Linear Machine Research Articles

Orbitofrontal Cortex Functional Connectivity-Based Classification for Chronic Insomnia Disorder Patients With Depression Symptoms.

Depression is a common comorbid symptom in patients with chronic insomnia disorder (CID). Previous neuroimaging studies found that the orbital frontal cortex (OFC) might be the core brain region linking insomnia and depression. Here, we used a machine learning approach to differentiate CID patients with depressive symptoms from CID patients without depressive symptoms based on OFC functional connectivity. Seventy patients with CID were recruited and subdivided into CID with high depressive symptom (CID-HD) and low depressive symptom (CID-LD) groups. The OFC functional connectivity (FC) network was constructed using the altered structure of the OFC region as a seed. A linear kernel SVM-based machine learning approach was carried out to classify the CID-HD and CID-LD groups based on OFC FC features. The predict model was further verified in a new cohort of CID group (n = 68). The classification model based on the OFC FC pattern showed a total accuracy of 76.92% (p = 0.0009). The area under the receiver operating characteristic curve of the classification model was 0.84. The OFC functional connectivity with reward network, salience network and default mode network contributed the highest weights to the prediction model. These results were further validated in an independent CID group with high and low depressive symptom (accuracy = 67.9%). These findings provide a potential biomarker for early diagnosis and intervention in CID patients comorbid with depression based on an OFC FC-based machine learning approach.

Optimal Reference Primary Flux Based Model Predictive Control of Linear Induction Machine With MTPA and Field-Weakening Operations for Urban Transit

The linear induction machine (LIM) in urban transit suffers from poor efficiency and thrust decay due to the large air-gap length, longitudinal end effect, and the light load operating conditions. Maximum thrust per ampere (MTPA) is one of the effective methods to improve the motor efficiency by minimizing the conduction loss. However, when MTPA is introduced in model predictive control as an additional constraint, traditional methods inevitably need to add more weighting factors in the cost function. To further improve the performance of LIM drives and eliminate the weighting factors, this article proposes an optimal reference primary flux based model predictive control (ORPF-MPC) for LIM. The ORPF containing MTPA condition is derived and introduced as the only criterion in cost function to achieve the control goal without using any weighting factors. Furthermore, the ORPF is extended to the field-weakening operation, which significantly widens the operating range and nearly achieves the optimal thrust capability to mitigate the thrust decay of LIM. Finally, the proposed ORPF-MPC is implemented with simplified arbitrary double vectors to remove the restriction on the selection of voltage vector pairs and further reduce the flux and thrust ripples. The proposed method is comprehensively investigated on one 3 kW arc induction machine, and its effectiveness is fully validated through both simulation and experimental results.

Assessment of Electrical Power Generation of Wave Energy Converters With Wave-to-Wire Modeling

Direct-drive wave energy converter (WEC) and buoy control algorithms have shown great potential for renewable wave energy extraction in ideal conditions. However the actual power take-off (PTO) impacts are barely considered in the WEC design. This paper highlights the demands of designing the WEC wave-to-wire control from a global point of view by studying the actual PTO impacts. A permanent magnet linear electrical machine (LEM) PTO unit is simulated and controlled to fulfill the WEC buoy control requirements. Several state of the art control algorithms, which include singular-arc (SA) control, shape-based (SB) control, model predictive control (MPC), and proportional-derivative (PD) control, are applied to maximize the wave energy production (mechanical energy). Multiple types of electrical PTOs, including ideal PTO, unlimited PTO and limited PTO, are all implemented to evaluate WEC wave-to-wire performances. Further, the PTO copper loss model and the PTO actual efficiency maps are introduced and studied to improve the electrical PTO operation efficiency. To further assess the control schemes in various wave conditions, one-year PacWave ground-truth data is applied as well. Numerical simulations are conducted using MATLAB/Simulink and the Simscape toolbox. The electrical PTO unit is composed of a LEM, an ideal inverter, and an ideal energy storage system. The results show that the actual PTO will impact the constrained controls (MPC and SB) less comapring to unconstrained controls (SA and PD). Although SB can produce the maximum energy with the limited PTOs, it is not robust for all wave conditions. At the end of the paper, the possible solutions for improving the WEC wave-to-wire performances are also provided.

A Phase-Domain Model of Dual Three-Phase Segmented Powered Linear PMSM for Hardware-Assisted Real-Time Simulation

In this article, a hardware-assisted real-time simulation phase-domain (PD) model of a dual three-phase segmented powered linear permanent magnet synchronous machine (SP-LPMSM) used in electromagnetic drive is established. The model is able to consider the end-effect, saturation effect, coupling effect, and open-phase fault conditions, which are specially occurs in the segmented structure in SP-LPMSM. The unbalanced inductances and saturation effect caused by the segmented structure are investigated and a simplified look-up table for inductances is introduced. Additionally, to accurately evaluate the coupling effect, the permanent magnet (PM) flux linkage is decoupled into the product of unit PM flux linkage and coupling coefficient. In order to avoid the numerical impulse caused by the derivative in the progress of calculating back-EMF, sigmoid function is adopted to express the coupling coefficient. Meanwhile, to make the PD-model compatible with the open-circuit fault conditions, incidence matrices and current constraint matrices are introduced to unify the PD-models with different winding connection types and open circuit faults into one general form. The PD-model realized in field-programmable gate arrays (FPGA) based hardware is validated in comparison with finite-element analysis results.

The second dimension of spatial association

• The concept of the second dimension of spatial associations (SDA) is proposed. • SDA can extract in-depth and multi-scale geoinformation outside sample locations. • SDA-based statistical, machine learning, and geostatistical models are developed. • Spatial prediction accuracy is critically improved by SDA models compared with FDA. • SDA models provide smooth, effective, and low-uncertainty spatial predictions. A reasonable and adequate understanding of spatial association between geographical variables is the basis of spatial statistical inference and geocomputation, such as spatial prediction. Most of the current models for exploring spatial association of variables are constructed using data at sample locations. In this study, approaches for exploring spatial association using observations at sample locations are defined as the first dimension of spatial association (FDA). However, geographical information outside sample locations is usually missing in current models. To address this issue, this study proposes the concept of the second dimension of spatial association (SDA), which is an approach that extracts geographical information at locations outside samples for exploring spatial association. Based on the concept of SDA, three SDA models, including SDA-based multivariate linear regression, machine learning (i.e., random forest), and geostatistical models (i.e., random forest kriging), are developed for examining spatial association and predicting spatial distributions of trace elements, including Cr and Cu, in a mining region in Western Australia. Model accuracy is evaluated by comparing with corresponding FDA models. A new R package “SecDim” is developed to conduct SDA models. Results show that SDA models have a series of advantages in examining spatial association compared with FDA models. First, the accuracy of spatial prediction can be critically improved by SDA compared with the FDA, although identical explanatory variables and models are used for the modeling and prediction. Second, SDA can effectively indicate the multi-scale effects and diverse information within explanatory variables of the geographical environment at local ranges using the second dimension variables. Third, SDA can avoid underestimating high values and overestimating low values in the general FDA-based statistical models, machine learning, and geostatistical models. Finally, SDA models provide more smooth spatial predictions across space than that predicted by FDA models and avoid massive fluctuations at local ranges. The concept of SDA provides new insight into geographical information-based spatial association. SDA and multiple types of SDA models have great potential for more accurate and effective spatial statistical inference and geocomputation in diverse fields.

Cogging Force Optimization of Double-Sided Tubular Linear Machine With Tooth-Cutting

The double-sided tubular linear machine (DSTLM) has a more compact structure than general single-sided motors, resulting in a higher power density, typically at the cost of a higher cogging force. To reduce this aspect and make it more suitable for applications in wave energy conversion systems or elevator driver systems, a new tooth-cutting topology is proposed in this article. The Schwarz–Christoffel mapping method is used to model and analyze the magnetic field and cogging force of the DSTLM. Based on the analytical model and the relationships between the cutting dimensions and the force performances, optimization is carried out by using a differential evolution algorithm. The optimization and the experiment results show that the proposed tooth-cutting topology can effectively reduce the cogging force at the small cost of weakening average thrust.

Control System Upgradation with Linear Actuator in Coil Unloading Machine

Abstract: In this paper, we are going to see about the project is to unload the coil for the dispatching process from metal plate to place in thermo coal with the help of linear actuator. The process of coil unloading is automated with the help of HMI, sensors, motors and linear actuator. This process is carried out through ladder logic methodology. The linear actuator application machine process is the machine based on the PLC (PROGRAMMABLE LOGIC CONTROLLER) controller. Earlier PLC controllers are huge programmed size, spare parts are rare, inbuilt programming that could not change as per the requirements and the absolute models which is difficult to learn. The cam shaft mechanism takes over time to assemble the coil position. And also coils pick up and drop step by step process is difficult.To overcome this problem used the advanced PLC controllers model SIEMENS-S7 1200 instead of earlier telemechanic PLC controller because in this SIEMENS-S7 1200 model PLC the HMI is used to display the coil positions.

Insights to estimate exposure to regulated and non-regulated disinfection by-products in drinking water

BackgroundKnowledge about human exposure and health effects associated with non-routinely monitored disinfection by-products (DBPs) in drinking water is sparse.ObjectiveTo provide insights to estimate exposure to regulated and non-regulated DBPs in drinking water.MethodsWe collected tap water from homes (N = 42), bottled water (N = 10), filtered tap water with domestic activated carbon jars (N = 6) and reverse osmosis (N = 5), and urine (N = 39) samples of participants from Barcelona, Spain. We analyzed 11 haloacetic acids (HAAs), 4 trihalomethanes (THMs), 4 haloacetonitriles (HANs), 2 haloketones, chlorate, chlorite, and trichloronitromethane in water and HAAs in urine samples. Personal information on water intake and socio-demographics was ascertained in the study population (N = 39) through questionnaires. Statistical models were developed based on THMs as explanatory variables using multivariate linear regression and machine learning techniques to predict non-regulated DBPs.ResultsChlorate, THMs, HAAs, and HANs were quantified in 98–100% tap water samples with median concentration of 214, 42, 18, and 3.2 μg/L, respectively. Multivariate linear regression models had similar or higher goodness of fit (R2) compared to machine learning models. Multivariate linear models for dichloro-, trichloro-, and bromodichloroacetic acid, dichloroacetonitrile, bromochloroacetonitrile, dibromoacetonitrile, trichloropropnanone, and chlorite showed good predictive ability (R2 = 0.8–0.9) as 80–90% of total variance could be explained by THM concentrations. Activated carbon filters reduced DBP concentrations to a variable extent (27–80%), and reverse osmosis reduced DBP concentrations ≥98%. Only chlorate was detected in bottled water samples (N = 3), with median = 13.0 µg/L. Creatinine-adjusted trichloroacetic acid was the most frequently detected HAA in urine samples (69.2%), and moderately correlated with estimated drinking water intake (r = 0.48).SignificanceFindings provide valuable insights for DBP exposure assessment in epidemiological studies. Validation of predictive models in a larger number of samples and replication in different settings is warranted.Impact statementOur study focused on assessing and describing the occurrence of several classes of DBPs in drinking water and developing exposure models of good predictive ability for non-regulated DBPs.

Comparing Multiple Linear Regression and Machine Learning in Predicting Diabetic Urine Albumin-Creatinine Ratio in a 4-Year Follow-Up Study.

The urine albumin–creatinine ratio (uACR) is a warning for the deterioration of renal function in type 2 diabetes (T2D). The early detection of ACR has become an important issue. Multiple linear regression (MLR) has traditionally been used to explore the relationships between risk factors and endpoints. Recently, machine learning (ML) methods have been widely applied in medicine. In the present study, four ML methods were used to predict the uACR in a T2D cohort. We hypothesized that (1) ML outperforms traditional MLR and (2) different ranks of the importance of the risk factors will be obtained. A total of 1147 patients with T2D were followed up for four years. MLR, classification and regression tree, random forest, stochastic gradient boosting, and eXtreme gradient boosting methods were used. Our findings show that the prediction errors of the ML methods are smaller than those of MLR, which indicates that ML is more accurate. The first six most important factors were baseline creatinine level, systolic and diastolic blood pressure, glycated hemoglobin, and fasting plasma glucose. In conclusion, ML might be more accurate in predicting uACR in a T2D cohort than the traditional MLR, and the baseline creatinine level is the most important predictor, which is followed by systolic and diastolic blood pressure, glycated hemoglobin, and fasting plasma glucose in Chinese patients with T2D.

Predicting Depression in Adolescents Using Mobile and Wearable Sensors: Multimodal Machine Learning-Based Exploratory Study.

BackgroundDepression levels in adolescents have trended upward over the past several years. According to a 2020 survey by the National Survey on Drug Use and Health, 4.1 million US adolescents have experienced at least one major depressive episode. This number constitutes approximately 16% of adolescents aged 12 to 17 years. However, only 32.3% of adolescents received some form of specialized or nonspecialized treatment. Identifying worsening symptoms earlier using mobile and wearable sensors may lead to earlier intervention. Most studies on predicting depression using sensor-based data are geared toward the adult population. Very few studies look into predicting depression in adolescents.ObjectiveThe aim of our work was to study passively sensed data from adolescents with depression and investigate the predictive capabilities of 2 machine learning approaches to predict depression scores and change in depression levels in adolescents. This work also provided an in-depth analysis of sensor features that serve as key indicators of change in depressive symptoms and the effect of variation of data samples on model accuracy levels.MethodsThis study included 55 adolescents with symptoms of depression aged 12 to 17 years. Each participant was passively monitored through smartphone sensors and Fitbit wearable devices for 24 weeks. Passive sensors collected call, conversation, location, and heart rate information daily. Following data preprocessing, 67% (37/55) of the participants in the aggregated data set were analyzed. Weekly Patient Health Questionnaire-9 surveys answered by participants served as the ground truth. We applied regression-based approaches to predict the Patient Health Questionnaire-9 depression score and change in depression severity. These approaches were consolidated using universal and personalized modeling strategies. The universal strategies consisted of Leave One Participant Out and Leave Week X Out. The personalized strategy models were based on Accumulated Weeks and Leave One Week One User Instance Out. Linear and nonlinear machine learning algorithms were trained to model the data.ResultsWe observed that personalized approaches performed better on adolescent depression prediction compared with universal approaches. The best models were able to predict depression score and weekly change in depression level with root mean squared errors of 2.83 and 3.21, respectively, following the Accumulated Weeks personalized modeling strategy. Our feature importance investigation showed that the contribution of screen-, call-, and location-based features influenced optimal models and were predictive of adolescent depression.ConclusionsThis study provides insight into the feasibility of using passively sensed data for predicting adolescent depression. We demonstrated prediction capabilities in terms of depression score and change in depression level. The prediction results revealed that personalized models performed better on adolescents than universal approaches. Feature importance provided a better understanding of depression and sensor data. Our findings can help in the development of advanced adolescent depression predictions.

Correlation between air pollution and prevalence of conjunctivitis in South Korea using analysis of public big data

This study investigated how changes in weather factors affect the prevalence of conjunctivitis using public big data in South Korea. A total of 1,428 public big data entries from January 2013 to December 2019 were collected. Disease data and basic climate/air pollutant concentration records were collected from nationally provided big data. Meteorological factors affecting eye diseases were identified using multiple linear regression and machine learning analysis methods such as extreme gradient boosting (XGBoost), decision tree, and random forest. The prediction model with the best performance was XGBoost (1.180), followed by multiple regression (1.195), random forest (1.206), and decision tree (1.544) when using root mean square error (RMSE) values. With the XGBoost model, province was the most important variable (0.352), followed by month (0.289) and carbon monoxide exposure (0.133). Other air pollutants including sulfur dioxide, PM10, nitrogen dioxides, and ozone showed low associations with conjunctivitis. We identified factors associated with conjunctivitis using traditional multiple regression analysis and machine learning techniques. Regional factors were important for the prevalence of conjunctivitis as well as the atmosphere and air quality factors.

Metabolomic Changes Upon Conjugated Linoleic Acid Supplementation and Predictions of Body Composition Responsiveness.

Conjugated linoleic acid (CLA) may optimize body composition, yet mechanisms underlining its benefits are not clear in humans. We aimed to reveal the CLA-induced changes in the plasma metabolome associated with body composition improvement and the predictive performance of baseline metabolome on intervention responsiveness. Plasma metabolome from overnight fasted samples at pre- and post-intervention of 65 participants in a 12-week randomized, placebo-controlled trial (3.2 g/day CLA vs 3.2 g/day sunflower oil) were analyzed using untargeted LC-MS metabolomics. Mixed linear model and machine learning were applied to assess differential metabolites between treatments, and to identify optimal panel (based on baseline conventional variables vs metabolites) predicting responders of CLA-derived body composition improvement (increased muscle variables or decreased adiposity variables) based on dual-energy x-ray absorptiometry. Compared with placebo, CLA altered 57 metabolites (P < 0.10) enriched in lipids/lipid-like molecules including glycerophospholipids (n = 7), fatty acyls (n = 6), and sphingolipids (n = 3). CLA-upregulated cholic acid (or downregulated aminopyrrolnitrin) was inversely correlated with changes in muscle and adiposity variables. Inter-individual variability in response to CLA-derived body composition change. The areas under the curves of optimal metabolite panels were higher than those of optimal conventional panels in predicting favorable response of waist circumference (0.93 [0.82-1.00] vs 0.64 [0.43-0.85]), visceral adiposity index (0.95 [0.88-1.00] vs 0.58 [0.35-0.80]), total fat mass (0.94 [0.86-1.00] vs 0.69 [0.51-0.88]) and appendicular fat mass (0.97 [0.92-1.00] vs 0.73 [0.55-0.91]) upon CLA supplementation (all FDR P < 0.05). Post-intervention metabolite alterations were identified, involving in lipid/energy metabolism, associated with body composition changes. Baseline metabolite profiling enhanced the prediction accuracy for responsiveness of CLA-induced body composition benefits.

Efficiency of the Adjusted Binary Classification (ABC) Approach in Osteometric Sex Estimation: A Comparative Study of Different Linear Machine Learning Algorithms and Training Sample Sizes.

Simple SummaryThis study adopts a dynamic methodology to explore challenges to the practical application of the adjusted binary classification (ABC) approach, which are related to the unmodifiable characteristics of data used in its development, such as intrasexual variation (sexual dimorphism) of variables and methodological factors such as the selected classification algorithm and sample size. The adequacy of a training dataset’s size was judged relative to the classification performance in an independent test set. Finding an optimal classifier was also addressed in this study, wherein the results demonstrate that both statistical modeling and machine learning techniques perform almost equally in the univariate models; however, differences are evident in the multivariate model due to the different number of variables included via the feature selection process, as well as the effect of inadequate training sample size relative to the test set. This approach is particularly useful when quick classification/prediction is required for making real-time forensic decisions.The adjusted binary classification (ABC) approach was proposed to assure that the binary classification model reaches a particular accuracy level. The present study evaluated the ABC for osteometric sex classification using multiple machine learning (ML) techniques: linear discriminant analysis (LDA), boosted generalized linear model (GLMB), support vector machine (SVM), and logistic regression (LR). We used 13 femoral measurements of 300 individuals from a modern Turkish population sample and split data into two sets: training (n = 240) and testing (n = 60). Then, the five best-performing measurements were selected for training univariate models, while pools of these variables were used for the multivariable models. ML classifier type did not affect the performance of unadjusted models. The accuracy of univariate models was 82–87%, while that of multivariate models was 89–90%. After applying ABC to the crossvalidation set, the accuracy and the positive and negative predictive values for uni- and multivariate models were ≥95%. Sex could be estimated for 28–75% of individuals using univariate models but with an obvious sexing bias, likely caused by different degrees of sexual dimorphism and between-group overlap. However, using multivariate models, we minimized the bias and properly classified 81–87% of individuals. A similar performance was also noted in the testing sample (except for FEB), with accuracies of 96–100%, and a proportion of classified individuals between 30% and 82% in univariate models, and between 90% and 91% in multivariate models. When considering different training sample sizes, we demonstrated that LR was the most sensitive with limited sample sizes (n < 150), while GLMB was the most stable classifier.

Abstract 1942: Interpretation of machine learning methods for the prediction of breast and prostate cancer risk

Abstract Neural networks are powerful tools for modeling genetic contributions to cancer risk that can theoretically capture nonlinear, epistatic interactions between disease-associated loci in heritable cancers. Recent studies applying neural networks and other machine learning methods to complex trait risk prediction from single nucleotide polymorphism (SNP) array data have shown promise in improving risk stratification. However, current performance gains for neural networks when compared to traditional polygenic risk scoring (PRS) approaches and other nonlinear and linear machine learning methods have been modest. Moreover, there remains substantial debate as to the effect of capturing epistatic interactions between SNPs in risk modeling. Central to the debate has been the difficulty in interpreting the complex, nonlinear mapping learned by neural networks and other nonlinear modeling approaches. To decipher the importance of capturing nonlinear interactions in cancer risk modeling, we first applied several PRS approaches to the prediction of breast and prostate cancer status for individuals in the Discovery, Biology, and Risk of Inherited Variants in Breast Cancer (DRIVE) and Elucidating Loci Involved in Prostate Cancer Susceptibility (ELLIPSE) datasets respectively. Consistent with previous studies of complex disease prediction using machine learning, we noted greater predictive capability upon inclusion of more loci in our modeling but only small performance gains when nonlinearity was captured in both cancer types. We then applied machine learning interpretation methods to derive a score for each variant per method, including several neural network interpretation methods which, to our knowledge, have not been applied in this context. We noted varying degrees of concordance between the scores assigned by each method. Finally, we performed pairwise in silico perturbations on salient features and techniques from network biology to identify epistatic interactions between loci captured by each model. Our work represents a comprehensive study of methods for inferring both variant level and epistatic interaction contributions to cancer risk. Citation Format: Adam Klie, James Talwar, Meghana Pagadala, Hannah Carter. Interpretation of machine learning methods for the prediction of breast and prostate cancer risk [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1942.

Development of Surface Properties of High-Carbon Steel by Pulse Plasma Technique

The surface of DIN 115CrV3 (high carbon steel) was chosen for the plasma pulse process. Different process parameters such as distance and number of pulses were tested. The effects of pulse plasma treatment on the morphology, phase, and crystallographic orientation of DIN 115CrV3 steel were analyzed. The cross section of modified surfaces was examined by optical microscope and scanning electron microscopy (SEM). The structure in the steel surface layer changes to nitride and an austenite phase is formed. The formed structure is composed of columnar and fine grains. X-ray diffraction (XRD) analysis was performed on the modified surface. The existence of complex phases and hard phases was shown in the modification layer. The microhardness was investigated before and after treatment using a microhardness tester. The hardness of treated steels was increased from 250 HV to 1600 HV after pulse plasma treatment. Wear testing of the modifed surface used a CSM linear wear test machine. The reciprocating wear method was preferred. Wear rate was calculated using CSM wear software. The friction coefficient and wear values were varied in accordance with parameters. The friction coefficient of treated samples was decreased, while the wear rate increased to four times than that of untreated samples. The mechanical properties of the surface, such as hardness and wear resistance, were improved after pulse plasma treatment. The SEM–energy-dispersive spectroscopy (EDS) analyses, atomic force microscopy (AFM), and three-dimensional (3D) profilometer techniques were used to investigate the worn surfaces of specimens. The surface of the treated sample was shown to experience less deformation during the wear test.

Hyp-Net: Automated detection of hypertension using deep convolutional neural network and Gabor transform techniques with ballistocardiogram signals

Hypertension (HPT) is a physiological abnormality characterized by high blood pressure, headache, wooziness, and fainting that may lead to various heart, kidney, or brain diseases. Detection and continuous monitoring of HPT by sphygmometer is arduous and hectic. Nowadays, ballistocardiogram (BCG) signals are used to determine HPT as it indicates the vibration of the heart. Usual linear and nonlinear hand-crafted machine learning methods are subjective, involve decomposition of signal, features elicitation, selection, and classification steps. In this work, a completely automated HPT detection system is proposed using time–frequency (T-F) spectral images and a convolutional neural network (CNN) for the accurate detection of HPT using BCG signals. The BCG signals are subjected to Gabor transform (GT), smoothed pseudo-Wigner Ville distribution (SPWVD), and short-time Fourier transform (STFT) techniques to obtain T-F spectral images. These T-F spectral images are fed to benchmarked Alex-Net, Res-Net50, and proposed CNN (Hyp-Net) to develop the automated HPT detection model with a 10-fold cross-validation scheme. The proposed Hyp-Net obtained the highest detection accuracy of 97.65% with GT-based spectral images. In comparison to Alex-Net and Res-Net50 pre-trained models, our developed Hyp-Net needs minimal learnable parameters, makes it computationally fast and more efficient. This shows that our proposed model has outperformed other two transfer learning methods. The experimental results with collected BCG signals from a public dataset are provided to show the effectiveness of the presented technique for automated detection of HPT.

DESIGN AND IMPLEMENTATION OF A PORTABLE AUTOMATED AUDIOMETER FOR HEARING CLASSIFICATION USING MACHINE LEARNING APPROACHES

Audiometric tests can identify the hearing loss at specific frequencies using the audiogram. The aim and objectives of the study were (i) to develop an automated audiometer for self-diagnosing the hearing ability of the patient; (ii) to extract the features from the acoustic signals and to classify the normal and profound hearing loss patients using different machine learning algorithms; (iii) to validate the hearing loss classification using six-frequency average (6-FA) method based on simple linear regression analysis and machine learning algorithms. The study is conducted among 150 patients, including 75 patients with normal hearing ability and 75 patients with profound hearing loss. The total population of 150 underwent audiometric test both in the soundproof audiometric room and in the normal field environment. Based on the patient response, the intensity and frequency are changed automatically, and the audiogram is plotted by the principle of Artificial Neural Network learning procedures. The overall accuracy produced by classification of normal and profound hearing loss patients using Support Vector Machine (SVM), k-Nearest Neighbor classifier, and Naïve Bayes classifier is 97%, 96%, and 95%, respectively. The results indicated that the SVM classifier outperforms the other two classifiers well. The preliminary audiometric test can be performed remotely and then consulted with an audiologist. Thus, the patient could operate the developed prototype independently and get a consultation from trained medical personnel.

Design and optimization of a wave energy converter for drifting sensor platforms in realistic ocean waves

One of the biggest challenges in converting wave energy is to enable the use of low frequency waves, since the highest waves in typical sea states have low frequencies, as can be seen from the corresponding wave spectra, such as the Pierson–Moskowitz or JONSWAP spectra. In this paper, we show that this challenge is indeed achievable for the operation of small autonomous drifting sensor platforms. We present the design and optimization of a compact wave energy converter that freely floats in random sea waves. An optimization of the dynamical behavior as well as the electromagnetic power take-off is conducted based on simulations and experiments. The platform has compact dimensions of 50 cm draft and 50 cm diameter, which leads to special requirements for size and appearance. To meet these requirements, a two-body self-reacting point absorber is designed and a flux switching permanent magnet linear machine is developed for the power take-off. The developed system is validated by experiments in a wave flume and the linear generator is analyzed on a test bench. A coupled model is used to simulate and optimize the corresponding mechanical system, which leads to an increased output power from below 10 mW for the simulated initial setup to a power output of more than 100 mW in the simulation. Simulations and experiments are performed for regular and random waves in order to provide realistic approximations of the total output power.

Multi‐physics analysis of double‐sided tubular linear machine with misalignment faults

Multi air-gap tubular linear machines usually take the form of a double-sided tubular linear machine (DSTLM). Such machines have a higher space utilisation and higher power density than their single air-gap counterparts, but at the same time the structure is more complex. It is traditionally very difficult to mechanically implement a movement between two slotted stators, whilst maintaining an even air gap distribution. Thermally, the inner stator is another complex aspect of a DSTLM. However, very little understanding of these two aspects and how they influence the performances of tubular machines exists today. Therefore in this paper, the aim is to accurately analyse the influence of temperature rise and eccentricity fault on the performances of DSTLM. A two-way electromagnetic-fluid-thermal multi-physics coupling model of the DSTLM is established. The variations of electromagnetic, fluid and thermal characteristics with the eccentricity are systematically analysed. A physical prototype is manufactured and the analysed results are verified by the experiments.

Thermal Analysis of a Novel Linear Oscillating Machine Based on Direct Oil-Cooling Windings

High-frequency electromagnetic linear machine is one major component for linear eletro-hydraulic actuator that in turn is extrimely important for the flight control of more-electrical or all-electrical aircrafts. The employment of high-frequency linear machine helps to achieve high reliability and dynamic performance of the actuators. However, the capability of heat dissipation of the linear machine unavoidably constrains the input current, which in turn depresses its output performance. Therefore, this paper proposes a novel electromagnetic linear machine based on the self-powered direct oil-cooling windings. Specifically, the windings are immersed into the oil directly, which helps to decrease the temperature of the linear machine significantly and thus increase its output performance. In addition, the oil is pushed by the motion of the mover itself. No additional pump is required any more. Thus, the system can be simplified greatly. The design concept of high-frequency electromagnetic linear machine along with its operating principle is presented. A high-precision three-dimensional (3D) piecewise modeling method is proposed to simultaneously consider axial, radial and circumferential equivalent thermal circuits. Numerical simulation is conduced to validated the analytical model. A research prototype of the linear oscillating machine based on direct oil-cooling windings is developed and empolyed for experimental study, and the experimental result verifies the mathematical model of the thermal field well.