Supervised Machine-Learning Methodology for Industrial Robot Positional Health Using Artificial Neural Networks, Discrete Wavelet Transform, and Nonlinear Indicators.

Robotic systems are a fundamental part of modern industrial development. In this regard, they are required for long periods, in repetitive processes that must comply with strict tolerance ranges. Hence, the positional accuracy of the robots is critical, since degradation of this can represent a considerable loss of resources. In recent years, prognosis and health management (PHM) methodologies, based on machine and deep learning, have been applied to robots, in order to diagnose and detect faults and identify the degradation of robot positional accuracy, using external measurement systems, such as lasers and cameras; however, their implementation is complex in industrial environments. In this respect, this paper proposes a method based on discrete wavelet transform, nonlinear indices, principal component analysis, and artificial neural networks, in order to detect a positional deviation in robot joints, by analyzing the currents of the actuators. The results show that the proposed methodology allows classification of the robot positional degradation with an accuracy of 100%, using its current signals. The early detection of robot positional degradation, allows the implementation of PHM strategies on time, and prevents losses in manufacturing processes.

Analysis of Vibration Signals Based on Machine Learning for Crack Detection in a Low-Power Wind Turbine.

Currently, renewable energies, including wind energy, have been experiencing significant growth. Wind energy is transformed into electric energy through the use of wind turbines (WTs), which are located outdoors, making them susceptible to harsh weather conditions. These conditions can cause different types of damage to WTs, degrading their lifetime and efficiency, and, consequently, raising their operating costs. Therefore, condition monitoring and the detection of early damages are crucial. One of the failures that can occur in WTs is the occurrence of cracks in their blades. These cracks can lead to the further deterioration of the blade if they are not detected in time, resulting in increased repair costs. To effectively schedule maintenance, it is necessary not only to detect the presence of a crack, but also to assess its level of severity. This work studies the vibration signals caused by cracks in a WT blade, for which four conditions (healthy, light, intermediate, and severe cracks) are analyzed under three wind velocities. In general, as the proposed method is based on machine learning, the vibration signal analysis consists of three stages. Firstly, for feature extraction, statistical and harmonic indices are obtained; then, the one-way analysis of variance (ANOVA) is used for the feature selection stage; and, finally, the k-nearest neighbors algorithm is used for automatic classification. Neural networks, decision trees, and support vector machines are also used for comparison purposes. Promising results are obtained with an accuracy higher than 99.5%.

Short-Circuited Turn Fault Diagnosis in Transformers by Using Vibration Signals, Statistical Time Features, and Support Vector Machines on FPGA.

One of the most critical devices in an electrical system is the transformer. It is continuously under different electrical and mechanical stresses that can produce failures in its components and other electrical network devices. The short-circuited turns (SCTs) are a common winding failure. This type of fault has been widely studied in literature employing the vibration signals produced in the transformer. Although promising results have been obtained, it is not a trivial task if different severity levels and a common high-level noise are considered. This paper presents a methodology based on statistical time features (STFs) and support vector machines (SVM) to diagnose a transformer under several SCTs conditions. As STFs, 19 indicators from the transformer vibration signals are computed; then, the most discriminant features are selected using the Fisher score analysis, and the linear discriminant analysis is used for dimension reduction. Finally, a support vector machine classifier is employed to carry out the diagnosis in an automatic way. Once the methodology has been developed, it is implemented on a field-programmable gate array (FPGA) to provide a system-on-a-chip solution. A modified transformer capable of emulating different SCTs severities is employed to validate and test the methodology and its FPGA implementation. Results demonstrate the effectiveness of the proposal for diagnosing the transformer condition as an accuracy of 96.82% is obtained.

An Explainable Machine Learning Approach Based on Statistical Indexes and SVM for Stress Detection in Automobile Drivers Using Electromyographic Signals.

The economic and personal consequences that a car accident generates for society have been increasing in recent years. One of the causes that can generate a car accident is the stress level the driver has; consequently, the detection of stress events is a highly desirable task. In this article, the efficacy that statistical time features (STFs), such as root mean square, mean, variance, and standard deviation, among others, can reach in detecting stress events using electromyographical signals in drivers is investigated, since they can measure subtle changes that a signal can have. The obtained results show that the variance and standard deviation coupled with a support vector machine classifier with a cubic kernel are effective for detecting stress events where an AUC of 0.97 is reached. In this sense, since SVM has different kernels that can be trained, they are used to find out which one has the best efficacy using the STFs as feature inputs and a training strategy; thus, information about model explain ability can be determined. The explainability of the machine learning algorithm allows generating a deeper comprehension about the model efficacy and what model should be selected depending on the features used to its development.

ECG-Based Identification of Sudden Cardiac Death through Sparse Representations.

Sudden Cardiac Death (SCD) is an unexpected sudden death due to a loss of heart function and represents more than 50% of the deaths from cardiovascular diseases. Since cardiovascular problems change the features in the electrical signal of the heart, if significant changes are found with respect to a reference signal (healthy), then it is possible to indicate in advance a possible SCD occurrence. This work proposes SCD identification using Electrocardiogram (ECG) signals and a sparse representation technique. Moreover, the use of fixed feature ranking is avoided by considering a dictionary as a flexible set of features where each sparse representation could be seen as a dynamic feature extraction process. In this way, the involved features may differ within the dictionary's margin of similarity, which is better-suited to the large number of variations that an ECG signal contains. The experiments were carried out using the ECG signals from the MIT/BIH-SCDH and the MIT/BIH-NSR databases. The results show that it is possible to achieve a detection 30 min before the SCD event occurs, reaching an an accuracy of 95.3% under the common scheme, and 80.5% under the proposed multi-class scheme, thus being suitable for detecting a SCD episode in advance.

Natural Frequencies Identification by FEM Applied to a 2-DOF Planar Robot and Its Validation Using MUSIC Algorithm.

In this paper, the natural frequencies (NFs) identification by finite element method (FEM) is applied to a two degrees-of-freedom (2-DOF) planar robot, and its validation through a novel experimental methodology, the Multiple Signal Classification (MUSIC) algorithm, is presented. The experimental platforms are two different 2-DOF planar robots with different materials for the links and different types of actuators. The FEM is carried out using ANSYS™ software for the experiments, with vibration signal analysis by MUSIC algorithm. The advantages of the MUSIC algorithm against the commonly used fast Fourier transform (FFT) method are also presented for a synthetic signal contaminated by three different noise levels. The analytical and experimental results show that the proposed methodology identifies the NFs of a high-resolution robot even when they are very closed and when the signal is embedded in high-level noise. Furthermore, the results show that the proposed methodology can obtain a high-frequency resolution with a short sample data set. Identifying the NFs of robots is useful for avoiding such frequencies in the path planning and in the selection of controller gains that establish the bandwidth.

Convolutional Neural Network and Motor Current Signature Analysis during the Transient State for Detection of Broken Rotor Bars in Induction Motors.

Although induction motors (IMs) are robust and reliable electrical machines, they can suffer different faults due to usual operating conditions such as abrupt changes in the mechanical load, voltage, and current power quality problems, as well as due to extended operating conditions. In the literature, different faults have been investigated; however, the broken rotor bar has become one of the most studied faults since the IM can operate with apparent normality but the consequences can be catastrophic if the fault is not detected in low-severity stages. In this work, a methodology based on convolutional neural networks (CNNs) for automatic detection of broken rotor bars by considering different severity levels is proposed. To exploit the capabilities of CNNs to carry out automatic image classification, the short-time Fourier transform-based time-frequency plane and the motor current signature analysis (MCSA) approach for current signals in the transient state are first used. In the experimentation, four IM conditions were considered: half-broken rotor bar, one broken rotor bar, two broken rotor bars, and a healthy rotor. The results demonstrate the effectiveness of the proposal, achieving 100% of accuracy in the diagnosis task for all the study cases.

A New Methodology Based on EMD and Nonlinear Measurements for Sudden Cardiac Death Detection.

Heart diseases are among the most common death causes in the population. Particularly, sudden cardiac death (SCD) is the cause of 10% of the deaths around the world. For this reason, it is necessary to develop new methodologies that can predict this event in the earliest possible stage. This work presents a novel methodology to predict when a person can develop an SCD episode before it occurs. It is based on the adroit combination of the empirical mode decomposition, nonlinear measurements, such as the Higuchi fractal and permutation entropy, and a neural network. The obtained results show that the proposed methodology is capable of detecting an SCD episode 25 min before it appears with a 94% accuracy. The main benefits of the proposal are: (1) an improved detection time of 25% compared with previously published works, (2) moderate computational complexity since only two features are used, and (3) it uses the raw ECG without any preprocessing stage, unlike recent previous works.

A Novel Wavelet Transform-Homogeneity Model for Sudden Cardiac Death Prediction Using ECG Signals.

Sudden cardiac death (SCD) is one of the main causes of death among people. A new methodology is presented for predicting the SCD based on ECG signals employing the wavelet packet transform (WPT), a signal processing technique, homogeneity index (HI), a nonlinear measurement for time series signals, and the Enhanced Probabilistic Neural Network classification algorithm. The effectiveness and usefulness of the proposed method is evaluated using a database of measured ECG data acquired from 20 SCD and 18 normal patients. The proposed methodology presents the following significant advantages: (1) compared with previous works, the proposed methodology achieves a higher accuracy using a single nonlinear feature, HI, thus requiring low computational resource for predicting an SCD onset in real-time, unlike other methodologies proposed in the literature where a large number of nonlinear features are used to predict an SCD event; (2) it is capable of predicting the risk of developing an SCD event up to 20 min prior to the onset with a high accuracy of 95.8%, superseding the prior 12 min prediction time reported recently, and (3) it uses the ECG signal directly without the need for transforming the signal to a heart rate variability signal, thus saving time in the processing.

EEMD-MUSIC-based analysis for natural frequencies identification of structures using artificial and natural excitations.

This paper presents a new EEMD-MUSIC- (ensemble empirical mode decomposition-multiple signal classification-) based methodology to identify modal frequencies in structures ranging from free and ambient vibration signals produced by artificial and natural excitations and also considering several factors as nonstationary effects, close modal frequencies, and noisy environments, which are common situations where several techniques reported in literature fail. The EEMD and MUSIC methods are used to decompose the vibration signal into a set of IMFs (intrinsic mode functions) and to identify the natural frequencies of a structure, respectively. The effectiveness of the proposed methodology has been validated and tested with synthetic signals and under real operating conditions. The experiments are focused on extracting the natural frequencies of a truss-type scaled structure and of a bridge used for both highway traffic and pedestrians. Results show the proposed methodology as a suitable solution for natural frequencies identification of structures from free and ambient vibration signals.

A New dispersion entropy and fuzzy logic system methodology for automated classification of dementia stages using electroencephalograms.

A new EEG-based methodology is presented for differential diagnosis of the Alzheimer's disease (AD), Mild Cognitive Impairment (MCI), and healthy subjects employing the discrete wavelet transform (DWT), dispersion entropy index (DEI), a recently-proposed nonlinear measurement, and a fuzzy logic-based classification algorithm. The effectiveness and usefulness of the proposed methodology are evaluated by employing a database of measured EEG data acquired from 135 subjects, 45 MCI, 45 AD and 45 healthy subjects. The proposed methodology differentiates MCI and AD patients from HC subjects with an accuracy of 82.6-86.9%, sensitivity of 91 %, and specificity of 87 %.

A novel methodology for automated differential diagnosis of mild cognitive impairment and the Alzheimer's disease using EEG signals.

BACKGROUND: EEG signals obtained from Mild Cognitive Impairment (MCI) and the Alzheimer's disease (AD) patients are visually indistinguishable. NEW METHOD: A new methodology is presented for differential diagnosis of MCI and the AD through adroit integration of a new signal processing technique, the integrated multiple signal classification and empirical wavelet transform (MUSIC-EWT), different nonlinear features such as fractality dimension (FD) from the chaos theory, and a classification algorithm, the enhanced probabilistic neural network model of Ahmadlou and Adeli using the EEG signals. RESULTS: Three different FD measures are investigated: Box dimension (BD), Higuchi's FD (HFD), and Katz's FD (KFD) along with another measure of the self-similarities of the signals known as the Hurst exponent (HE). The accuracy of the proposed method was verified using the monitored EEG signals from 37 MCI and 37 AD patients. COMPARISON WITH EXISTING METHODS: The proposed method is compared with other methodologies presented in the literature recently. CONCLUSIONS: It was demonstrated that the proposed method, MUSIC-EWT algorithm combined with nonlinear features BD and HE, and the EPNN classifier can be employed for differential diagnosis of MCI and AD patients with an accuracy of 90.3%.

A new methodology for automated diagnosis of mild cognitive impairment (MCI) using magnetoencephalography (MEG).

Mild cognitive impairment (MCI) is a cognitive disorder characterized by memory impairment, greater than expected by age. A new methodology is presented to identify MCI patients during a working memory task using MEG signals. The methodology consists of four steps: In step 1, the complete ensemble empirical mode decomposition (CEEMD) is used to decompose the MEG signal into a set of adaptive sub-bands according to its contained frequency information. In step 2, a nonlinear dynamics measure based on permutation entropy (PE) analysis is employed to analyze the sub-bands and detect features to be used for MCI detection. In step 3, an analysis of variation (ANOVA) is used for feature selection. In step 4, the enhanced probabilistic neural network (EPNN) classifier is applied to the selected features to distinguish between MCI and healthy patients. The usefulness and effectiveness of the proposed methodology are validated using the sensed MEG data obtained experimentally from 18 MCI and 19 control patients.