Next Issue
Volume 15, August-2
Previous Issue
Volume 15, July-2
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
7.0
2.9
Journals
Electronics
Volume 15
Issue 15
electronics-logo

Journal Menu

► Journal Menu

Journal Browser

► Journal Browser
share announcement

Electronics, Volume 15, Issue 15 (August-1 2026) – 252 articles

Cover Story (view full-size image): This research presents a robust framework for condition monitoring of railway rolling stock, focusing on Y-21 bogie wheelsets. By integrating Empirical Mode Decomposition (EMD) with Support Vector Machines (SVMs), the study categorizes vibration signals into four health states. The methodology employs ANOVA to select the most discriminative predictors: Kurtosis, Crest Factor, and Shape Factor Results show that an SVM model with a Fine Gaussian kernel achieves a diagnostic accuracy of 99.9%. This approach optimizes predictive maintenance by ensuring high precision with a parsimonious set of variables. View this paper
  • Issues are regarded as officially published after their release is announced to the table of contents alert mailing list.
  • You may sign up for e-mail alerts to receive table of contents of newly released issues.
  • PDF is the official format for papers published in both, html and pdf forms. To view the papers in pdf format, click on the "PDF Full-text" link, and use the free Adobe Reader to open them.
Order results
Result details
Section
Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
26 pages, 4273 KB  
Article
An EMF-Aware Intelligent Framework for Adaptive SSB Periodicity Control in 5G Networks
by Keze Li, Michael S. Mollel, Olaoluwa Popoola, Muhammad Ali Imran and Yusuf Sambo
Electronics 2026, 15(15), 3491; https://doi.org/10.3390/electronics15153491 - 6 Aug 2026
Viewed by 289
Abstract
Synchronization Signal Blocks (SSBs) support initial access, synchronization, and beam management in fifth-generation networks. However, the periodic transmission of SSBs contributes to background electromagnetic field (EMF) exposure, while increasing the SSB periodicity may increase the waiting time experienced by newly arriving User Equipment [...] Read more.
Synchronization Signal Blocks (SSBs) support initial access, synchronization, and beam management in fifth-generation networks. However, the periodic transmission of SSBs contributes to background electromagnetic field (EMF) exposure, while increasing the SSB periodicity may increase the waiting time experienced by newly arriving User Equipment (UE). This paper proposes an EMF-aware adaptive SSB periodicity control framework that integrates Long Short-Term Memory (LSTM)-based demand prediction with a Proximal Policy Optimization (PPO) controller. Aggregated Internet usage measurements collected by Telecom Italia in Milan in 2013 at 10 min resolution are transformed into a traffic-derived UE arrival proxy. The LSTM model forecasts the proxy for the subsequent control interval, and the PPO policy selects a SSB periodicity from 5, 10, 20, 40, 80, and 160 ms. The optimization reward combines normalized SSB EMF power density and normalized aggregate UE waiting time, thereby avoiding dimensional and numerical-scale inconsistencies between the two objectives. The LSTM and PPO models are developed using six chronological weeks of data and evaluated without parameter updates over a seven-day held-out period. The prediction-driven controller achieves a mean SSB EMF power density of 3.67×10−5 W/m2 and a weekly average waiting time of 19.22 ms per person. Relative to the fixed 20 and 40 ms configurations, the proposed controller reduces mean EMF power density by approximately 65.4% and 30.8%, respectively. Its waiting time is close to that of the fixed 40 ms configuration and substantially lower than that of the fixed 80 ms configuration. Although the fixed 80 ms configuration provides lower EMF power density, it incurs considerably greater waiting time. The results show that the proposed framework provides an adaptive intermediate operating point between the lower waiting time of short fixed periodicities and the lower EMF exposure of long fixed periodicities. Full article
(This article belongs to the Special Issue Advances in 5G and Beyond Mobile Communication)
►▼ Show Figures

Figure 1

22 pages, 5276 KB  
Article
Research on an Improved Multi-Model Dynamic Fusion Classification Technique
by Xiang Wan, Youxing He, Xionghai Rao, Yijian Qiu, Ruijian Cheng, Jiang Wei, Xiangping Cheng, Tianci Li and Manqing Zhu
Electronics 2026, 15(15), 3490; https://doi.org/10.3390/electronics15153490 - 6 Aug 2026
Viewed by 382
Abstract
Existing multi-model fusion methods generally adopt a global, fixed fusion strategy, applying all base models and weighting rules uniformly to all samples to be classified and all categories, thereby lacking adaptability to specific samples and category-specific targeting. Most traditional dynamic selection and dynamic [...] Read more.
Existing multi-model fusion methods generally adopt a global, fixed fusion strategy, applying all base models and weighting rules uniformly to all samples to be classified and all categories, thereby lacking adaptability to specific samples and category-specific targeting. Most traditional dynamic selection and dynamic weighting fusion methods only implement global parameter adjustments at the sample level, without considering the significant differences in category-specific capabilities among the base models. When a single base model exhibits superior recognition capabilities for only certain categories, its prediction accuracy and confidence for the remaining categories are low. If all base models are fused directly, poor-quality class predictions can cause negative interference and even dominate the final decision, leading to classification errors. Furthermore, traditional dynamic fusion suffers from computational redundancy, difficulty in suppressing interference from low-confidence samples, and the challenge of balancing dynamic optimization with inference efficiency. To address these issues, this paper proposes an improved multi-model dynamic fusion classification technique that differs from the traditional global dynamic fusion paradigm. By constructing a voting matrix, contribution weights, and a matrix of effective category voting weights, this method establishes a category-level model performance evaluation and differentiated weighting mechanism. This enables the precise selection of superior base models for each sample and category, thereby filtering out interference from low-confidence and suboptimal category predictions. At the same time, in the network architecture design, lightweight models are organized into a branch structure, and effective branches are dynamically activated as needed to participate in decision-making, significantly reducing the computational overhead of inference. To validate the fusion classification technique proposed in this paper, for the experiments, we selected mainstream lightweight models such as MobileNetV2, EfficientNetB0, ShuffleNetv2, MNASNet 0.75, and MobileNetV3_Small for evaluation on the NEU dataset and NASA’s Milling Data Set. The experimental results demonstrate that the fusion method proposed in this paper can fully aggregate the category-specific strengths of different lightweight models, effectively mitigate the risk of misclassification associated with traditional fusion methods, and enhance model robustness while ensuring high classification accuracy. It achieves classification performance comparable to that of large deep models with extremely low computational overhead. This method is not only suitable for application in multiple-criteria decision-making but can also be implemented and extended to multi-source/multi-modal data fusion and deep neural networks, making it of practical value. Full article
(This article belongs to the Special Issue Multimodal Learning and Transfer Learning)
►▼ Show Figures

Figure 1

27 pages, 6897 KB  
Article
Lightweight Small-Object Detection for Urban UAV Imagery with Content-Aware Feature Reconstruction and Gradient-Adaptive Localization
by Zefeng Zhao, Fanyu Meng and Jing Bian
Electronics 2026, 15(15), 3489; https://doi.org/10.3390/electronics15153489 - 6 Aug 2026
Viewed by 570
Abstract
Small-object detection in unmanned aerial vehicle (UAV) imagery is hindered by limited target pixels, dense spatial distributions, background interference, and scale-sensitive bounding-box regression. This study develops a lightweight YOLO11n configuration in which established SCSA recalibration, DySample reconstruction, decoupled prediction, and SimOTA assignment act [...] Read more.
Small-object detection in unmanned aerial vehicle (UAV) imagery is hindered by limited target pixels, dense spatial distributions, background interference, and scale-sensitive bounding-box regression. This study develops a lightweight YOLO11n configuration in which established SCSA recalibration, DySample reconstruction, decoupled prediction, and SimOTA assignment act at successive stages of the detection pipeline. Its principal methodological contribution is a gradient-adaptive WIoU–NWD objective that uses previously observed regression-gradient fluctuations to balance overlap-oriented and distribution-based localization without altering the inference graph. On the official VisDrone2019-DET test-dev server, the 2.48 M-parameter model achieves 45.9% mAP@0.5, 27.0% mAP@0.5:0.95, and 20.2% APsmall, improving the YOLO11n baseline by 2.3, 1.4, and 2.4 percentage points, respectively; it also improves mAP@0.5/mAP@0.5:0.95 by 2.2/1.3 points on the vehicle-focused UAVDT benchmark and reaches 40.2 FPS on a Jetson Orin NX in 15 W mode using TensorRT FP16. The two benchmarks mainly represent urban, traffic, and low-altitude surveillance imagery; consequently, the cross-dataset result supports transfer within these conditions rather than universal generalization to all UAV applications. These results support a compact single-pass accuracy–efficiency trade-off for resource-constrained UAV perception, while the modest margin over fixed loss weighting indicates that the adaptive mechanism should be interpreted as an incremental, primarily small-object localization improvement rather than a complete solution to regression instability. Full article
(This article belongs to the Section Artificial Intelligence)
►▼ Show Figures

Figure 1

25 pages, 857 KB  
Article
Frequency–Time Domain Alignment for Cross-Subject Time-Series Classification Under Distribution Shift
by Dong-Hyun Won and Kwang-Seong Shin
Electronics 2026, 15(15), 3488; https://doi.org/10.3390/electronics15153488 - 6 Aug 2026
Viewed by 352
Abstract
Background: Distribution shift is one of the most pervasive data-related challenges for time-series classification. In sensor-based Human Activity Recognition (HAR), a model trained on one population of subjects often degrades substantially when deployed on another, even with identical hardware and activity labels. This [...] Read more.
Background: Distribution shift is one of the most pervasive data-related challenges for time-series classification. In sensor-based Human Activity Recognition (HAR), a model trained on one population of subjects often degrades substantially when deployed on another, even with identical hardware and activity labels. This paper addresses that setting specifically: cross-subject distribution shift. Methods: We propose FTDA (Frequency–Time Domain Alignment), an unsupervised domain-adaptation method that processes raw signals and their FFT-magnitude spectra through a dual-branch encoder, combines a Gradient-Reversal adversarial loss on the joint feature with a Multi-Kernel MMD on the frequency branch, and applies a symmetric-KL time–frequency consistency loss on unlabeled target data. Results: On UCI HAR, over four cross-subject transfer tasks and 10 random seeds (n=40 paired observations), FTDA reaches 95.01% target accuracy against 92.46% for Source-only, 92.29% for DANN and 93.32% for a RAINCOAT-style baseline. Because the tasks share subjects and two of them are mirror configurations, no pooled significance test is offered; inference is confined to within-task paired tests (10 seeds each, Holm-corrected), under which FTDA is ahead of every single-view baseline on all four tasks, while the difference from the two dual-view competitors (+1.69 pp against RAINCOAT-style, descriptive task-level interval [−0.40,+3.78]) is not established. Over a common set of seeds, controls widened to FTDA’s parameter and multiply-accumulate budgets shifted accuracy by at most +0.35 pp against DANN (intervals including zero), where FTDA gains +2.97 pp on the same runs, so added capacity alone does not appear to explain the improvement, and a target-only reference model, which is optimistic because it is evaluated on the windows it was trained on, sits 0.70 pp above FTDA. An ablation shows that the frequency branch is useful chiefly because it enables a coupled mechanism of cross-view consistency and auxiliary source supervision (+2.02 pp for that final step) rather than concatenated features (+1.53 pp); for the frequency MMD term, no measurable benefit was detected (−0.57 pp, 95% CI [−1.28,+0.13]), and it can be removed. On a second dataset, HHAR, the cross-subject advantage does not reproduce under cross-device shift: we detect no difference from time-only adversarial baselines (paired difference +0.48 pp against DANN, 95% CI [−1.20,+2.16]; exploratory, since all device pairs share the same subjects), so the benefit of the spectral view does not appear to carry across shift types. Conclusions: Under cross-subject shift, the operative ingredient is not the spectral view as an extra set of features but the coupled consistency-and-auxiliary mechanism that the second view makes possible; the alignment objectives themselves contribute little, and no measurable benefit was detected for the frequency MMD term. FTDA realizes this view at a modest 0.31 M/7.55 MMAC footprint with no OT solver in the training loop and no hyper-parameter that requires target labels. Transfer across devices and operating environments is not evaluated here and remains open for future work. Full article
(This article belongs to the Special Issue Data-Related Challenges in Machine Learning: Theory and Application)
►▼ Show Figures

Figure 1

19 pages, 949 KB  
Article
A Lightweight Security Authentication Scheme Based on Multiple Factors in the Industrial Internet
by Hanjun Gao, Haozhong Shi, Shuai Jian and Gang Shen
Electronics 2026, 15(15), 3487; https://doi.org/10.3390/electronics15153487 - 6 Aug 2026
Viewed by 270
Abstract
The authentication mechanism can ensure the security of data in the industrial Internet. However, existing authentication schemes mainly target traditional network entities and lack specific designs for AI service security, leaving models at risk of integrity breaches and malicious tampering. To address this, [...] Read more.
The authentication mechanism can ensure the security of data in the industrial Internet. However, existing authentication schemes mainly target traditional network entities and lack specific designs for AI service security, leaving models at risk of integrity breaches and malicious tampering. To address this, this paper proposes a novel three-party authentication scheme for the AI service environment. This scheme establishes secure two-way authentication among users, gateway nodes, and AI service providers, while ensuring user anonymity and forward security, and providing comprehensive resistance to diverse attacks such as insider attacks and DDoS. By organically combining biometric verification with lightweight cryptographic operations, our scheme maintains operational efficiency while enhancing security. A detailed analysis shows that its computational overhead is approximately 0.065 milliseconds and the communication overhead is 2304 bits, significantly outperforming existing schemes. Full article
(This article belongs to the Special Issue Advances in Cryptography and Secure Protocols for Modern Information Systems)
►▼ Show Figures

Figure 1

29 pages, 3582 KB  
Article
FGC-ConvNeXt: Frequency-Guided Collaborative Modeling of Spatial and Frequency Features for Han Stone Relief Image Classification
by Hua Wei, Junxiang Diao, Wenlin Jin, Hongsheng Liu, Zhihua Diao, Lijuan Zhang and Shuang Liang
Electronics 2026, 15(15), 3486; https://doi.org/10.3390/electronics15153486 - 6 Aug 2026
Viewed by 279
Abstract
Han stone reliefs contain rich historical and artistic information, but their images often exhibit severe weathering, rubbing noise, and complex carving textures, making fine-grained classification challenging. To address the strong coupling between global semantic structures and local material textures, this paper proposes Frequency-Guided [...] Read more.
Han stone reliefs contain rich historical and artistic information, but their images often exhibit severe weathering, rubbing noise, and complex carving textures, making fine-grained classification challenging. To address the strong coupling between global semantic structures and local material textures, this paper proposes Frequency-Guided Collaborative ConvNeXt (FGC-ConvNeXt), a dual-branch spatial–frequency classification framework. The spatial branch extracts multi-scale semantic features using ConvNeXt-Tiny, while the frequency branch applies Fast Fourier Transform to model complementary structural and texture-sensitive information. A fixed frequency threshold of τ = 0.22 is used for frequency decomposition, and both branches are projected into a 512-dimensional feature space. The Spatial–Frequency Gated Fusion module dynamically adjusts their contributions at the sample level, while frequency-domain auxiliary supervision with a loss weight of λ = 0.4 prevents branch degradation during joint optimization. Experiments were conducted on a self-constructed four-class dataset containing 6368 Han stone relief images with an input resolution of 224 × 224 pixels. Over three independent runs, FGC-ConvNeXt achieved an accuracy of 97.6 ± 0.2%, a precision of 97.8 ± 0.2%, a recall of 97.4 ± 0.3%, and an F1-score of 97.6 ± 0.2%, outperforming the ConvNeXt-Tiny baseline by 4.1 percentage points in accuracy. The proposed model contains 29.8 M parameters. Visualization analyses provide qualitative support for more separated feature distributions and more concentrated activation patterns, while the robustness experiments indicate improved classification stability under the evaluated conditions. Full article
(This article belongs to the Section Artificial Intelligence)
►▼ Show Figures

Figure 1

18 pages, 780 KB  
Systematic Review
Generative AI in Manufacturing and Industrial Contexts: A Systematic Review of Applications, Challenges, and Future Directions
by Galina Ilieva and Yuliy Iliev
Electronics 2026, 15(15), 3485; https://doi.org/10.3390/electronics15153485 - 6 Aug 2026
Viewed by 617
Abstract
Generative artificial intelligence (GAI) is expanding from model-centered research into engineering and manufacturing activities, but its scope and maturity remain uneven. This PRISMA-guided bibliometric and abstract-level thematic review maps peer-reviewed industrial GAI research published from 2022 to 4 June 2026. Searches of Scopus, [...] Read more.
Generative artificial intelligence (GAI) is expanding from model-centered research into engineering and manufacturing activities, but its scope and maturity remain uneven. This PRISMA-guided bibliometric and abstract-level thematic review maps peer-reviewed industrial GAI research published from 2022 to 4 June 2026. Searches of Scopus, Web of Science, and the ACM Digital Library identified 492 records; 119 duplicates and 121 ineligible records were removed, leaving 252 studies. Keyword normalization, co-occurrence analysis, dominant and secondary thematic coding, and an abstract-reported evidence characterization were applied. The corpus shows two connected trajectories: engineering generation based on generative models for design, topology, materials, and electronics, and knowledge-intensive industrial intelligence based on large language models, retrieval-augmented generation, knowledge graphs, agents, and human–AI collaboration. Most studies report empirical or computational evaluation (72.2%), but 84.5% remain research-stage; only 0.8% indicate operational industrial evidence in their abstracts. The findings, therefore, distinguish publication activity from deployment maturity. Priority requirements for adoption include domain-grounded data, verification, manufacturability checks, traceability, cybersecurity, intellectual property protection, system integration, workforce preparation, and human accountability. This review contributes a reproducible cross-domain map, an overlap-aware synthesis, and stakeholder-specific guidance for trustworthy industrial GAI. Full article
(This article belongs to the Special Issue Generative AI and Its Transformative Potential, 2nd Edition)
►▼ Show Figures

Figure 1

29 pages, 1544 KB  
Article
NERFlow: A Workflow-Based Subsystem of FIT4NER for LLM-Assisted Medical Named Entity Recognition
by Florian Freund, Philippe Tamla, Bao Tran and Matthias Hemmje
Electronics 2026, 15(15), 3484; https://doi.org/10.3390/electronics15153484 - 6 Aug 2026
Viewed by 268
Abstract
Preparing training data for domain-specific medical Named Entity Recognition (NER) involves a trade-off between annotation quality, expert effort, and data privacy: manual annotation is costly, whereas cloud-based Large Language Models (LLMs) raise concerns about the control of sensitive clinical text. This article introduces [...] Read more.
Preparing training data for domain-specific medical Named Entity Recognition (NER) involves a trade-off between annotation quality, expert effort, and data privacy: manual annotation is costly, whereas cloud-based Large Language Models (LLMs) raise concerns about the control of sensitive clinical text. This article introduces NERFlow, a workflow-driven subsystem of the FIT4NER project whose contribution is an abstraction layer that makes rule-based, model-based, and LLM-based annotation interchangeable and comparable within one configurable workflow environment. Open-source LLMs are integrated as exchangeable annotation services, deployable locally or in cloud-agnostic infrastructures via Kubernetes, and embedded into the KM-EP knowledge management system. NERFlow was evaluated qualitatively, through a cognitive walkthrough, an IEEE 1028 technical review, and a user-centered survey with 18 participants, and quantitatively on the CRAFT corpus with seven open-source and hosted LLMs run through an identical pipeline. The results support its use as LLM-assisted pre-annotation with expert correction, with locally deployable open-source models as the more reliable basis for reproducible operation. Full article
(This article belongs to the Special Issue Formal Methods and Intelligent Systems: Trends and Advances in Theoretical and Applied Informatics)
►▼ Show Figures

Figure 1

36 pages, 1172 KB  
Article
Causal Benefit-Aware Recommendation for Personalized Learning-Path Features: A Targeting-Policy Framework with Provable Guarantees and Randomized Evaluation
by Yanfen Huang, Lin Wang, Weihua Bai, Teng Zhou and Xinyang Wang
Electronics 2026, 15(15), 3483; https://doi.org/10.3390/electronics15153483 - 6 Aug 2026
Viewed by 328
Abstract
Educational platforms increasingly personalize which AI learning-path features (adaptive homework, learner choice) each student receives. The natural correlational baseline ranks students by predicted performance—deliver the feature to those expected to do well—a heuristic that need not identify who actually benefits. We formalize feature [...] Read more.
Educational platforms increasingly personalize which AI learning-path features (adaptive homework, learner choice) each student receives. The natural correlational baseline ranks students by predicted performance—deliver the feature to those expected to do well—a heuristic that need not identify who actually benefits. We formalize feature recommendation as a causal targeting-policy problem: rank students by the estimated conditional average treatment effect (CATE) of a feature and recommend to the top of the ranking. We prove three results: (i) causal top-CATE targeting maximizes policy value at any budget and weakly dominates predictive (outcome-based) targeting, strictly when the two rankings disagree; (ii) a split-sample doubly robust evaluation of targeting quality is leakage-free (null-exact in finite samples), whereas the naive in-sample version is optimistically biased; and (iii) greedily targeting by CATE traces the optimal cost–benefit (Qini) frontier, with the deployment rule “recommend when τ^>0.” We validate the method on 17 randomized embedded experiments from the ASSISTments platform. Because the 40 held-out splits re-partition the same students, we do not treat them as independent replicates: we calibrate every headline comparison against a within-experiment permutation null and an experiment-clustered bootstrap. Under that calibrated inference, causal targeting outperforms predictive targeting for adaptive homework at every budget (permutation p≤0.005, the resolution floor of 200 replicates; Holm-corrected p≤0.040), while for learner choice the same contrast is directionally consistent but not statistically significant (permutation p=0.23–0.38; clustered p=0.42). The direction is stable in both families: no leave-one-experiment-out refit reverses its sign. Predictive targeting is nonetheless the one rule that is reliably worse than the alternatives, because it recommends the feature to high-performing students who benefit least—realized benefit falls monotonically across predicted-performance deciles (from +0.087 in the lowest to −0.030 in the highest). Against a fuller baseline suite, causal (CATE) targeting does not beat random, a simple risk-based rule (target low performers), or treating everyone. Indeed, the estimated benefit ranking is close to noise—split-half rank agreement is ρ≈0.002–0.008 and its calibration slope is 0.018, far below the ideal of 1—so the gain over predictive targeting comes from avoiding an actively harmful ordering rather than from recovering individual benefit. A fairness analysis shows why this matters: predictive targeting is regressive, concentrating feature access on high-ability students, whereas causal and risk-based targeting reverse that gradient in this corpus; no policy differentiates by neighborhood opportunity zone. Group-conditional policy values, however, are not individually distinguishable from zero once dependence across students and experiments is accounted for; what survives resampling is the allocation itself—predictive targeting directs 0.33 fewer of its recommendations to low-ability than to high-ability students (95% CI [−0.46,−0.01], experiment-clustered)—so we frame the fairness result as improved access, not established equity gains. The actionable finding is therefore narrow and specific: outcome-based targeting systematically mis-allocates learning-path features and should be replaced by some benefit-aware rule; whether that rule needs to be a learned CATE model, rather than a simple risk-based heuristic, is not established by this corpus. Full article
►▼ Show Figures

Figure 1

34 pages, 3408 KB  
Article
RapproX: An Adaptive Approximate Adder with Lookback for Efficient Edge AI via Memristive In-Memory Computing
by Lukas Rapp, Leandro Borzyk, Fabian Seiler, Nima Amirafshar and Nima TaheriNejad
Electronics 2026, 15(15), 3482; https://doi.org/10.3390/electronics15153482 - 6 Aug 2026
Viewed by 402
Abstract
As silicon scaling nears its physical limits and digital systems process ever-growing amounts of data, integrating computation directly within memory is emerging as a key strategy to overcome the constraints of conventional Von Neumann architectures. Approximate In-Memory Computation (IMC) with memristors offers a [...] Read more.
As silicon scaling nears its physical limits and digital systems process ever-growing amounts of data, integrating computation directly within memory is emerging as a key strategy to overcome the constraints of conventional Von Neumann architectures. Approximate In-Memory Computation (IMC) with memristors offers a promising path toward energy-efficient processing for data-intensive applications. Recent adaptive approximate adders exploit operand magnitude to dynamically switch between exact and approximate computation, but typically ignore carry propagation across approximation boundaries, which can significantly degrade application-level robustness. This work introduces RapproX, a family of adaptive memristive approximate adders featuring a lightweight carry lookback mechanism that approximates carry interaction between exact and approximate regions. The proposed approach improves arithmetic robustness while introducing only minimal overhead and enabling resource-efficient implementations through memristor reuse. Experimental results demonstrate that the proposed approaches achieve superior arithmetic quality compared to State-of-the-Art (SoA) memristive approximate adders. More importantly, the carry lookback mechanism translates into substantial application-level benefits. In image processing, RapproX reduces energy consumption by up to 30.9% compared to the most competitive SoA design and by 50.3% compared to exact computation while maintaining roughly 43 dB Peak Signal-to-Noise Ratio (PSNR). Across a range of machine-learning workloads, including k-means, AlexNet on MNIST, and multiple CIFAR-10 models, RapproX preserves near-exact inference accuracy for the evaluated models at low-to-moderate k and maintains the energy advantages of adaptive approximation, while SoA approximations degrade markedly under the same conditions. These simulation-based results suggest that lightweight carry-aware approximation can improve the robustness of adaptive approximate in-memory computing with only marginal hardware overhead. Full article
(This article belongs to the Special Issue Emerging Computing Paradigms for Efficient Edge AI Acceleration)
►▼ Show Figures

Figure 1

35 pages, 1124 KB  
Article
A Granularity-Centered Taxonomy of Personalized Federated Learning
by Ei Ei Nyein Chan, Sergei Chuprov, Pretom Roy Ovi and Kamrul Hasan
Electronics 2026, 15(15), 3481; https://doi.org/10.3390/electronics15153481 - 6 Aug 2026
Viewed by 517
Abstract
Personalized Federated Learning (PFL) has emerged as a key approach to address performance degradation in FL systems under heterogeneous client data. While existing surveys typically categorize PFL methods based on optimization strategies or system-level mechanisms, they often overlook a fundamental question: where is [...] Read more.
Personalized Federated Learning (PFL) has emerged as a key approach to address performance degradation in FL systems under heterogeneous client data. While existing surveys typically categorize PFL methods based on optimization strategies or system-level mechanisms, they often overlook a fundamental question: where is personalization embedded within the model architecture? In this survey, we bridge this knowledge gap and introduce a granularity-centered taxonomy that organizes PFL approaches according to the structural depth of personalization, ranging from head-layer and layer-wise adaptation to model-wise and parameter-wise customization. This novel perspective helps practitioners select appropriate personalization strategies based on model architecture, data heterogeneity, and system constraints. By analyzing representative works published between 2016 and 2026, we identify recurring design bottlenecks and highlight key opportunities for improving personalization across different granularity levels. While existing surveys focus solely on algorithms, our taxonomy provides practitioners and researchers with the informed guidance needed to design scalable and effective PFL systems. Building upon this analysis, we also formulate and discuss several unaddressed open challenges currently present in the field, highlighting specific roadblocks and unresolved questions that must be addressed to drive the future evolution of PFL methodologies. Full article
(This article belongs to the Special Issue Cybersecurity Solutions for Intelligent Systems)
►▼ Show Figures

Figure 1

24 pages, 3543 KB  
Article
A Driving-Primitive-Based Framework for Modeling the Evolution of Unprotected Left-Turn Interactions Using UAV Trajectory Data
by Yibo Xu, Amin Moeinaddini, Yichuan Peng, Yajie Zou and Shubo Wu
Electronics 2026, 15(15), 3480; https://doi.org/10.3390/electronics15153480 - 6 Aug 2026
Viewed by 272
Abstract
Unprotected left turns at urban intersections require drivers to continually regulate their driving behavior while negotiating conflicts with opposing through traffic. Existing studies have mostly examined such maneuvers through gap-acceptance decisions and surrogate conflict indicators, which cannot reflect how driving behavior changes for [...] Read more.
Unprotected left turns at urban intersections require drivers to continually regulate their driving behavior while negotiating conflicts with opposing through traffic. Existing studies have mostly examined such maneuvers through gap-acceptance decisions and surrogate conflict indicators, which cannot reflect how driving behavior changes for unprotected left-turn interaction events. To explore the evolution of unprotected left-turn interactions, this study develops a data-driven framework to decompose continuous left-turn trajectories into interpretable, variable-length driving primitives. Using high-resolution unmanned aerial vehicle trajectory data, 2490 valid left-turning and opposing-through vehicle interaction events were extracted. A Non-Homogeneous Hidden Markov Model was adopted to segment each interaction event into driving primitives. These primitives are then clustered using Time-Series K-Means with Dynamic Time Warping. The clustering results yielded six behavior patterns: cautious turning, low-speed waiting, accelerating departure, deceleration observation, intensive turning, and steady driving. These patterns were then mapped back to the temporal sequence of each interaction event to construct behavioral transition chains. The results demonstrate that an unprotected left turn typically evolves as an ordered combination of the identified behavior patterns, from waiting and observation to turning and departure, and that transitions among these patterns are associated with changing opposing-traffic conditions. Full article
(This article belongs to the Special Issue Advanced Intelligent Control and Optimization for Autonomous and Electric Vehicles)
►▼ Show Figures

Figure 1

19 pages, 5944 KB  
Article
Modeling Human–Fire–Agent Interactions for Subway Fire Evacuation: A Case Study of Lumu Metro Station in Suzhou
by Guojing Hu, Rui Qiang, Zhe Li, Weike Lu and Yinnan Yuan
Electronics 2026, 15(15), 3479; https://doi.org/10.3390/electronics15153479 - 6 Aug 2026
Viewed by 367
Abstract
Metro stations, while essential for urban transportation, pose unique evacuation challenges due to confined layouts and high densities; existing models often struggle to accurately capture individual pedestrian behaviors and the dynamic spread of fires. This study introduces a human–fire–agent interaction model designed to [...] Read more.
Metro stations, while essential for urban transportation, pose unique evacuation challenges due to confined layouts and high densities; existing models often struggle to accurately capture individual pedestrian behaviors and the dynamic spread of fires. This study introduces a human–fire–agent interaction model designed to enhance the understanding and simulation of critical interactions among pedestrians, fire dynamics, and the underground environment of a metro station. The model integrates social force modeling and fluid dynamics to accurately represent pedestrian behavior and fire spread, for a more complete analysis of evacuation scenarios. Using Lumu Station in Suzhou as a case study, this study develops a detailed simulation framework implemented in an integrated PyroSim-Python-AnyLogic platform to model the evacuation process. The framework is employed to evaluate the effectiveness of turnstile reversal strategies—an approach that involves temporarily reversing the direction of turnstiles to facilitate faster evacuation during emergencies. Beyond mitigation, this study extends to the preparedness phase by functioning as a high-fidelity digital twin. It enables immersive “Serious Game” training and provides a quantitative tool for railway managers, decision-makers, and engineers to optimize operating procedures and performance-based station designs. Full article
(This article belongs to the Topic Data-Driven Optimization for Smart Urban Mobility)
►▼ Show Figures

Figure 1

28 pages, 6390 KB  
Article
A Kubernetes-Deployed Tamper-Evident Media-Evidence Provenance Pipeline with Hybrid Blockchain/IPFS Anchoring and Queue-Mediated Ingress
by Haoliang Wang, Zarina Shukur, Khairul Akram Zainol Ariffin and Lili Wang
Electronics 2026, 15(15), 3478; https://doi.org/10.3390/electronics15153478 - 6 Aug 2026
Viewed by 380
Abstract
High-stakes online assessment produces suspicious-event records, yet storage placement and burst admission remain insufficiently characterized. This article presents a Kubernetes-deployed provenance pipeline integrating Hyperledger Fabric, IPFS, SHA-256 verification, and RabbitMQ ingress. Media objects are retained in IPFS, while compact semantics, CIDs, and verification [...] Read more.
High-stakes online assessment produces suspicious-event records, yet storage placement and burst admission remain insufficiently characterized. This article presents a Kubernetes-deployed provenance pipeline integrating Hyperledger Fabric, IPFS, SHA-256 verification, and RabbitMQ ingress. Media objects are retained in IPFS, while compact semantics, CIDs, and verification anchors are committed to Fabric. At 5 TPS, five 300-transaction runs completed under both direct large-payload and compact-anchor ledger conditions. At 20 TPS, compact anchoring of a pre-retained IPFS object sustained 19.40 ± 0.00 TPS with 0.22 ± 0.00 s mean ledger latency; direct large-payload submission achieved 14.20 ± 0.40 TPS with 31.62 ± 3.35 s latency. A write-load sweep identified 100 TPS as the transition point; higher loads were delay-dominated. Three 1000-VU PTS runs reduced mean response time from 4.76 ± 0.78 s to 2.40 ± 0.02 s, with similar failure rates. In three 30-message consumer-enabled runs, all 90 messages reached IPFS retention, Fabric commit, query visibility, and manual acknowledgement without retry, producer failure, or dead-letter outcome. Mean queue-drain and total completion times were 61.80 ± 0.71 s and 62.99 ± 0.15 s. The results isolate Fabric transaction-path payload cost and distinguish queue acceptance from downstream completion. Full article
(This article belongs to the Section Computer Science & Engineering)
►▼ Show Figures

Figure 1

30 pages, 1345 KB  
Review
RGB-Only rPPG for Contactless SpO2 Estimation: A Critical Review of Optical Constraints, Calibration Dependence, and Clinical Validation
by Sajana Islam Moon and Nam Kyu Kwon
Electronics 2026, 15(15), 3477; https://doi.org/10.3390/electronics15153477 - 6 Aug 2026
Viewed by 526
Abstract
This review focuses exclusively on RGB camera-based remote photoplethysmography (rPPG) for contactless estimation of peripheral oxygen saturation (SpO2). Unlike heart-rate estimation, which mainly depends on recovering pulse periodicity, SpO2 estimation requires oxygenation-sensitive optical contrast, stable channel-wise AC/DC information, and calibration [...] Read more.
This review focuses exclusively on RGB camera-based remote photoplethysmography (rPPG) for contactless estimation of peripheral oxygen saturation (SpO2). Unlike heart-rate estimation, which mainly depends on recovering pulse periodicity, SpO2 estimation requires oxygenation-sensitive optical contrast, stable channel-wise AC/DC information, and calibration and validation against a physiological reference. The reviewed literature is organized around the main stages of an RGB-only SpO2 pipeline. Existing methods are further discussed as ratio-based or Beer–Lambert-inspired approaches, classical regression models, deep learning and spatiotemporal methods, and hybrid or physics-informed designs. This critical review shows that RGB-based systems can estimate oxygenation-related signals in controlled settings. However, current evidence is limited by device-dependent camera responses, narrow oxygen saturation ranges, weak validation, limited demographic reporting, and insufficient hypoxemia testing. The main conclusion of this review is that RGB-only SpO2 estimation is useful under controlled conditions, but it should not be treated as an absolute, clinically interpretable SpO2 measurement method without stronger validation. Future work should focus on wider SpO2 distributions, transparent calibration procedures, cross-device testing protocols, subgroup reporting, and a clear distinction between oxygenation trend monitoring and clinically interpretable absolute SpO2 measurement. Full article
(This article belongs to the Special Issue Advanced Technologies in Signal and Image Processing)
►▼ Show Figures

Figure 1

37 pages, 17322 KB  
Article
Frequency-Domain Multi-Objective Decoupled Control of a Harmonic Impedance Measurement Device for an Energy-Storage-Integrated Grid-Connected System
by Binghan Sun, Mingli Wu, Qiujiang Liu, Liran Wu, Tingting He, Muchen Wang and Jingjing Ye
Electronics 2026, 15(15), 3476; https://doi.org/10.3390/electronics15153476 - 6 Aug 2026
Viewed by 1377
Abstract
Energy-storage-integrated renewable energy stations contain both grid-following and grid-forming converters, which makes their wideband impedance behaviour difficult to predict. This paper studies a field-oriented harmonic impedance measurement method for such stations. A three-phase cascaded H-bridge converter is used to develop a harmonic impedance [...] Read more.
Energy-storage-integrated renewable energy stations contain both grid-following and grid-forming converters, which makes their wideband impedance behaviour difficult to predict. This paper studies a field-oriented harmonic impedance measurement method for such stations. A three-phase cascaded H-bridge converter is used to develop a harmonic impedance measurement device connected to the 35 kV bus. For the measured system, a generalized Norton equivalent model is established for the grid-following part, while a generalized Thevenin equivalent model is derived for the grid-forming energy-storage part. A frequency-domain multi-objective target-oriented decoupled control strategy is then designed for the measurement device. The strategy assigns fundamental power synchronisation, harmonic disturbance injection, and submodule capacitor voltage balancing to different frequency components. This design prevents the fundamental current loop from suppressing the injected harmonic current. Simulation studies are carried out under four conditions, including no harmonic injection and 5th-, 25th-, and 99th-order harmonic injections. The baseline case shows no clear commanded-frequency current component. Under harmonic commands, the device injects the corresponding current components while maintaining stable multilevel bridge-port voltage. The device also demonstrates robust impedance-identification performance under grid background harmonics, measurement noise, and equivalent-impedance variations. The results indicate that the proposed device has application capability for wideband impedance measurement and stability analysis in energy-storage-integrated grid-connected systems. Full article
(This article belongs to the Special Issue Electrical Energy Storage Systems and Grid Services)
►▼ Show Figures

Figure 1

31 pages, 1747 KB  
Review
A Comprehensive Review About Human Digital Twins and AI-Powered Wearables for the Oil and Gas Industry
by Saul Davila-Gonzalez and Sergio Martin
Electronics 2026, 15(15), 3475; https://doi.org/10.3390/electronics15153475 - 6 Aug 2026
Viewed by 431
Abstract
Construction activities within the Oil and Gas Industry present many challenges and inherent dangers for workers. Many current solutions available in the market lack real-time insights and predictive capabilities for proactive decision-making and operational safety. This industry demands robust solutions to enhance safety, [...] Read more.
Construction activities within the Oil and Gas Industry present many challenges and inherent dangers for workers. Many current solutions available in the market lack real-time insights and predictive capabilities for proactive decision-making and operational safety. This industry demands robust solutions to enhance safety, improve security, and increase productivity, and this is where Human Digital Twins (HDTs), Wearables, and Artificial Intelligence (AI) play an important role. HDTs aim to create a dynamic digital replica of workers, integrating data from wearables, IIoT sensors, information systems, and any other data source available, enabling continuous monitoring of physiological and cognitive conditions. Complementing HDTs, AI-powered wearables collect enriched data from embedded sensors, such as heart rate, blood oxygen, or fatigue, and integrate deep learning algorithms to predict potential incidents directly on the edge before they happen, thanks to their high computing capabilities. This review represents the state-of-the-art for HDTs and AI-powered wearables for the Oil and Gas industry. It explores current technological developments, their applications, the challenges and limitations of deploying them into a real-world project, and also synthesizes current research and industry practices, highlighting their potential in human health and safety. Additionally, it identifies future research directions to overcome existing barriers. Full article
(This article belongs to the Special Issue Industrial Internet of Things and Smart Technologies in Industry 4.0/5.0: Advanced Monitoring and Automation Trends and Applications)
►▼ Show Figures

Figure 1

19 pages, 3780 KB  
Article
The Impact of Covariates on Zero-Shot Building Energy Forecasting Using Chronos-2 Foundation Model
by Amedeo Buonanno, Salvatore Fabozzi, Maria Valenti and Giorgio Graditi
Electronics 2026, 15(15), 3474; https://doi.org/10.3390/electronics15153474 - 6 Aug 2026
Cited by 1 | Viewed by 354
Abstract
Foundation models for time series forecasting have recently been applied to energy prediction tasks, where they can produce accurate forecasts without task-specific training. This study investigates the impact of two types of covariates, meteorological variables and calendar-based day type indicators, on the forecasting [...] Read more.
Foundation models for time series forecasting have recently been applied to energy prediction tasks, where they can produce accurate forecasts without task-specific training. This study investigates the impact of two types of covariates, meteorological variables and calendar-based day type indicators, on the forecasting performance of Chronos-2, a state-of-the-art foundation model, in building energy consumption prediction. Using real-world monitoring data from two non-residential buildings at the ENEA Research Centre in Portici, Italy, we systematically evaluate seven configurations combining past and future covariates across multiple observation window lengths (7–28 days). Future meteorological covariates are derived from historical weather forecasts rather than observed weather data, ensuring that the evaluation reflects realistic operational forecasting conditions. The results show that incorporating day type indicators as both past and future covariates consistently delivers the highest forecasting accuracy, reducing CV-RMSE from 14.58% for the covariate-free baseline to 10.41% with a 28-day observation window. A day-stratified analysis further reveals that these improvements are concentrated on regime transition days, for which recent load history alone provides limited information about the operating conditions of the day being forecast. By contrast, meteorological variables, whether obtained from weather forecasts or historical observations, yield only marginal performance gains, suggesting that calendar-driven operational schedules are the primary determinants of energy demand in the buildings considered. These findings provide practical guidance for deploying foundation models in real-world energy building management systems and show that covariate selection is a key determinant of forecasting performance. Full article
(This article belongs to the Special Issue Advanced Technologies in Power Electronics)
►▼ Show Figures

Figure 1

19 pages, 2436 KB  
Article
Two-Dimensional DOA Estimation Based on Dual-Branch CNN
by Fangyu Liu, Guimei Zheng, Yuwei Song, Yujie Bai and He Zheng
Electronics 2026, 15(15), 3473; https://doi.org/10.3390/electronics15153473 - 6 Aug 2026
Viewed by 342
Abstract
Direction-of-arrival (DOA) estimation is a core research topic in array signal processing, and two-dimensional (2-D) DOA estimation can jointly acquire the azimuth and elevation of incoming signals, which bears great practical value. Traditional subspace and sparse reconstruction algorithms are plagued by heavy computation [...] Read more.
Direction-of-arrival (DOA) estimation is a core research topic in array signal processing, and two-dimensional (2-D) DOA estimation can jointly acquire the azimuth and elevation of incoming signals, which bears great practical value. Traditional subspace and sparse reconstruction algorithms are plagued by heavy computation and deteriorated accuracy under imperfect array manifolds, low signal-to-noise ratios (SNRs) and insufficient snapshots. To enhance estimation robustness and inference speed simultaneously, this paper presents a dual-branch convolutional neural network (CNN) for 2-D DOA estimation based on uniform rectangular arrays. The network takes the sample covariance matrix of array received data as input. A shared feature encoder with residual blocks and channel-attention modules extracts common spatial features, followed by two prediction heads with independent parameters for elevation and azimuth estimation. Because each branch has a 61-dimensional output while two sources may be simultaneously present, the angle estimation is formulated as multi-label classification using sigmoid outputs and weighted binary cross-entropy. Simulations covering diverse SNRs, snapshot counts, angular intervals and off-grid cases verify that the proposed network obtains smaller root mean square errors than methods with multiple signal classification (MUSIC), estimation of signal parameters via rotational invariance techniques (ESPRIT) and ordinary CNN methods, with millisecond-level inference latency. This framework offers an efficient, high-precision real-time 2-D DOA estimation scheme for complicated electromagnetic scenes. Full article
(This article belongs to the Section Circuit and Signal Processing)
►▼ Show Figures

Figure 1

27 pages, 10745 KB  
Article
A Wide-Frequency Stability Characteristic Domain Method for Small-Signal Stability Analysis of Grid-Forming Direct-Drive Wind Turbines
by Huajia Wang, Yan Zhang, Wenjun Cao, Fan Xiao, Danwen Yu, Qingqing Zhang and Wenjun Peng
Electronics 2026, 15(15), 3472; https://doi.org/10.3390/electronics15153472 - 6 Aug 2026
Viewed by 334
Abstract
To accurately evaluate the small-signal stability of grid-forming (GFM) direct-drive wind turbines over time-varying operating conditions and broad frequency ranges, this paper proposes a wide-frequency stability characteristic domain method. Unlike grid-following turbines, GFM control relies on power-loop-driven self-synchronization rather than a phase-locked loop, [...] Read more.
To accurately evaluate the small-signal stability of grid-forming (GFM) direct-drive wind turbines over time-varying operating conditions and broad frequency ranges, this paper proposes a wide-frequency stability characteristic domain method. Unlike grid-following turbines, GFM control relies on power-loop-driven self-synchronization rather than a phase-locked loop, which introduces multi-time-scale couplings among the virtual power angle, inner control loops, digital control delay, and weak-grid impedance. A parametric admittance model embedded with continuous operating-point variables is therefore established to characterize the converter wide-frequency dynamics. By combining physical power-transmission constraints with closed-loop pole-based small-signal stability criteria, a wide-frequency stability characteristic domain is constructed to map the stable, unstable, and physically infeasible regions over the continuous operating space. The boundary evolution under varying grid impedances and VSG control parameters is further analyzed. The experimental cases verify the operating-region transition predicted by the proposed domain. In addition, the supplementary high-frequency pole analysis shows that, when digital delay is considered, another high-frequency mode may become weakly damped or unstable under different grid-impedance conditions. These results indicate that the proposed framework captures both the low-frequency boundary-crossing behavior observed in the experiments and the potential high-frequency instability risk introduced by converter digital dynamics. Full article
►▼ Show Figures

Figure 1

21 pages, 10378 KB  
Article
Improved YOLOv11 with Information Integration Attention for Multi-Organ Apple Disease Detection Throughout the Whole Growth Period
by Yuanyuan Zhang, Jiya Tian and Duanyang Zhang
Electronics 2026, 15(15), 3471; https://doi.org/10.3390/electronics15153471 - 6 Aug 2026
Viewed by 257
Abstract
Manual visual diagnosis of apple diseases suffers from low efficiency, strong subjectivity and poor scalability for large commercial orchards. Existing research mainly targets diseases on single plant organs, whereas full-growth-cycle detection has to cope with extreme multi-scale differences among lesions. For instance, Valsa [...] Read more.
Manual visual diagnosis of apple diseases suffers from low efficiency, strong subjectivity and poor scalability for large commercial orchards. Existing research mainly targets diseases on single plant organs, whereas full-growth-cycle detection has to cope with extreme multi-scale differences among lesions. For instance, Valsa canker on tree trunks leads to extensive cortical necrosis, while early-stage anthracnose on fruits appears as tiny spots spanning only a few pixels. These significant scale gaps necessitate robust spatial feature aggregation and anti-noise ability to resist complex background interference. Aiming to achieve rapid and precise detection of diseases on multiple apple organs including leaves, fruits, trunks and branches, this work presents an enhanced YOLOv11 model equipped with the Information Integration Attention (IIA) module. The IIA module is embedded into the key fusion layers of the backbone and neck networks. It strengthens the extraction of fine-grained lesion features, recovers spatial location information via a bidirectional attention mechanism, and suppresses noise induced by uneven lighting and intricate backgrounds. To guarantee stable convergence on low-resource computing devices, a tailored training scheme is designed. Experimental results on a seven-category dataset with 7406 images demonstrate that YOLOv11-IIA reaches a precision of 0.763, a recall of 0.819, mAP@50 of 0.857 and mAP@50-95 of 0.699, which achieves clear performance improvements over the original YOLOv11 (mAP@50 improved from 0.485 to 0.857) and other attention-augmented detectors. The model operates stably on an NVIDIA GTX 1050 4GB GPU with an inference speed of 16 FPS for 640 × 640 input images; comprehensive quantitative computational metrics including parameter count, FLOPs and memory consumption will be fully measured in subsequent UAV deployment experiments. The proposed method provides a reliable technical reference for intelligent apple disease monitoring in smart orchard systems. Full article
►▼ Show Figures

Figure 1

33 pages, 10301 KB  
Article
An Explainable Multi-Task Deep Learning Framework for Service-Gap Identification and Emerging Urban Prediction
by Abdulelah Algosaibi
Electronics 2026, 15(15), 3470; https://doi.org/10.3390/electronics15153470 - 6 Aug 2026
Viewed by 350
Abstract
Rapid urbanization has intensified pressure on public services, infrastructure systems, and spatial equity, highlighting the need for integrated approaches that jointly assess service deficits and emerging urban growth. This study proposes an explainable urban analytics framework for identifying service-gap risk and emerging urban [...] Read more.
Rapid urbanization has intensified pressure on public services, infrastructure systems, and spatial equity, highlighting the need for integrated approaches that jointly assess service deficits and emerging urban growth. This study proposes an explainable urban analytics framework for identifying service-gap risk and emerging urban patterns using harmonized spatial, service, population, and digital-readiness indicators. The framework integrates an MCP-enabled data harmonization pipeline, composite service-availability and population-adjusted service-stress features, and a Dual-Head MLP architecture that supports shared representation learning across two related prediction tasks. SHAP-based explainability is employed to interpret the relative contribution of service, stress, digital-readiness, and spatial-context features. The framework was evaluated using Saudi district-level data and proxy-based external city datasets to assess cross-city transferability. Compared with classical machine-learning and single-task neural baselines, the Dual-Head MLP demonstrated stronger task-wise predictive performance, while the external evaluation indicated stable but context-dependent transfer across heterogeneous urban settings. The findings suggest that the proposed framework can support urban planners and municipal decision-makers in prioritizing infrastructure investment, identifying underresourced areas, and interpreting early urban transformation patterns. However, the outputs should be regarded as proxy-based decision-support indicators rather than official administrative measures of service adequacy. Full article
(This article belongs to the Special Issue Data-Driven AI Approaches with Applications in Social Network, Media Analytics and Smart Cities)
►▼ Show Figures

Figure 1

12 pages, 4139 KB  
Article
A Capacitor-Less Low Dropout Regulator with Wide Input Range and High Power-Supply Rejection Ratio
by Chunlai Wang, Chen Yang, Yue Pei, Zhanqi Liu and Weihua Han
Electronics 2026, 15(15), 3469; https://doi.org/10.3390/electronics15153469 - 5 Aug 2026
Viewed by 324
Abstract
This paper presents a capacitor-less low-dropout regulator (LDO) with a wide input-voltage range and high power-supply rejection ratio (PSRR), designed and laid out in a 0.18 μm CMOS process. The design incorporates a pre-regulation circuit to suppress power-supply noise. The proposed LDO supports [...] Read more.
This paper presents a capacitor-less low-dropout regulator (LDO) with a wide input-voltage range and high power-supply rejection ratio (PSRR), designed and laid out in a 0.18 μm CMOS process. The design incorporates a pre-regulation circuit to suppress power-supply noise. The proposed LDO supports an input-voltage range from 3.3 V to 40 V while maintaining an output voltage of 3.3 V. The integrated output noise from 10 Hz to 100 kHz is 4.3 nV2. The proposed LDO achieves a PSRR of −109 dB at 1 kHz under a maximum load current of 100 mA, representing a 20 dB improvement compared with the same LDO without the pre-regulation circuit. Full article
(This article belongs to the Special Issue Innovative Applications of Semiconductor Materials and Devices)
►▼ Show Figures

Figure 1

54 pages, 4342 KB  
Article
SGC: Soft Gradient Collaboration for Backdoor Attacks in Self-Supervised Distillation
by Da Xiao, Tongke Fan, Ning Dong, Jianfei Tong and Yihong Zhang
Electronics 2026, 15(15), 3468; https://doi.org/10.3390/electronics15153468 - 5 Aug 2026
Viewed by 297
Abstract
Self-supervised knowledge distillation is widely used to compress reusable encoders, but an untrusted distillation implementation can itself become an attack surface. We study an algorithm-level threat in which the teacher encoder and user-visible distillation dataset remain unchanged, while malicious code internally generates trigger-bearing [...] Read more.
Self-supervised knowledge distillation is widely used to compress reusable encoders, but an untrusted distillation implementation can itself become an attack surface. We study an algorithm-level threat in which the teacher encoder and user-visible distillation dataset remain unchanged, while malicious code internally generates trigger-bearing views and optimizes an additional backdoor objective. To instantiate this threat, we propose soft gradient collaboration (SGC), which combines distribution-alignment-based distillation, target-representation-based backdoor design, and conflict-avoidance gradient collaboration to reduce interference with benign representation transfer while embedding a trigger-to-target association in the student encoder. Experiments on CIFAR-10 and STL-10 show that SGC maintains competitive downstream accuracy and effective non-target attack success. Quantitative CKA, feature-distribution, and class-structure analyses further indicate that SGC retains clean representations closer to benign distillation than fixed scalarization or removal of distribution alignment. Its no-defense attack success is not the highest among the compared attacks; instead, its main empirical advantage is stronger residual attack persistence after MIMIC, MKD, and SSLDefender. Under SSLDefender, SGC retains 9.12% non-target ASR on CIFAR-10 and 9.06% on STL-10, the highest residual values among the compared attacks. Additional experiments with a compact ResNet-18 student, multiple target classes and trigger configurations, and a supplemental CIFAR-100 setting broaden the empirical evaluation across student capacity, target semantics, trigger configurations, and label-space complexity. These results show that security assessment of self-supervised distillation should include executable training logic in addition to model weights and visible data. The concealment considered here is limited to dataset-only inspection and clean-output validation; SGC is not claimed to evade source-code auditing, runtime data-flow monitoring, or training-log inspection. Full article
(This article belongs to the Special Issue AI-Powered Cyber Security and Protection)
►▼ Show Figures

Figure 1

22 pages, 3850 KB  
Article
ESMP: Exploring Efficient and Stable Multicast on Multiple Communication Paths
by Xin Dong, Qiuling Yang and Deshun Li
Electronics 2026, 15(15), 3467; https://doi.org/10.3390/electronics15153467 - 5 Aug 2026
Viewed by 231
Abstract
Modern communication networks may provide several heterogeneous links between the same pair of devices, including Wi-Fi, 5G, Bluetooth, and SparkLink. Existing multicast schemes often use simple-graph abstractions and therefore cannot distinguish these parallel links. We present ESMP, a multi-graph-based heuristic framework for [...] Read more.
Modern communication networks may provide several heterogeneous links between the same pair of devices, including Wi-Fi, 5G, Bluetooth, and SparkLink. Existing multicast schemes often use simple-graph abstractions and therefore cannot distinguish these parallel links. We present ESMP, a multi-graph-based heuristic framework for efficient and stable multicast construction over heterogeneous parallel communication links. ESMP represents parallel channels as edges with delay and stability attributes. We show that an aggregate-edge-delay-constrained decision variant of the formulation is NP-hard. The framework includes six polynomial-time heuristics: delay-based DMA and DSMA, stability-based SMA and SDMA, and stability-delay-ratio-based RMA and MRMA. Each algorithm derives a metric-specific graph from the original multi-graph and constructs a tree according to its delay-stability preference. We also develop local adjustment strategies for vertex joins, vertex exits, and link dynamics. Experiments on connected synthetic multi-graphs reveal distinct metric preferences. Delay-oriented methods reduce delay, stability-oriented methods improve stability, and ratio-based methods provide stability-aware trade-offs at relatively low delay. In particular, RMA favors low delay, whereas MRMA uses pair-level average stability-delay information and shows comparatively favorable stability preservation and tree compactness in the evaluated scenarios. These findings characterize heuristic behavior in the evaluated synthetic settings and do not establish general optimality. Full article
►▼ Show Figures

Figure 1

16 pages, 1078 KB  
Article
Efficiency Analysis of Predictive Current Control of a Multiphase Drive
by Esteban Marsal, Francisco Colodro and Antonio Visioli
Electronics 2026, 15(15), 3466; https://doi.org/10.3390/electronics15153466 - 5 Aug 2026
Viewed by 244
Abstract
In Finite Control Set Model Predictive Control for multiphase drives, weighting factors govern the relative priority among the multiple objectives included in the cost function. Although their influence on control performance, harmonic characteristics, and switching behavior has been extensively studied, their impact on [...] Read more.
In Finite Control Set Model Predictive Control for multiphase drives, weighting factors govern the relative priority among the multiple objectives included in the cost function. Although their influence on control performance, harmonic characteristics, and switching behavior has been extensively studied, their impact on energy efficiency remains largely unexplored. This paper addresses this gap through a comprehensive experimental assessment of the influence of cost function weighting factors on the energy efficiency of a five-phase induction motor drive. A dedicated laboratory test bench is used to evaluate drive efficiency across a broad range of operating conditions by systematically varying mechanical speed, load level and weighting factor combinations. The results reveal that cost function weighting factor tuning leads to efficiency variations of up to 7.6 percentage points depending on the operating regime. Furthermore, the study identifies an optimal parameter region that provides consistently high efficiency across the tested conditions. Full article
(This article belongs to the Special Issue Advanced Control Techniques for Power Converter and Drives, 2nd Edition)
►▼ Show Figures

Figure 1

26 pages, 3205 KB  
Article
Improved DPT-Hybrid for Monocular Depth Estimation with Geometry-Enhanced Encoding and Structure-Aware Gated Fusion
by Wei Liu, Shilei Hu, Yi Qin, Shengkai Hong and Dehua Zhang
Electronics 2026, 15(15), 3465; https://doi.org/10.3390/electronics15153465 - 5 Aug 2026
Viewed by 348
Abstract
Monocular depth estimation aims to recover dense 3D scene geometry from a single RGB image and plays an important role in autonomous driving, robotic perception, augmented reality, and 3D reconstruction. Although Transformer-based dense prediction models have achieved strong performance, existing DPT-Hybrid frameworks still [...] Read more.
Monocular depth estimation aims to recover dense 3D scene geometry from a single RGB image and plays an important role in autonomous driving, robotic perception, augmented reality, and 3D reconstruction. Although Transformer-based dense prediction models have achieved strong performance, existing DPT-Hybrid frameworks still suffer from three limitations: insufficient local geometric modeling in shallow stages, inadequate cross-scale fusion for preserving fine structures, and training objectives that only weakly constrain structural consistency. To address these issues, we propose a structure-aware enhanced DPT-Hybrid framework. First, a geometry-enhanced encoder introduces lightweight depth-wise separable convolution branches into shallow Transformer stages to better capture local edge and texture cues while preserving global contextual modeling. Second, a Structure-Aware Cross-Scale Gated Attention Fusion (S-GAF) module is proposed to improve decoder-side feature aggregation by jointly modeling channel-wise and spatial importance with an auxiliary RGB-gradient input. Third, joint structure–geometric consistency loss combines scale-invariant logarithmic loss, gradient consistency loss, and edge-focused loss to improve pixel-level accuracy, geometric plausibility, and boundary sharpness. Experiments on NYUv2 and KITTI demonstrate that the proposed method achieves lower single-run error metrics than the controlled DPT-Hybrid baseline under the evaluated settings. On NYUv2, our method achieves an absolute relative error (AbsRel) of 0.099 and an RMSE of 0.334. On KITTI, it achieves an AbsRel of 0.058 and an RMSE of 2.455. The proposed method introduces only modest additional complexity while producing more accurate and structurally sharper depth predictions. Full article
(This article belongs to the Section Computer Science & Engineering)
►▼ Show Figures

Figure 1

24 pages, 1934 KB  
Article
Advanced Adaptive Scheduling for Autonomous Driving in Beyond-5G/6G Networks
by Athanasios Kanavos, Gerasimos Papanikolaou-Ntais and Alexandros Kaloxylos
Electronics 2026, 15(15), 3464; https://doi.org/10.3390/electronics15153464 - 5 Aug 2026
Viewed by 340
Abstract
The shift toward autonomous driving is steadily reducing the need for human intervention in vehicular operations, but sustaining this shift depends on 6G networks that can reliably multiplex diverse, delay-critical services under deterministic latency, throughput, and reliability constraints. Static over-provisioning can meet the [...] Read more.
The shift toward autonomous driving is steadily reducing the need for human intervention in vehicular operations, but sustaining this shift depends on 6G networks that can reliably multiplex diverse, delay-critical services under deterministic latency, throughput, and reliability constraints. Static over-provisioning can meet the stringent Quality of Service (QoS) requirements of critical traffic, but at the cost of resource starvation for co-existing non-critical services, and it degrades further under adverse channel conditions or network congestion where spectrum must be used efficiently. This motivates dynamic, service-aware scheduling as a core requirement for multi-service 6G vehicular architectures. This paper presents SOVANET+, an extended scheduling technique that jointly accounts for service criticality (critical vs. non-critical), network load, and wireless link quality to allocate resources adaptively across coexisting Vehicle-to-Everything (V2X) services. We evaluate SOVANET+ through extensive simulations of a congested single-cell urban-grid deployment supporting delay-critical automated driving services. Results show that SOVANET+ achieves lower latency and jitter, higher throughput, and improved uplink and downlink reliability compared to existing scheduling approaches, while scaling effectively to large numbers of connected autonomous vehicles, supporting its viability for next-generation intelligent transportation systems. Full article
(This article belongs to the Special Issue Advances in 6G Wireless Communication Technologies)
►▼ Show Figures

Figure 1

25 pages, 1838 KB  
Article
Prescribed Performance Control for Electro-Hydrostatic Erecting System Based on Dual-RISE Scheme
by Weilin Zhu, Xiaowei Yang, Xiaochuan Yu and Jianyong Yao
Electronics 2026, 15(15), 3463; https://doi.org/10.3390/electronics15153463 - 5 Aug 2026
Viewed by 283
Abstract
Unmodeled uncertainties, such as friction and stage-change collision of the hydraulic cylinder, along with system disturbances, exist in the multi-link erecting system and impede high-precision erecting angle tracking. To tackle these challenges, this study develops a novel control framework characterized by asymptotic prescribed [...] Read more.
Unmodeled uncertainties, such as friction and stage-change collision of the hydraulic cylinder, along with system disturbances, exist in the multi-link erecting system and impede high-precision erecting angle tracking. To tackle these challenges, this study develops a novel control framework characterized by asymptotic prescribed performance based on a distributed dual robust integral of the sign of the error (Dual-RISE) for the electro-hydrostatic multi-link erecting system. First, a precise system model is established by integrating complex multi-link kinematics with the pressure-flow dynamics of the two-stage hydraulic cylinder. A prescribed performance function (PPF) and nonlinear error transformation are then introduced to strictly constrain the tracking error within predefined transient and steady-state boundaries. The proposed framework integrates a distributed dual-loop RISE architecture to simultaneously reject matched and unmatched uncertainties, mathematically enforcing semi-global asymptotic convergence of the tracking error to zero without requiring infinite high-gain feedback. Comparative experiments with Dual-RISE and VFPI controllers demonstrate superior tracking accuracy and boundary protection under different erecting conditions. Full article
(This article belongs to the Special Issue Robust Control of Dynamic Systems)
►▼ Show Figures

Figure 1

29 pages, 722 KB  
Article
TMCAS: Efficient Large Language Model-Assisted Topic Modeling for Civil Aviation Safety Reports
by Xiangge Li, Haofeng Wang, Xiuting Zhou, Yan Ren, Zhi Tian, Weidong Liang and Gengsong Wang
Electronics 2026, 15(15), 3462; https://doi.org/10.3390/electronics15153462 - 5 Aug 2026
Viewed by 295
Abstract
Voluntary safety reports provide valuable information for identifying potential risks and improving safety management in civil aviation. However, these reports are often large in volume, unstructured in format, and rich in domain-specific terminology, making manual analysis costly, inefficient, and difficult to scale. To [...] Read more.
Voluntary safety reports provide valuable information for identifying potential risks and improving safety management in civil aviation. However, these reports are often large in volume, unstructured in format, and rich in domain-specific terminology, making manual analysis costly, inefficient, and difficult to scale. To address these challenges, this paper proposes TMCAS, an efficient large language model-assisted topic modeling framework for civil aviation safety reports. The proposed framework combines domain-adapted text embeddings, density-based clustering, representative sampling, noise repair, and large language model-based topic generation. Specifically, a contrastive learning-based fine-tuning strategy is introduced to enhance the semantic representation of aviation safety texts. An HDBSCAN-based clustering and sampling mechanism is then designed to select representative reports and reduce the computational cost of large language model inference, while a noise-repair strategy is used to improve topic coverage. Finally, large language models are employed to generate interpretable sentence-level topic labels and descriptions. Experiments demonstrate that TMCAS achieves superior clustering and interpretability while substantially reducing inference cost compared with document-wise LLM baselines. Full article
(This article belongs to the Section Artificial Intelligence)
►▼ Show Figures

Figure 1

Show export options expand_more Show export options expand_less
Select all
Export citation of selected articles as:
 
Displaying articles 1-252
Previous Issue
Next Issue
Back to TopTop