Normalized Root Mean Research Articles

UWB indoor localization method based on neural network multi-classification for NLOS distance correction

It is well known that ultra-wideband (UWB) is widely used in building indoor positioning systems (IPS) because of its unique advantages. However, compared with the line-of-sight environment (LOS), UWB localization on none-line-of-sight (NLOS) channels has certain limitations, which will reduce the UWB ranging accuracy and location reliability in indoor environment. In this paper, a neural network (NN)-enhanced UWB positioning method is proposed. It can improve positioning performance by using the received channel impulse response (CIR) and UWB raw ranging data to classify the channel conditions and predict the distance. By training CNN-LSTM and MLP neural networks, the proposed method can alleviate the deterioration of localization performance caused by NLOS. The experimental results showed that the average NLOS recognition accuracy of five different obstacles including wooden doors, concrete walls, metal shelves, human body and glass windows reaches up to 92.36 %. In addition, the average root mean square error (RMSE) between the predicted distance and the true distance was 0.3123 m. The indoor positioning test was carried out by weighted least squares (WLS) and the average positioning error under three trajectories was 0.1223 m, which improved the performance by 83.56 % compared with the original UWB positioning system, thus proving its ability to reduce positioning degradation.

Studying Past Earthquakes with Modern Techniques: Ground-Motion Simulations for the 11 January 1693 Noto Earthquake in Italy

Abstract The 1693 Noto earthquake, which struck on 11 January, is one of the Italy’s largest and most devastating earthquakes. According to the Italian Parametric Earthquake Catalogue, it reached a maximum intensity of 11 on the Mercalli–Cancani–Sieberg scale and had an estimated magnitude of M 7.3. Nevertheless, its precise location and source definition remain subjects to debate due to the complexity of the seismic sequence and lack of geological evidence. A series of potential seismic sources differing in location, dimension, and kinematics have been proposed in the literature based on seismotectonic data and interpretations. The goal of this work is to perform a retrospective experiment to verify which of the proposed seismic sources have a better fit with the observed intensity data. To do so, novel simulation techniques are used to study this historical earthquake. We generated ground-motion scenarios for each proposed source model through a stochastic finite-fault simulation approach. Then, the simulated ground-motion parameters were converted to intensities using two different ground-motion intensity conversion equations for Italy. Finally, we compared these converted intensities with the observed intensity data in terms of normalized root mean square errors and converted intensities from ground-motion models. Our results generally show good consistency between converted intensities from the simulated and predicted ground motions, whereas the observed intensities fit better to converted ones from the peak ground velocity rather than peak ground acceleration. Our analysis reveals that the source model reproducing the best of the macroseismic data of the Noto earthquake is the Canicattì–Villasmundo fault system with a magnitude of 7.1.

Predictive models for estimating the sugar content and organic acids in processed mangoes based on the initial content

SummaryThe quality of processed products can be adversely affected by uncontrollable batches of mangoes, which exhibit heterogeneous characteristics. This study aimed to establish predictive models for sugar and organic acid contents (dependent variables) in processed products using the initial compositions of fresh mangoes. Three mango cultivars (cv. ‘Kent’, cv. ‘Keo Romeat’, and cv. ‘Keo Chen’) were classified as low‐density and high‐density groups. Each group of mangoes at the green‐mature, mid‐ripe, and ripe stages was processed into pasteurised purees, dried slices, and mango chips. Prediction models were established using a mix of simple linear regression (SLR) based on the initial content and Analysis of Variance (ANOVA) to identify the impact of qualitative variables (ripening stage, cultivar‐density, and processing technique). In processed mangoes, 13% sucrose content was estimated to accumulate with the three qualitative variables, whereas glucose and fructose contents decreased from their initial levels by 10% and 7%, respectively. Processing techniques can predict the ratio of sugars/acids (S/A) in processed products, regardless of the ripening stage or cultivar‐density. Similar to S/A, citric acid and malic acid contents in mango products were significantly increased by processing techniques. The initial content and processes were insufficient to predict the final contents of the same parameters in processed mangoes; therefore, some models need to include the effects of ripening stage and cultivar‐density to improve the prediction. These relevant explanatory variables contributed significantly to the development of the models, resulting the accuracy of predictive models with normalised root mean square errors (NRMSEs) lower than 10%, except for malic acid (14.04%). In conclusion, it is feasible to estimate the sugar and acidity levels in processed mangoes, offering promising possibilities for ensuring consistent quality of mango‐based products.

Investigating the Use of Traveltime and Reflection Tomography for Deep Learning-Based Sound-Speed Estimation in Ultrasound Computed Tomography.

Ultrasound computed tomography (USCT) quantifies acoustic tissue properties such as the speed-of-sound (SOS). Although full-waveform inversion (FWI) is an effective method for accurate SOS reconstruction, it can be computationally challenging for large-scale problems. Deep learning-based image-to-image learned reconstruction (IILR) methods can offer computationally efficient alternatives. This study investigates the impact of the chosen input modalities on IILR methods for high-resolution SOS reconstruction in USCT. The selected modalities are traveltime tomography (TT) and reflection tomography (RT), which produce a low-resolution SOS map and a reflectivity map, respectively. These modalities have been chosen for their lower computational cost relative to FWI and their capacity to provide complementary information: TT offers a direct SOS measure, while RT reveals tissue boundary information. Systematic analyses were facilitated by employing a virtual USCT imaging system with anatomically realistic numerical breast phantoms. Within this testbed, a supervised convolutional neural network (CNN) was trained to map dual-channel (TT and RT images) to a high-resolution SOS map. Single-input CNNs were trained separately using inputs from each modality alone (TT or RT) for comparison. The accuracy of the methods was systematically assessed using normalized root mean squared error (NRMSE), structural similarity index measure (SSIM), and peak signal-to-noise ratio (PSNR). For tumor detection performance, receiver operating characteristic analysis was employed. The dual-channel IILR method was also tested on clinical human breast data. Ensemble average of the NRMSE, SSIM, and PSNR evaluated on this clinical dataset were 0.2355, 0.8845, and 28.33 dB, respectively.

Monitoring aboveground organs biomass of wheat and maize: A novel model combining ensemble learning and allometric theory

Accurate monitoring of crop organ biomass facilitates optimizing agronomic strategies to maximize yield or economic benefit. Unmanned aerial vehicle (UAV) is extensively employed for aboveground biomass (AGB) monitoring at the farm scale, but previous studies have mostly concentrated on total AGB rather than individual organ biomass. Furthermore, film-mulched crops, a widely used cropping pattern in northwest China, have received less attention for AGB monitoring. We aim to develop a novel model to precisely estimate the AGB of leaf (AGBLeaf), stem (AGBStem), and reproductive organs (AGBR) by UAV for film-mulched wheat and maize. The maize-wheat rotation field experiments with treatments of five nitrogen application amounts and three planting densities were conducted from 2021 to 2023, respectively. Firstly, we constructed allometric models at jointing, heading (tasseling), and grain filling stages by ground sampling data in 2021–2022. Next, the input feature set was obtained by feature selection methods (Lasso and Boruta) using UAV image data, and three traditional methods (partial least squares, ridge regression, and support vector machine) and three ensemble learning models (random forest, extreme gradient boosting, and local cascade ensemble (LCE)) were trained for AGBLeaf inversion based on the physically-based PROSAIL model simulation dataset. Finally, the optimal AGBLeaf inversion hybrid model was coupled with the allometric model to estimate the AGBStem and AGBR in 2022–2023. The results indicated that both wheat and maize organ biomass conformed to the allometric pattern. While feature selection helped reduce computation and complexity, but didn’t improve monitoring accuracy. The normalized root mean square error (NRMSE) of the optimal hybrid model (PROSAIL + Boruta + LCE) on the measured wheat and maize AGBLeaf datasets were 12.72 %–24.93 % and 19.65 %–25.16 %, respectively. After coupling the allometric model, the coefficient of determination (R2) of wheat and maize AGBStem were 0.64–0.85 and 0.63–0.68, and the NRMSE were 15.05 %–25.28 % and 24.10 %–27.06 %, respectively; and the corresponding R2 of AGBR was 0.67–0.76 and 0.72, and the NRMSE were 16.81 %–22.12 % and 21.66 %, for wheat and maize, respectively. Overall, the novel model performed well in film-mulched wheat and maize, providing a cost-effective approach for organ biomass monitoring. In the future, further validation of the model’s transferability is necessary to increase the potential for generalization in production practice.

Musculoskeletal models determine the effect of a soft active exosuit on muscle activations and forces during lifting and lowering tasks

Exosuits have the potential to mitigate musculoskeletal stress and prevent back injuries during industrial tasks. This study aimed to 1) validate the implementation of a soft active exosuit into a musculoskeletal model of the spine by comparing model predicted muscle activations versus corresponding surface EMG measurements, and 2) evaluate the effect of the exosuit on peak back and hip muscle forces. Fourteen healthy participants performed squat and stoop lift and lower tasks with boxes of 6 and 10 kg, with and without wearing a 2.7 kg soft active exosuit. Participant-specific musculoskeletal models, which included the exosuit, were created in OpenSim. Model validation focused on the back and hip extensors, where temporal agreement between EMG and model estimated muscle activity was generally strong to excellent (average cross-correlation coefficients ranging from 0.84 to 0.98). Root mean square errors of muscle activity (0.05–0.10) were similar with and without the exosuit, and compared well to prior model validation studies without the exosuit (average root mean square errors ranging from 0.05 to 0.19). In terms of performance, the exosuit reduced the estimated peak erector spinae forces during lifting and lowering phases across all lifting tasks but reduced peak hip extensor muscles forces only in a squat lift task of 10 kg. These reductions in total peak muscle forces were approximately 1.7–4.2 times greater than the corresponding exosuit assistance force, which were 146 ± 19 N and 102 ± 14 N at the times of peak erector spinae forces in lifting and lowering, respectively. Overall, the results support the hypothesis that exosuits reduce soft tissue loading, and thereby potentially reduce fatigue and injury risk during manual materials handling tasks. Incorporating exosuits into musculoskeletal models is a valid approach to understand the impact of exosuit assistance on muscle activity and forces.

Influence of brick laying height on biomechanical load in masons: Cross-sectional field study with technical measurements.

Work-related musculoskeletal disorders (WMSDs) located in the low back and neck/shoulder regions are major concerns for both workers, workplaces, and society. Masons are prone to WMSD, because their work is characterized by repetitive work and high physical workload. However, the knowledge on the physical workload during bricklaying is primarily based on subjective measurements. This cross-sectional field study with technical measurements aimed to quantify physical workload in terms of muscular activity and degree of forward bending during bricklaying at different working heights among masons, i.e., knee, hip, shoulder, and above shoulder height. Twelve male (36.1±16.1 years) experienced masons participated in a cross-sectional field study with technical measurements. Surface electromyography from erector spinae longissimus and upper trapezius muscles and an inertial measurement unit-sensor placed on the upper back were used to assess the physical workload (level of muscle activation and degree of forward bending) different bricklaying heights. Manual video analysis was used to determine duration of work tasks, frequency, type, and working height. The working heights were categorized as 'knee', 'hip', 'shoulder', and 'above shoulder'. The 95 percentiles of the normalized Root Mean Square (RMSn) values were extracted assess from erector spinae and trapezius recordings to assess strenuous level muscle of muscle activation. The RMSn of dominant erector spinae muscle increased from hip- to shoulder height (from 26.6 to 29.6, P < 0.0001), but not from hip to above shoulder height and decreased from hip to knee height (from 26.6 to 18.9, P < 0.0001). For the dominant trapezius muscle, the RMSn increased from hip- to shoulder- and above shoulder height (from 13.9 to 19.7 and 24.0, respectively, P < 0.0001) but decreased from hip- to knee height (from 13.9 to 11.5, P < 0.0001). Compared to hip height (27.9°), an increased forward bending was detected during bricklaying at knee height (34.5°, P < 0.0001) and a decreased degree of forward bending at shoulder- and above shoulder height (17.6° and 12.5°, P < 0.0001, respectively). Based on technical measurements, bricklaying at hip height showed the best compromise between muscular load and degree of forward bending. This study contributes to the development of the work environment for masons and can help guide preventive initiatives to reduce physical workload.

Fractional calculus integration for improved ECG modeling: A McSharry model expansion

This study introduces a new method for modeling electrocardiogram (ECG)11- CR: Compression Ratio; FDE: Fractional Differential Equation; GA: Genetic Algorithm; IDE: Integer Differential Equation; MAE: Mean Absolute Error; MSE: Mean Square Error; NRMSE: Normalized Root Mean Square Error; PC: Predictor Corrector; PRD: Percentage Root Mean Square Difference; QS: Quality Score; RMSE: Root Mean Square Error waveforms using Fractional Differential Equations (FDEs). By incorporating fractional calculus into the well-established McSharry model, the proposed approach achieves improved representation and high precision for a wide range of ECG waveforms. The research focuses on the impact of integrating fractional derivatives into Integer Differential Equation (IDE) models, enhancing the fidelity of ECG signal modeling.To optimize the model's unknown parameters, a combination of the Predictor-Corrector method for solving FDEs and genetic algorithms for optimization is utilized. The effectiveness of the fractional-order model is assessed through distortion metrics, providing a comprehensive evaluation of the modeling quality.Comparisons show that the fractional-order model outperforms the traditional McSharry IDE model in modeling quality and compression efficiency. It improves modeling quality by 48.40 % in MSE and compression efficiency by 23.18 % when applied on five beat types of MIT/BIH arrhythmia database. The fractional-order model demonstrates enhanced flexibility while preserving essential McSharry model characteristics, with fractional orders (α) ranging from 0.96 to 0.99 across five beat types.

Estimating Ground-Level NO2 Concentrations Using Machine Learning Exclusively with Remote Sensing and ERA5 Data: The Mexico City Case Study

This study explores the estimation of ground-level NO2 concentrations in Mexico City using an integrated approach of machine learning (ML) and remote sensing data. We used the NO2 measurements from the Sentinel-5P satellite, along with ERA5 meteorological data, to evaluate a pre-trained machine learing model. Our findings indicate that the model captures the spatial and temporal variability of NO2 concentrations across the urban landscape. Key meteorological parameters, such as temperature and wind speed, were identified as significant factors influencing NO2 levels. The model’s adaptability was further tested by incorporating additional variables, such as atmospheric boundary layer height. In order to compare the model’s performance to alternative ML models, we estimated the ground-level NO2 using the state-of-the-art TimeGPT. The results demonstrate that our baseline model has the best performance with a mean normalised root mean square error of 84.47%. This research underscores the potential of combining satellite observations with ML for scalable air quality monitoring, particularly in low- and middle-income countries with limited ground-based infrastructure. The study provides critical insights for air quality management and policy-making, aiming to mitigate the adverse health and environmental impacts of NO2 pollution.

Deep learning infused SIRVD model for COVID-19 prediction: XGBoost-SIRVD-LSTM approach.

The global impact of the ongoing COVID-19 pandemic, while somewhat contained, remains a critical challenge that has tested the resilience of humanity. Accurate and timely prediction of COVID-19 transmission dynamics and future trends is essential for informed decision-making in public health. Deep learning and mathematical models have emerged as promising tools, yet concerns regarding accuracy persist. This research suggests a novel model for forecasting the COVID-19's future trajectory. The model combines the benefits of machine learning models and mathematical models. The SIRVD model, a mathematical based model that depicts the reach of the infection via population, serves as basis for the proposed model. A deep prediction model for COVID-19 using XGBoost-SIRVD-LSTM is presented. The suggested approach combines Susceptible-Infected-Recovered-Vaccinated-Deceased (SIRVD), and a deep learning model, which includes Long Short-Term Memory (LSTM) and other prediction models, including feature selection using XGBoost method. The model keeps track of changes in each group's membership over time. To increase the SIRVD model's accuracy, machine learning is applied. The key properties for forecasting the spread of the infection are found using a method called feature selection. Then, in order to learn from these features and create predictions, a model involving deep learning is applied. The performance of the model proposed was assessed with prediction metrics such as R 2, root mean square error (RMSE), mean absolute percentage error (MAPE), and normalized root mean square error (NRMSE). The results are also validated to those of other prediction models. The empirical results show that the suggested model outperforms similar models. Findings suggest its potential as a valuable tool for pandemic management and public health decision-making.

Application of convolutional long short-term memory for spatio-temporal forecastings of hydrocarbon saturations and pressure in oil fields

A machine learning architecture composed of convolutional long short-term memory (convLSTM) is developed to predict spatiotemporal parameters, that is, saturation and pressure in an oil & gas field namely, the SACROC field in Texas, USA. The spatial parameters are recorded at the end of each month for 360 months, approximately 83% of which is used for training and the rest 17% is kept for testing. The samples for the convLSTM models are prepared by choosing ten consecutive frames as input and ten consecutive frames shifted forward by one frame as output. A workflow is provided describing the entire process of data extraction, preprocessing, sample preparation, training, testing of machine learning models, and error analysis. The structural similarity index measure (SSIM) and normalized root mean squared error (NRMSE) are calculated during training and testing for quantitative error analysis. During the training phase, the predicted frames demonstrate a strong resemblance to the ground truth frames. Moreover, it exhibits superior performance during the initial months of testing. As testing progresses into the later months, deviations increase as errors accumulate over time for relying on past predictions. In summary, convLSTM for spatio-temporal prediction offers significant potential for forecasting and optimizing hydrocarbon recovery from porous media.

Improving GNSS-IR Sea Surface Height Accuracy Based on a New Ionospheric Stratified Elevation Angle Correction Model

Approximately 71% of the Earth’s surface is covered by vast oceans. With the exacerbation of global climate change, high-precision monitoring of sea surface height variations is of vital importance for constructing global ocean gravity fields and preventing natural disasters in the marine system. Global Navigation Satellite System Interferometry Reflectometry (GNSS-IR) sea surface altimetry is a method of inferring sea surface height based on the signal-to-noise ratio of satellite signals. It enables the retrieval of sea surface height variations with high precision. However, navigation satellite signals are influenced by the ionosphere during propagation, leading to deviations in the measured values of satellite elevation angles from their true values, which significantly affects the accuracy of GNSS-IR sea surface altimetry. Based on this, the contents of this paper are as follows: Firstly, a new ionospheric stratified elevation angle correction model (ISEACM) was developed by integrating the International Reference Ionosphere Model (IRI) and ray tracing methods. This model aims to improve the accuracy of GNSS-IR sea surface altimetry by correcting the ionospheric refraction effects on satellite elevation angles. Secondly, four GNSS stations (TAR0, PTLD, GOM1, and TPW2) were selected globally, and the corrected sea surface height values obtained using ISEACM were compared with observed values from tide gauge stations. The calculated average Root Mean Square Error (RMSE) and Pearson Correlation Coefficient (PCC) were 0.20 m and 0.83, respectively, indicating the effectiveness of ISEACM in sea surface height retrieval. Thirdly, a comparative analysis was conducted between sea surface height retrieval before and after correction using ISEACM. The optimal RMSE and PCC values with tide gauge station observations were 0.15 m and 0.90, respectively, representing a 20.00% improvement in RMSE and a 4.00% improvement in correlation coefficient compared to traditional GNSS-IR retrieval heights. These experimental results demonstrate that correction with ISEACM can effectively enhance the precision of GNSS-IR sea surface altimetry, which is crucial for accurate sea surface height measurements.

Strain-mediated reservoir computing with temporal and spatial co-multiplexing in multiferroic heterostructures

Physical reservoir computing (PRC), a brain-inspired computing method known for its efficient information processing and low training requirements, has attracted significant attention. The key factor lies in the number of computational nodes within the reservoir for its computational capability. Here, we explore co-multiplexing reservoirs that leverage both temporal and spatial strategies. Temporal multiplexing virtually expands the node count through the use of masking techniques, while spatial multiplexing utilizes multiple physical locations (e.g., Hall bars) to achieve an increase in the number of real nodes. Our experiment employs a strain-mediated reservoir based on multiferroic heterostructures. By applying a single voltage across the PMN-PT substrate (acting as global input) and measuring the output Hall voltages from four Hall bars (real nodes), we achieve significant efficiency gains. This co-multiplexing approach results in a reduction in the normalized root mean square error from 0.5 to 0.23 for a 20-step prediction task of a Mackey–Glass chaotic time series. Furthermore, the single input and four independent outputs lead to a fourfold reduction in energy consumption compared to the strain-mediated PRC with temporal multiplexing solely. This research paves the way for future energy saving PRC implementations utilizing co-multiplexing, promoting a resource-efficient paradigm in reservoir computing.

Hourly Downscaling of Daily Climate Projection Data for the Nakdong River Basin using the Nearest Neighbor Search Method

Hydrological and water quality models have been continuously improved for assessing the impacts of climate change. Accordingly, meteorological data with high spatial and temporal resolutions are required. Current climate change scenario-based projections provide meteorological data on monthly and daily scales, which are too coarse for regional climate change assessments. In this study, we propose a temporal downscaling method based on the nearest neighbor search methodology for temporal downscaling from daily time-step meteorological data to hourly time-step data. To verify the temporal downscaling method, historical meteorological data from weather stations located in the Nakdong River basin were used, and the normalized root mean square error (NRMSE) between observational and simulated data was calculated. Consequently, the simulation errors were approximately 2–4% for temperature and precipitation, and approximately 7–16% for wind speed, relative humidity, and solar radiation. Next, using the temporal downscaling method, hourly future climate projection data were derived from the daily SSP5-8.5 climate scenario projections retrieved from weather stations in the target area. According to the downscaled hourly climate projection results, under the SSP5-8.5 climate change scenario, the future annual maximum temperature is projected to increase by up to 7.0 ℃ and 5.6 ℃ at Daegu and Busan weather stations, compared to the reference period maximum temperatures of 36.2 ℃ and 33 ℃, respectively. The annual precipitation duration is projected to decrease by up to 103.3 h (21.25%) and 157.2 h (27.57%), compared to the reference period durations of 486.2 h and 570.2 h, respectively. Annual precipitation intensity is projected to increase by up to 0.71 mm/h (33.5%) and 0.83 mm/h (31.2%), compared to the reference period precipitation intensities of 2.12 mm/h and 2.66 mm/h, respectively. The temporal downscaling method presented in this study is expected to contribute to the preparation of meteorological input data to effectively simulate detailed hourly rainfall runoff and the behavior of water quality factors, thereby improving the assessment of climate change impacts.

Multi-metric comparison of machine learning imputation methods with application to breast cancer survival

Handling missing data in clinical prognostic studies is an essential yet challenging task. This study aimed to provide a comprehensive assessment of the effectiveness and reliability of different machine learning (ML) imputation methods across various analytical perspectives. Specifically, it focused on three distinct classes of performance metrics used to evaluate ML imputation methods: post-imputation bias of regression estimates, post-imputation predictive accuracy, and substantive model-free metrics. As an illustration, we applied data from a real-world breast cancer survival study. This comprehensive approach aimed to provide a thorough assessment of the effectiveness and reliability of ML imputation methods across various analytical perspectives. A simulated dataset with 30% Missing At Random (MAR) values was used. A number of single imputation (SI) methods - specifically KNN, missMDA, CART, missForest, missRanger, missCforest - and multiple imputation (MI) methods - specifically miceCART and miceRF - were evaluated. The performance metrics used were Gower’s distance, estimation bias, empirical standard error, coverage rate, length of confidence interval, predictive accuracy, proportion of falsely classified (PFC), normalized root mean squared error (NRMSE), AUC, and C-index scores. The analysis revealed that in terms of Gower’s distance, CART and missForest were the most accurate, while missMDA and CART excelled for binary covariates; missForest and miceCART were superior for continuous covariates. When assessing bias and accuracy in regression estimates, miceCART and miceRF exhibited the least bias. Overall, the various imputation methods demonstrated greater efficiency than complete-case analysis (CCA), with MICE methods providing optimal confidence interval coverage. In terms of predictive accuracy for Cox models, missMDA and missForest had superior AUC and C-index scores. Despite offering better predictive accuracy, the study found that SI methods introduced more bias into the regression coefficients compared to MI methods. This study underlines the importance of selecting appropriate imputation methods based on study goals and data types in time-to-event research. The varying effectiveness of methods across the different performance metrics studied highlights the value of using advanced machine learning algorithms within a multiple imputation framework to enhance research integrity and the robustness of findings.

Myo-regressor Deep Informed Neural NetwOrk (Myo-DINO) for fast MR parameters mapping in neuromuscular disorders

Magnetic Resonance (MR) parameters mapping in muscle Magnetic Resonance Imaging (mMRI) is predominantly performed using pattern recognition-based algorithms, which are characterised by high computational costs and scalability issues in the context of multi-parametric mapping.Deep Learning (DL) has been demonstrated to be a robust and efficient method for rapid MR parameters mapping. However, its application in mMRI domain to investigate Neuromuscular Disorders (NMDs) has not yet been explored. In addition, data-driven DL models suffered in interpretation and explainability of the learning process. We developed a Physics Informed Neural Network called Myo-Regressor Deep Informed Neural NetwOrk (Myo-DINO) for efficient and explainable Fat Fraction (FF), water-T2 (wT2) and B1 mapping from a cohort of NMDs.A total of 2165 slices (232 subjects) from Multi-Echo Spin Echo (MESE) images were selected as the input dataset for which FF, wT2,B1 ground truth maps were computed using the MyoQMRI toolbox. This toolbox exploits the Extended Phase Graph (EPG) theory with a two-component model (water and fat signal) and slice profile to simulate the signal evolution in the MESE framework. A customized U-Net architecture was implemented as the Myo-DINO architecture. The squared L2 norm loss was complemented by two distinct physics models to define two ‘Physics-Informed’ loss functions: Cycling Loss 1 embedded a mono-exponential model to describe the relaxation of water protons, while Cycling Loss 2 incorporated the EPG theory with slice profile to model the magnetization dephasing under the effect of gradients and RF pulses. The Myo-DINO was trained with the hyperparameter value of the 'Physics-Informed' component held constant, i.e. λmodel = 1, while different hyperparameter values (λcnn) were applied to the squared L2 norm component in both the cycling loss. In particular, hard (λcnn=10), normal (λcnn=1) and self-supervised (λcnn=0) constraints were applied to gradually decrease the impact of the squared L2 norm component on the ‘Physics Informed’ term during the Myo-DINO training process.Myo-DINO achieved higher performance with Cycling Loss 2 for FF, wT2 and B1 prediction. In particular, high reconstruction similarity and quality (Structural Similarity Index > 0.92, Peak Signal to Noise ratio > 30.0 db) and small reconstruction error (Normalized Root Mean Squared Error < 0.038) to the reference maps were shown with self-supervised weighting of the Cycling Loss 2. In addition muscle-wise FF, wT2 and B1 predicted values showed good agreement with the reference values. The Myo-DINO has been demonstrated to be a robust and efficient workflow for MR parameters mapping in the context of mMRI. This provides preliminary evidence that it can be an effective alternative to the reference post-processing algorithm. In addition, our results demonstrate that Cycling Loss 2, which incorporates the Extended Phase Graph (EPG) model, provides the most robust and relevant physical constraints for Myo-DINO in this multi-parameter regression task. The use of Cycling Loss 2 with self-supervised constraint improved the explainability of the learning process because the network acquired domain knowledge solely in accordance with the assumptions of the EPG model.

Unlocking the potential of unlabeled data: Self-supervised machine learning for battery aging diagnosis with real-world field data

Accurate aging diagnosis is crucial for the health and safety management of lithium-ion batteries in electric vehicles. Despite significant advancements achieved by data-driven methods, diagnosis accuracy remains constrained by the high costs of check-up tests and the scarcity of labeled data. This paper presents a framework utilizing self-supervised machine learning to harness the potential of unlabeled data for diagnosing battery aging in electric vehicles during field operations. We validate our method using battery degradation datasets collected over more than two years from twenty real-world electric vehicles. Our analysis comprehensively addresses cell inconsistencies, physical interpretations, and charging uncertainties in real-world applications. This is achieved through self-supervised feature extraction using random short charging sequences in the main peak of incremental capacity curves. By leveraging inexpensive unlabeled data in a self-supervised approach, our method demonstrates improvements in average root mean square errors of 74.54% and 60.50% in the best and worst cases, respectively, compared to the supervised benchmark. This work underscores the potential of employing low-cost unlabeled data with self-supervised machine learning for effective battery health and safety management in real-world scenarios.

Collaborative-driven reservoir formation pressure prediction using GAN-ML models and well logging data

Formation pressure is the foundation and prerequisite for achieving efficient drilling and risk control in complex geological formations. However, intelligent prediction of formation pressure faces challenges such as insufficient sample sizes, poor prediction performance, and low prediction accuracy, which severely restrict the realization of safe and efficient drilling. To address this, this paper proposes a new method based on Generative Adversarial Networks (GAN) and Machine Learning (ML), aiming to improve adaptation to complex geological conditions and effectively capture the nonlinear relationship between logging data and formation pressure for intelligent prediction of formation pressure. The research results show that the GAN technology successfully amplified the 125 cable-measured formation pressure data by 204.8 times. Compared to machine learning predictions without using GAN technology, the constructed GAN-ML model reduced the average root mean square error by 0.59 MPa and the average absolute error by 0.71 MPa in the training set. The GAN-ML model predicted the formation pressure of adjacent wells with an average root mean square error of 1.59 MPa and an average absolute error of 1.96 MPa compared to the cable-measured formation pressure values. This achievement significantly enhances the robustness and generalization performance of intelligent models, demonstrating the potential of GAN-ML in addressing small sample data problems. It opens a new technical path for data-driven intelligent prediction and provides innovative data analysis and prediction methods for the field of geological exploration.

Sunspot number-based neural network model for global solar radiation estimation in Ghardaïa

In this investigation, the estimation of global solar radiation was meticulously carried out within Ghardaïa city, a region situated in Southern Algeria, utilizing a sophisticated multilayer perceptron (MLP) neural network architecture. This research primarily concentrated on developing a predictive model based on a singular input parameter, specifically, the sunspot numbers, to forecast global solar radiation levels. The model's formulation was rooted in empirical data collected over an extensive period from 1984 to 2000, which was used for training the neural network. To assess the model's predictive accuracy and robustness, data from the years 2001 to 2004 were employed for validation purposes. The outcomes of this study were highly satisfactory, indicating that the MLP-based model possesses a significant predictive capability for Diffuse Global Solar Radiation (DGSR). This is substantiated by robust statistical metrics, including a normalized Root Mean Square Error (nRMSE) of 0.076, reflecting the model's accuracy in prediction, and a correlation coefficient (R) of 93.16%, denoting a strong correlation between the predicted and observed values. These results underscore the model's efficacy and potential application in accurately estimating global solar radiation in the specified region.

Design and Testing of a 2-DOF Adaptive Profiling Header for Forage Harvesters

The existing forage harvester header cannot automatically adjust the height and inclination during operation, resulting in uneven stubble height of forage, which, in turn, affects the efficiency of harvesting and the quality of forage regeneration. To address this issue, this study conducted the design and experimentation of a 2-degrees-of-freedom (DOF) profiling header. Firstly, this study designed an adaptive profiling header with 2-DOF adjustment, which was realized by the height adjustment mechanism and the tilt angle adjustment mechanism. The relationship model between the profiling device and the attitude of the header was established so that the header can acquire ground undulation in real time through the angle sensor of the profiling device. In order to verify the rationality of the header design, a co-simulation model of ADAMS and MATLAB/Simulink was built, and the header attitude control system was designed based on the fuzzy PID algorithm. The co-simulation results show that the header height (H) is always kept around 150 mm during the forwarding process of the harvester, with a maximum error of 5.8 mm, and the average relative error (REH) and root mean square error (RMSEH) were 1.4% and 2.6 mm, respectively, and the maximum error of the tilt angle (γ) of the header is 0.53° and the RMSEγ is 0.22°, which indicates that the header profiling mechanism can accurately reflect the undulation of the terrain and the header attitude control system has good robustness. Finally, the test platform was built and tested in a grassland. The test results show that the average height of the header is 149.8 mm, the maximum error is 7.5 mm, and the REH and RMSEH are 3.4% and 5.3 mm, respectively. The average error of the header inclination is 0.34°, and the maximum error is 0.57°. The test results indicate that the header can realize the adaptive adjustment of height and inclination, and the control system has high precision, stability and reliability, meeting the demand of automatic regulation of header attitude of a forage harvester.