features of the dynamics of the system, namely temperature-dependent regions with different dynamics of increments of a number of hydrogen bonds and regions of stable oscillation of increments of a number of hydrophobic–polar interactions. Henceforth, we pave the path for automatic data-mining methods for the analysis of biological molecules with the intermediate step of the application of recurrence plot analysis, as the generalization of recurrence plot applications to other (biological molecules) datasets is straightforward. [ABSTRACT FROM AUTHOR]

Preparing regular time series optical remote sensing data is a difficult task due to the influences of frequently cloudy and rainy days. The irregular data and their forms severely limit the data’s ability to be analyzed and modeled for vegetation classification. However, how irregular time series data affect vegetation classification in deep learning models is poorly understood. To address these questions, this research preprocessed the 2019–2021 time series of Sentinel-2 in both unequal and equal intervals, and transformed them into an image through recurrence plot for each pixel. The initial one-dimension time series (1DTS) and recurrence plot data were then used as input data for three deep learning methods (i.e. Conv1D model based on one-dimensional convolution, GoogLeNet model based on two-dimensional convolution, and CGNet model which fused Conv1D and GoogLeNet) for vegetation classification, respectively. The class separability of the features generated by each model was evaluated and the importance of spectral and temporal features was further examined through gradient backpropagation. The equal-interval time series data significantly improved the classification accuracy with 0.04, 0.13, and 0.09 for Conv1D, GoogLeNet, and CGNet, respectively. The CGNet achieved the highest classification accuracy, indicating that the information from 1DTS and recurrence plot can be a good complementary for vegetation classification. The importance of spectral bands and time showed that the Sentinel-2 red edge-1 spectral band played a critical role in the identification of eucalyptus, loquat, and honey pomelo, but the importance order of bands varied in different vegetation types in GoogLeNet. The time importance varied across different vegetation types but is similar in these deep learning models. This study quantified the impacts of organizational form (1DTS and recurrence plot) of time series data on different models. This research is valuable for us to choose appropriate data structures and efficient deep learning models for vegetation classification.

Large-span spatial structure damage identification is a challenging element of structural health monitoring. Compared with other buildings such as bridges and frames, space structures are characterized by large spans, many degrees of freedom and complex structures. Therefore, this paper proposes a new step-by-step damage identification method for spatial structures based on vibration signals. The method uses recurrence plot to process the structural vibration response to obtain nonlinear features. Through the nonlinear features reacting to different damage conditions of the structure and introducing convolutional neural network to realize the classification recognition problem under different damages. The feasibility analysis of step-by-step identification of damaged nodes and damaged rods is carried out with an orthogonal orthotropic quadrangular cone mesh structure model as an example. The optimized model training methods of data augmentation and migration learning are also introduced. An overall recognition accuracy of more than 89.7% is obtained. In order to realize the application of the proposed loss identification method in practical engineering, an operable GUI interface is constructed by encapsulating with programming technology. Afterwards, the complete step-by-step damage identification method from substructure to rod was verified by combining field tests and numerical simulations using a single-layer column surface mesh shell model consisting of 157 nodes and 414 rods. The results show that the damage recognition method has more than 85% recognition accuracy for structural damage. To explain the effectiveness of the convolutional neural network model training visualization of the recognition image features is performed using class activation heat maps. [ABSTRACT FROM AUTHOR]

The existence of chaos is particularly relevant, as the identification of a chaotic behavior in a time series could lead to reliable short-term forecasting. This paper evaluates the existence of nonlinearity and chaos in the underlying process of the spot prices of the Spanish electricity market. To this end, we used daily data spanning from 1 January 2013, to 31 March 2021 and we applied a comprehensive framework that encompassed a wide range of techniques. Nonlinearity was analyzed using the BDS method, while the existence of a chaotic structure was studied through Lyapunov exponents, recurrence plots, and quantitative recurrence analysis. While nonlinearity was detected in the underlying process, conclusive evidence supporting chaos was not found. In addition, the generalized autoregressive conditional heteroscedastic (GARCH) model accounts for part of the nonlinear structure that is unveiled in the electricity market. These findings hold substantial value for electricity market forecasters, traders, producers, and market regulators. [ABSTRACT FROM AUTHOR]

Electromyography (EMG) is used to identify neuromuscular illnesses, motor issues, nerve damage, and degenerative ailments. EMG signals are difficult to accurately classify because of their complexity, nonlinearity, and time-variable nature; therefore, proper feature extraction and classification algorithms must be used. The proposed study uses the recurrence plot (RP) approach combined with deep learning (DL) techniques for the purpose of normal and aggressive EMG physical action (PA) classification. The transfer learning (TL) approach Inception-ResNet-v2 is utilized for feature extraction here. The Inception-ResNet-v2 model's results demonstrate the excellent generalizability of the proposed method. Using only a few of the eight leads in the original EMG data, the model demonstrated the best accuracy (0.9841) and F1 score (0.99). These results show that the 2D RP-based technique has a significant clinical potential for classifying PA while requiring fewer EMG leads. [ABSTRACT FROM AUTHOR]

RPThe wheel condition monitoring when the train in operation is significant task to prevent the occurrence of unexpected event. In this study, the piezoelectric sensors were installed on the railway track to collect the dynamic voltage-and-strain signals when the train wheels pressed them. These one-dimensional time series signals were transformed to the two-dimensional Recurrence Plots (RP) images as an input data sets for two deep learning models, Xception and EfficientNet-B7. The binary classification, Normal or Faulty as the diagnostical output to indicate the health state of the train wheels in that time. Five metrics were selected to evaluate the performance of two models, namely Accuracy, Precision, Recall, Miss Rate, and AUC. The results show that both models perform the high accuracy of 91.1% to the wheel condition classification. Furthermore, EfficientNet-B7 shows better performance in Recall, Miss-rate, and AUC metrics than those of Xception to express the premium ability in defective wheel identification, which is crucial for this application. Therefore, the efficientNet-B7 is selected as a favorable machine learning classifier for the fault diagnosis of rolling stock wheels. It is significant contribution to train wheel condition monitoring and health management since it provides the effective diagnostic information for maintenance decision to decrease the occurrence of unexpected event. [ABSTRACT FROM AUTHOR]

In this paper, the vibration-based image representation and data fusion demonstrates distinctive benefit in feature extraction, yielding superior performance for damage identification in railway engineering. Specifically, based on vehicle-track coupled dynamics, the rail vibration datasets under diverse fastener damage conditions are generated. By converting 1-D vibration signals into 2-D grayscale images with recurrence plots (RPs) and the aid of conditional variational autoencoder (CVAE), the acceleration RPs and displacement RPs are fused for enhancing feature extraction. It is demonstrated that detecting the variation in texture patterns and color distribution of the vibration-based images facilitates effective damage identification, mitigating the sensitivity of damage recognition to the deterioration of track irregularity. The results show that the displacement RPs characterised by quasi-static features are more suitable for fastener damage identification. Further, by employing the data fusion that combines both the random dynamic features of the acceleration RPs and quasi-static features of the displacement RPs, the tolerance of measurement range for accurate fastener damage identification can be extended. The robustness of the proposed method is validated after testing different sampling frequencies and additional noise. [ABSTRACT FROM AUTHOR]

Spectral-based features, typically used in ASR systems, do not capture the phase information of speech signals. Thus, exploiting new features that do not ignore the phase of the signal can be a complementary approach to improve the performance of the feature extraction (FE) block of an ASR system. In this paper, we propose an adaptive FE method that uses the reconstructed phase space (RPS) and recurrence plot (RP) theories as its foundations. The RP transformation can reveal some important aspects of the dynamics of high-dimensional speech trajectories reconstructed in the RPS. In this work, after transforming the speech signal to the image-like RP domain as a matrix, we apply a powerful wavelet-based FE method. We use a two-dimensional adaptive wavelet transform, implemented through a customized filter bank, to extract some beneficial dynamical features from the RP matrix for the ASR task. We evaluate the resulting features in an ASR task alone and in combination with the traditional MFCCs. Using the TIMIT speech corpus, the combination of the proposed and MFCC features results in a relative improvement of 7.79% in phoneme recognition accuracy rate compared to using only the MFCC features. [ABSTRACT FROM AUTHOR]

Recurrence lacunarity has been recently proposed to detect dynamical state transitions over various temporal scales. In this paper, we combine suggested distribution moments and introduce multifractal recurrence lacunarity to unearth rich information of trajectories in phase space. By considering generalized moments, it provides an enhanced measurement to account for differences of black pixels in the recurrence plot at various scales. Numerical simulations have proved that the proposed method is able to differentiate varying types of time series and provide further insights of inherent features including stochastic series, chaotic maps and series contaminated interference components. In real-world applications, it performs well on quantifying the subtle structural changes of financial time series. In addition, it is intriguing to confirm that corrugation signals possess much more vivid information of heterogeneity in terms of recurrence plots than normal ones. [ABSTRACT FROM AUTHOR]

This paper proposes a novel multivariate performance diagnostic approach for the Oracle database systems to detect performance degradation and crashes during database operations and maintenance. It was based on three technologies: image encoding, image concatenation, and deep convolutional network. Instantaneous variation and magnitude information of the time series were acquired by Heatmap and Recurrence Plot (RP) from databases. Moreover, the Heatmap and RP were concatenated in order to fully extract complementary information. Finally, the concatenated images of Heatmap and RP were used to train the VGG16 model for database performance diagnosis. The quantitative analysis demonstrated a good increment of accuracy in Heatmap and RP based on the same deep learning networks compared with other image encodings of Gramian Angular Difference Field (GADF), Gramian Angular Summation Fields (GASF), and Markov Transition Fields (MTF). Meanwhile, concatenated images of Heatmap and RP can improve the accuracy by 1% compared to single Heatmap input. The result of the trained model applied to multivariate database diagnosis shows an accuracy of 95.3%. Thus, our method can more effectively and accurately diagnose the performance of the Oracle database systems.

Early and highly precise detection is essential for delaying the progression of coronary artery disease (CAD). Previous methods primarily based on single-modal data inherently lack sufficient information that compromises detection precision. This paper proposes a novel multi-modal learning method aimed to enhance CAD detection by integrating ECG, PCG, and coupling signals. A novel coupling signal is initially generated by operating the deconvolution of ECG and PCG. Then, various entropy features are extracted from ECG, PCG, and its coupling signals, as well as recurrence deep features also encoded by integrating recurrence plots and a parallel-input 2-D CNN. After feature reduction and selection, final classification is performed by combining optimal multi-modal features and support vector machine. This method was validated on simultaneously recorded standard lead-II ECG and PCG signals from 199 subjects. The experimental results demonstrate that the proposed multi-modal method by integrating all signals achieved a notable enhancement in detection performance with best accuracy of 95.96%, notably outperforming results of single-modal and joint analysis with accuracies of 80.41%, 86.51%, 91.44%, and 90.42% using ECG, PCG, coupling signal, and joint ECG and PCG, respectively. This indicates that our multi-modal method provides more sufficient information for CAD detection, with the coupling information playing an important role in classification.

Coronary artery disease (CAD) is an irreversible and fatal disease. It necessitates timely and precise diagnosis to slow CAD progression. Electrocardiogram (ECG) and phonocardiogram (PCG), conveying abundant disease-related information, are prevalent clinical techniques for early CAD diagnosis. Nevertheless, most previous methods have relied on single-modal data, restricting their diagnosis precision due to suffering from information shortages. To address this issue and capture adequate information, the development of a multi-modal method becomes imperative. In this study, a novel multi-modal learning method is proposed to integrate both ECG and PCG for CAD detection. Along with deconvolution operation, a novel ECG-PCG coupling signal is evaluated initially to enrich the diagnosis information. After constructing a modified recurrence plot, we build a parallel CNN network to encode multi-modal information, involving ECG, PCG and ECG-PCG coupling deep-coding features. To remove irrelevant information while preserving discriminative features, we add an autoencoder network to compress feature dimension. Final CAD classification is conducted by combining support vector machine and optimal multi-modal features. The experiment is validated on 199 simultaneously recorded ECG and PCG signals from non-CAD and CAD subjects, and achieves high performance with accuracy, sensitivity, specificity and f1-score of 98.49%, 98.57%,98.57% and 98.89%, respectively. The result demonstrates the superiority of the proposed multi-modal method in overcoming information shortages of single-modal signals and outperforming existing models in CAD detection. This study highlights the potential of multi-modal deep-coding information, and offers a wider insight to enhance CAD diagnosis.

Atrial fibrillation (AF) is the most prevalent form of arrhythmia, with a rising incidence and prevalence worldwide, posing significant implications for public health. In this paper, we introduce an approach that combines the Recurrence Plot (RP) technique and the ResNet architecture to predict AF. Our method involves three main steps: using wavelet filtering to remove noise interference; generating RPs through phase space reconstruction; and employing a multi-level chained residual network for AF prediction. To validate our approach, we established a comprehensive database consisting of electrocardiogram (ECG) recordings from 1008 AF patients and 48,292 Non-AF patients, with a total of 2067 and 93,129 ECGs, respectively. The experimental results demonstrated high levels of prediction precision (90.5%), recall (89.1%), F1 score (89.8%), accuracy (93.4%), and AUC (96%) on our dataset. Moreover, when tested on a publicly available AF dataset (AFPDB), our method achieved even higher prediction precision (94.8%), recall (99.4%), F1 score (97.0%), accuracy (97.0%), and AUC (99.7%). These findings suggest that our proposed method can effectively extract subtle information from ECG signals, leading to highly accurate AF predictions.

Friction and wear phenomenon is a complex nonlinear system, and it is also a significant problem in the process of metal cutting. In order to systematically analyze the friction and wear process of tool material-workpiece material friction pair in the cutting process of high hardness alloy steel under different lubrication conditions, the chaotic characteristics of friction process between high hardness alloy steel and cemented carbide under the lubrication C60 nano-particles fluid are studied based on the chaos theory. Firstly, the friction and wear experiments of the friction pair between high hardness alloy steel and cemented carbide tool are carried out based on the ring-block friction and wear tester, and the results of friction force signal in time domain and wear width are obtained. Then, the friction signals in time domain are processed and transformed based on phase space reconstruction and recurrence plot theory, and the recurrence plots of different experimental groups under different lubrication conditions are generated. The evolution law of recurrence plot is further observed and studied, and the recursive quantitative index is analyzed. Finally, the cutting experiments of tool wear are carried out. The results show that the proposed method can intuitively and accurately reveal the wear evolution process and the wear feature identification law of the tool material-high hardness alloy steel pair under different lubrication conditions. Meanwhile, it is found that when the concentration of C60 nanoparticles is 200~300 ppm, the stability of the friction pair system is best. The proposed method can provide a strategy for wear prediction in cutting process, and provide a theoretical basis and technical support for antifriction lubrication methods in practical cutting applications. [ABSTRACT FROM AUTHOR]

This paper proposes fault diagnosis methods aimed at proactively preventing potential safety issues in robot systems, particularly human coexistence robots (HCRs) used in industrial environments. The data were collected from durability tests of the driving module for HCRs, gathering time-series vibration data until the module failed. In this study, to apply classification methods in the absence of post-failure data, the initial 50% of the collected data were designated as the normal section, and the data from the 10 h immediately preceding the failure were selected as the fault section. To generate additional data for the limited fault dataset, the Wasserstein generative adversarial networks with gradient penalty (WGAN-GP) model was utilized and residual connections were added to the generator to maintain the basic structure while preventing the loss of key features of the data. Considering that the performance of image encoding techniques varies depending on the dataset type, this study applied and compared five image encoding methods and four CNN models to facilitate the selection of the most suitable algorithm. The time-series data were converted into image data using image encoding techniques including recurrence plot, Gramian angular field, Markov transition field, spectrogram, and scalogram. These images were then applied to CNN models, including VGGNet, GoogleNet, ResNet, and DenseNet, to calculate the accuracy of fault diagnosis and compare the performance of each model. The experimental results demonstrated significant improvements in diagnostic accuracy when employing the WGAN-GP model to generate fault data, and among the image encoding techniques and convolutional neural network models, spectrogram and DenseNet exhibited superior performance, respectively. [ABSTRACT FROM AUTHOR]

Statistical process control (SPC) charts are commonly used to monitor quality characteristics in manufacturing processes. When monitoring two or more related quality characteristics simultaneously, multivariate T 2 control charts are often employed. Like univariate control charts, control chart pattern recognition (CCPR) plays a crucial role in multivariate SPC. The presence of non-random patterns in T 2 control charts indicates that a process is influenced by one or more assignable causes and that corrective actions should be taken. In this study, we developed a deep learning-based classification model for recognizing control chart patterns in multivariate processes. To address the problem of the insufficient representation of one-dimensional (1D) data, we explore the advantages of using two-dimensional (2D) image data obtained from a threshold-free recurrence plot. A multi-channel deep convolutional neural network (MCDCNN) model was developed to incorporate both 1D and 2D representations of control chart data. This model was tested on multivariate processes with different covariance matrices and compared with other traditional algorithms. Moreover, the effects of imbalanced datasets and dataset size on classification performance were analyzed. Simulation studies revealed that the developed MCDCNN model outperforms other techniques in identifying multivariate non-random patterns. For the most significant one, our proposed MCDCNN method achieved a 10% improvement over traditional methods. The overall results suggest that the developed MCDCNN model can be beneficial for intelligent SPC. [ABSTRACT FROM AUTHOR]

OBJECTIVE: Recently, older people’s cardiovascular systems have been affected by aging-related changes. An electrocardiogram (ECG) provides information about cardiac health. Analyzing ECG signals can help doctors and researchers diagnose many deaths. ‎Besides direct ECG analysis, some measurements can be extracted from the ECG signals, and ‎one of the most important measurements is heart rate variability (HRV). Research and clinical domains can benefit from HRV measurement and analysis as a potential noninvasive tool for assessing autonomic nervous system activity. The HRV describes the variation between an ECG signal’s RR intervals over time and the change in that interval over time. An individual’s heart rate (HR) is a non-stationary signal, and its variation can indicate a medical condition or impending cardiac disease. Many factors, such as stress, gender, disease, and age, influence HRV. METHODS: The data for this study is taken from a standard database, the Fantasia Database, which contains 40 subjects, including two groups of 20 young subjects (21‒34 years old) and 20 older subjects (68‒85 years old). We used two non-linear methods, Poincare and Recurrence Quantification Analysis (RQA), to determine how different age groups affect HRV using Matlab and Kubios software. RESULTS: By analyzing some features extracted from this non-linear method based on a mathematical model and making a comparison, the results indicate that the SD1, SD2, SD1/SD2, and area of an ellipse (S) in Poincare will be lower in old people than in young people, but %REC, %DET, Lmean and Lmax will recur more often in older people than in younger ones. Poincare Plot and RQA show opposite correlations with aging. In addition, Poincaré’s plot showed that young people have a greater range of changes than the elderly. CONCLUSION: According to the result of this study, heart rate changes can be reduced by aging, and ignoring this issue could lead to cardiovascular disease in the future (Tab. 3, Fig. 7, Ref. 55). [ABSTRACT FROM AUTHOR]

For the power system to be stable and reliable, power quality disturbances (PQDs) must be classified. In this work, deep learning was implemented for the purpose of categorizing PQDs. The transfer learning techniques such as ResNet-50, AlexNet, and GoogLeNet were compared and evaluated for the suitability of classifying PQD signals. Accuracy, classification probability, and explainability through GradCAM- an explainable AI technique was evaluated as a grading reference for the comparative analysis. Examination of the three criteria revealed ResNet-50 as the best among all the three architectures for classifying PQD signals since depending on the accuracy. [ABSTRACT FROM AUTHOR]

The paper presents a system for analyzing cardiac activity with the possibility of continuous and remote monitoring. The created sensor mobile device monitors heart activity by means of the convenient and imperceptible registration of cardiac signals. At the same time, the behavior of the human body is also monitored through the accelerometer and gyroscope built into the device, thanks to which it is possible to signal in the event of loss of consciousness or fall (in patients with syncope). Conducting real-time cardio monitoring and the analysis of recordings using various mathematical methods (linear, non-linear, and graphical) enables the research, accurate diagnosis, timely assistance, and correct treatment of cardiovascular diseases. The paper examines the recordings of patients diagnosed with arrhythmia and syncope recorded by electrocardiography (ECG) sensors in real conditions. The obtained results are subjected to statistical analysis to determine the accuracy and significance of the obtained results. The studies show significant deviations in the patients with arrhythmia and syncope regarding the obtained values of the studied parameters of heart rate variability (HRV) from the accepted normal values (for example, the root mean square of successive differences between normal heartbeats (RMSSD) in healthy individuals is 24.02 ms, while, in patients with arrhythmia (6.09 ms) and syncope (5.21 ms), it is much lower). The obtained quantitative and graphic results identify some possible abnormalities and demonstrate disorders regarding the activity of the autonomic nervous system, which is directly related to the work of the heart.

The goal of the research is to describe the aggregation process inside the mucilage produced by plant seeds using molecular dynamics (MD) combined with time series algorithmic analysis based on the recurrence plots. The studied biological molecules model is seed mucilage composed of three main polysaccharides, i.e. pectins, hemicellulose, and cellulose. The modeling of biological molecules is based on the assumption that a classical–quantum passage underlies the aggregation process in the mucilage, resulting from non-covalent interactions, as they affect the macroscopic properties of the system. The applied recurrence plot approach is an important tool for time series analysis and data mining dedicated to analyzing time series data originating from complex, chaotic systems. In the current research, we demonstrated that advanced algorithmic analysis of seed mucilage data can reveal some features of the dynamics of the system, namely temperature-dependent regions with different dynamics of increments of a number of hydrogen bonds and regions of stable oscillation of increments of a number of hydrophobic–polar interactions. Henceforth, we pave the path for automatic data-mining methods for the analysis of biological molecules with the intermediate step of the application of recurrence plot analysis, as the generalization of recurrence plot applications to other (biological molecules) datasets is straightforward.

The existence of chaos is particularly relevant, as the identification of a chaotic behavior in a time series could lead to reliable short-term forecasting. This paper evaluates the existence of nonlinearity and chaos in the underlying process of the spot prices of the Spanish electricity market. To this end, we used daily data spanning from 1 January 2013, to 31 March 2021 and we applied a comprehensive framework that encompassed a wide range of techniques. Nonlinearity was analyzed using the BDS method, while the existence of a chaotic structure was studied through Lyapunov exponents, recurrence plots, and quantitative recurrence analysis. While nonlinearity was detected in the underlying process, conclusive evidence supporting chaos was not found. In addition, the generalized autoregressive conditional heteroscedastic (GARCH) model accounts for part of the nonlinear structure that is unveiled in the electricity market. These findings hold substantial value for electricity market forecasters, traders, producers, and market regulators.

In recent years, the use of intelligent methods for automatic detection of sleep stages in medical applications to increase diagnostic accuracy and reduce the workload of physicians in analyzing sleep data by visual inspection is one of the important issues. The most important step for the automatic classification of sleep stages is the extraction of useful features. In this paper, an EEG-based algorithm for automatic detection of sleep stages is presented using features extracted from the recurrence plot and artificial neural network. Due to the non-stationary of the EEG signal, the recurrence plot was used in this paper for nonlinear analysis and extraction of signal features. Various extracted features have different numerical ranges. Normalization was performed to prevent the undesirable effects of large values of data. As all normalized features could not correctly classify different stages of sleep, effective features were selected. The results of this paper show the selected features and the Multi-Layer Perceptron (MLP) neural network able to achieve the values of 98.54 ± 1.88%, 99.03 ± 1.43%, and 98.32 ± 2.11%, respectively, for specificity, sensitivity, and accuracy between the two types of sleep, i.e., Non-Rapid Eye Movement (Non-REM) and Rapid Eye Movement (REM). Also, the results show that the selection of Pz-Oz channel compared to Fpz-Cz channel leads us to a higher percentage for the separation of stages I-IV, awake, while the separation of REM stage using Fpz-Cz channel is better. The results show that the proposed method has a higher success rate in classifying sleep stages than previous studies. The proposed method could well identify and distinguish all stages of sleep at an acceptable level. In addition to saving time, automatic analysis of sleep stages can help better and more accurate diagnosis and reduce physicians' workload in analyzing sleep data through visual inspection.

Estimation of actual evapotranspiration (ETa) is challenging due to its complex interactions with surrounding environments; thus, understanding the complexity and predictability of ETa along with its influencing factors is essential to improve ETa estimation. Based on the FLUXNET and ChinaFLUX datasets, we first examined whether daily ETa exhibited chaotic behaviors, and then investigated how daily ETa complexity and predictability varied across climate regimes and ecosystems using the chaos theory. The results of recurrence plot and correlation dimension (CD) analysis suggested the existence of chaotic behaviors in daily ETa, implying the system controlling ETa processes was deterministic with a limited number of controlling variables. The recurrence quantification analysis further revealed a varying degree of complexity for ETa processes (e.g., determinism‐DET with a range from 0.01% to 40.7%) across the sites. Specifically, daily ETa featured stronger deterministic properties (e.g., higher DET and lower CD values) and thus higher predictability in humid and arid regions; whereas, higher degrees of daily ETa complexity emerged in sub‐humid and semi‐arid regions due to more complex interactions with environmental factors. Moreover, primary environmental controls on daily ETa complexity varied with climatic and land surface conditions. Results showed that energy and water supplies along with average daily temperature and relative humidity were the primary controls on the ETa complexity across the sites, while leaf area index also played an important role at drier sites. As a first attempt, this study provides additional avenues to understand the complexity of ETa processes from a global perspective using the chaos theory. Key Points: Chaotic behaviors of daily actual evapotranspiration (ETa) across 81 sites were investigated using the chaos theoryHigher degrees of daily ETa complexity emerged at sub‐humid and semi‐arid sites than those at humid and arid sitesEnvironmental controls on daily ETa complexity varied with climatic and land surface conditions [ABSTRACT FROM AUTHOR]

The Recurrence Quantification Analysis (RQA), a pattern recognition-based time series analysis method, can be successfully utilized for short, nonstationary, nonlinear, and chaotic time series. These RQA measures quantify several properties of time series, including predictability, regularity, stability, randomness, and complexity. In this regard, first, we analyzed the intraday seasonality with RQA and demonstrated how RQA measures change among the intraday periods by using 160 million row matched orders of 100 shares from Borsa Istanbul Equity Market between 2019M10 and 2020M02. We selected 50 stocks from the BIST50 Index group and 50 stocks from outside of the BIST100 Index group. Since these two share groups exhibit similar intraday RQA seasonality, our results are robust. Second, we explained intraday volatility with RQA measures and found a relationship between RQA measures and intraday volatility using a regression model. [ABSTRACT FROM AUTHOR]