Military Training Aircraft Structural Health Monitoring Leveraging an Innovative Biologically Inspired Feedback Mechanism for Neural Networks

Machines, Journal Year: 2025, Volume and Issue: 13(3), P. 179 - 179

Published: Feb. 24, 2025

Structural health monitoring (SHM) is crucial for ensuring the safety and longevity of military training aircraft, which face demanding conditions such as high maneuverability, variable loads, extreme environments, leading to structural fatigue. Traditional methods, modal analysis, often struggle handle multivariate complexity operational data variability. Recently, deep learning has emerged a promising alternative overcome these limitations. However, models typically operate in unidirectional manner, where feedback inputs neglected. In contrast, biological neurons utilize mechanisms refine adapt their responses natural ecosystems, enabling adaptive error correction. this context, study proposes an innovative Convolutional Neural Network with Reversed Mapping (CNN-RM) approach SHM, incorporates loops self-correcting mechanisms. Before feeding into CNN-RM, dataset reduced through time-series-to-images Continuous Wavelet Transform (CWT), followed by denoising CNN (DnCNN) mitigate complex behavior under various conditions. For application, utilizes massive collected from sensors installed on decommissioned aircraft previously used British Royal Air Force now housed laboratory environment. The results revealed that overall mean classification metrics 0.9673 (training) 0.9422 (testing), while CNN-MR, it 0.9764 0.9515 showing improvement 0.94% 1.00% testing. These highlight significant advancements recommending consideration neural models.

Language: Английский

---

Joint Image Processing with Learning-Driven Data Representation and Model Behavior for Non-Intrusive Anemia Diagnosis in Pediatric Patients

Journal of Imaging, Journal Year: 2024, Volume and Issue: 10(10), P. 245 - 245

Published: Oct. 2, 2024

Anemia diagnosis is crucial for pediatric patients due to its impact on growth and development. Traditional methods, like blood tests, are effective but pose challenges, such as discomfort, infection risk, frequent monitoring difficulties, underscoring the need non-intrusive diagnostic methods. In light of this, this study proposes a novel method that combines image processing with learning-driven data representation model behavior anemia in patients. The contributions threefold. First, it uses an image-processing pipeline extract 181 features from 13 categories, feature-selection process identifying most learning. Second, deep multilayered network based long short-term memory (LSTM) utilized train classifying images into anemic non-anemic cases, where hyperparameters optimized using Bayesian approaches. Third, trained LSTM integrated layer learning developed recurrent expansion rules, forming part new called (RexNet). RexNet designed learn representations akin traditional deep-learning methods while also understanding interaction between dependent independent variables. proposed approach applied three public datasets, namely conjunctival eye images, palmar fingernail children aged up 6 years. achieves overall evaluation 99.83 ± 0.02% across all classification metrics, demonstrating significant improvements results generalization compared networks existing This highlights RexNet's potential promising alternative blood-based diagnosis.

Language: Английский

---

Adaptive Fault-Tolerant Tracking Control for Multi-Joint Robot Manipulators via Neural Network-Based Synchronization

Sensors, Journal Year: 2024, Volume and Issue: 24(21), P. 6837 - 6837

Published: Oct. 24, 2024

In this paper, adaptive fault-tolerant control for multi-joint robot manipulators is proposed through the combination of synchronous techniques and neural networks. By using a synchronization technique, position error at each joint simultaneously approaches zero during convergence due to constraints imposed by controller. This aspect particularly important in control, as it enables rapidly effectively reduce impact faults, ensuring performance when faults occur. Additionally, network technique used compensate uncertainty, disturbances, system via online updating. Firstly, novel robust manipulator based on terminal sliding mode presented. Subsequently, enhance fault tolerance manipulator. Finally, simulation results 3-DOF are presented demonstrate effectiveness controller comparison traditional techniques.

Language: Английский

---

The Neural Frontier of Future Medical Imaging: A Review of Deep Learning for Brain Tumor Detection

Journal of Imaging, Journal Year: 2024, Volume and Issue: 11(1), P. 2 - 2

Published: Dec. 24, 2024

Brain tumor detection is crucial in medical research due to high mortality rates and treatment challenges. Early accurate diagnosis vital for improving patient outcomes, however, traditional methods, such as manual Magnetic Resonance Imaging (MRI) analysis, are often time-consuming error-prone. The rise of deep learning has led advanced models automated brain feature extraction, segmentation, classification. Despite these advancements, comprehensive reviews synthesizing recent findings remain scarce. By analyzing over 100 papers past half-decade (2019-2024), this review fills that gap, exploring the latest methods paradigms, summarizing key concepts, challenges, datasets, offering insights into future directions using learning. This also incorporates an analysis previous targets three main aspects: results revealed primarily focuses on Convolutional Neural Networks (CNNs) their variants, with a strong emphasis transfer pre-trained models. Other Generative Adversarial (GANs) Autoencoders, used while Recurrent (RNNs) employed time-sequence modeling. Some integrate Internet Things (IoT) frameworks or federated real-time diagnostics privacy, paired optimization algorithms. However, adoption eXplainable AI (XAI) remains limited, despite its importance building trust diagnostics. Finally, outlines opportunities, focusing image quality, underexplored techniques, expanding deeper representations model behavior recurrent expansion advance imaging

Language: Английский

---

Machine Learning‐Guided Design of 10 nm Junctionless Gate‐All‐Around Metal Oxide Semiconductor Field Effect Transistors for Nanoscaled Digital Circuits

physica status solidi (a), Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 12, 2024

In this paper, we introduce an innovative design approach based on combined numerical simulations and machine learning (ML) analysis to investigate the key parameters of ultra‐low scale junctionless gate‐all‐around (JLGAA) field‐effect transistor (FET) devices. To end, precise 3D models that incorporate quantum effects ballistic transport are employed simulate current–voltage ( I – V ) characteristics 10 nm‐scale JLGAA FET The influence parameter variations high‐k dielectric material subthreshold is thoroughly examined. Various ML algorithms were analyze classify influencing figures‐of‐merit (FoMs), swing (SS) factor ON / OFF ratio. obtained results highlight channel radius doping particularly important for affecting behavior. Similarly, these features also play a significant role in predicting current ratio values. Additionally, used determine optimal each figure merit (FoM) output value. context, effectively predicted both ratios SS classification, with Naive Bayes achieving accuracy 90.8% 92.6% SS, showcasing model's robustness classification tasks.

Language: Английский

---