Next Issue
Volume 14, July-1
Previous Issue
Volume 14, June-1
 
 

Electronics, Volume 14, Issue 12 (June-2 2025) – 192 articles

Cover Story (view full-size image): MultiAVSR presents a supervised multi-task framework that trains a single conformer encoder–transformer decoder network to perform audio, visual, and audio–visual speech recognition simultaneously. A shared encoder optimized with both CTC and attention objectives fosters rich cross-modal representations while keeping the model lightweight. Despite using modest computational resources and public training data. this approach achieves state-of-the-art peformance , significantly improving robustness to real-world data and reducing reliance on external language models. This framework represents notable advancment in fast, deployable, and privacy-preserving multimodal interfaces. 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
Select all
Export citation of selected articles as:
28 pages, 8607 KiB  
Article
Analysis of Grid-Connected Damping Characteristics of Virtual Synchronous Generator and Improvement Strategies
by Xudong Cao, Ruogu Zhang, Jun Li, Li Ji, Xueliang Wei, Jile Geng and Bowen Li
Electronics 2025, 14(12), 2501; https://doi.org/10.3390/electronics14122501 - 19 Jun 2025
Viewed by 173
Abstract
Focused on the contradiction between the steady-state error of active power and the dynamic oscillation caused by the virtual damping characteristics of the virtual synchronous generator (VSG) under disturbances during grid-connected operation, this article proposes an adaptive virtual inertia regulation and compensation method [...] Read more.
Focused on the contradiction between the steady-state error of active power and the dynamic oscillation caused by the virtual damping characteristics of the virtual synchronous generator (VSG) under disturbances during grid-connected operation, this article proposes an adaptive virtual inertia regulation and compensation method (PFFCVSG_AJ) based on an active power differential feedforward compensation strategy (PFFCVSG). Firstly, this article presents the working and control principles of VSG, analyzing its control mechanisms through a small-signal model. Models for VSG’s active power, reactive power, and virtual impedance components are established, with particular focus on the impact of the damping coefficient on active power regulation. Based on the PFFCVSG, an adaptive virtual inertia adjustment method is introduced to resolve the inherent inertia deficiency in PFFCVSG control, the influence of the moment of inertia on PFFCVSG is theoretically analyzed, and a dynamic adjustment mechanism for moment of inertia is developed based on the rate of change in frequency (RoCoF). Finally, simulation validation using MATLAB/Simulink (MathWorks, R2022b, Natick, MA, USA) demonstrates that the proposed PFFCVSG_AJ strategy effectively eliminates active power steady-state deviation, suppresses active power dynamic oscillation, and mitigates the frequency overshoot issue prevalent in traditional PFFCVSG. Experimental verification is conducted via a TMS320F28378DPTPS-based control platform, confirming the algorithm’s effectiveness under sudden load variations, and that the power quality of the power grid is not affected under the premise of efficient grid connection. Full article
(This article belongs to the Special Issue New Trends in Power Electronics for Microgrids)
Show Figures

Figure 1

16 pages, 1475 KiB  
Article
Power Loss Calculation in Oil-Type Distribution Transformers Supplying Nonlinear Loads
by Neda Miteva, Lior Sima and Kfir Jack Dagan
Electronics 2025, 14(12), 2500; https://doi.org/10.3390/electronics14122500 - 19 Jun 2025
Viewed by 163
Abstract
Power transformers are the most vital component in the electric grid. Their loss calculation is critical to transformer asset management and reflects on both operation and techno-economic assessment. Acknowledging the above, this paper presents an application of a novel loss calculation method to [...] Read more.
Power transformers are the most vital component in the electric grid. Their loss calculation is critical to transformer asset management and reflects on both operation and techno-economic assessment. Acknowledging the above, this paper presents an application of a novel loss calculation method to oil-type transformers supplying nonlinear loads. Unlike the methodology presented in std. C57.110-2018, the applied approach evaluates transformer loss components relying solely on readily available technical data. The method was experimentally validated using a full-scale 250kVA oil-type distribution transformer. Experimental results show close agreement with the theoretical model and feature errors smaller than 3%. Full article
(This article belongs to the Special Issue New Trends in Power Electronics for Microgrids)
Show Figures

Figure 1

25 pages, 547 KiB  
Article
An Interaction–Engagement–Intention Model: How Artificial Intelligence and Augmented Reality Transform the User–Platform Interaction Paradigm
by Zian Shah Kabir and Kyeong Kang
Electronics 2025, 14(12), 2499; https://doi.org/10.3390/electronics14122499 - 19 Jun 2025
Viewed by 214
Abstract
Interaction with mobile platforms changes users’ emotional and cognitive engagements through various stimuli cues that respond to behavioural intentions. Emerging technologies such as artificial intelligence (AI) and augmented reality (AR) foster more engagements and transform a new user–platform interaction paradigm in the e-commerce [...] Read more.
Interaction with mobile platforms changes users’ emotional and cognitive engagements through various stimuli cues that respond to behavioural intentions. Emerging technologies such as artificial intelligence (AI) and augmented reality (AR) foster more engagements and transform a new user–platform interaction paradigm in the e-commerce industry. This study signifies the effects of artificial intelligence and augmented reality in assessing user experience for mobile platforms. In this paper, we develop an interaction–engagement–intention model that considers users’ continuance intention based on perceived user experience. The proposed model uniquely explains a nuanced understanding of how the user–platform interactions evolve interactivity, product fit, artificial intelligence-driven recommendation, and online reviews in perceiving spatial presence and subjective norm. This paper explores the importance of attitude and trust as emotional states that influence the user’s behavioural responses. We validate the consequences of user–platform interactions toward continuance intention by conducting an online questionnaire survey and assessing user experience in augmented reality environments. The results contribute to adopting the co-created values of user–platform interactions through cognitive and emotional engagements that affect users’ continuance intention. The platform industry can apply the research outcomes by considering user experience and its implications to enhance the platforms’ capability with a broader aspect. Full article
Show Figures

Figure 1

21 pages, 1889 KiB  
Article
Optimizing Glioblastoma Multiforme Diagnosis: Semantic Segmentation and Survival Modeling Using MRI and Genotypic Data
by Yu-Hung Tsai, Wen-Yu Cheng, Bo-Hua Huang, Chiung-Chyi Shen and Meng-Hsiun Tsai
Electronics 2025, 14(12), 2498; https://doi.org/10.3390/electronics14122498 - 19 Jun 2025
Viewed by 215
Abstract
Glioblastoma multiforme (GBM) is the most aggressive and common primary brain tumor. Magnetic resonance imaging (MRI) provides detailed visualization of tumor morphology, edema, and necrosis. However, manually segmenting GBM from MRI scans is time-consuming, subjective, and prone to inter-observer variability. Therefore, automated and [...] Read more.
Glioblastoma multiforme (GBM) is the most aggressive and common primary brain tumor. Magnetic resonance imaging (MRI) provides detailed visualization of tumor morphology, edema, and necrosis. However, manually segmenting GBM from MRI scans is time-consuming, subjective, and prone to inter-observer variability. Therefore, automated and reliable segmentation methods are crucial for improving diagnostic accuracy. This study employs an image semantic segmentation model to segment brain tumors in MRI scans of GBM patients. The MRI recall images include T1-weighted imaging (T1WI) and fluid-attenuated inversion recovery (FLAIR) sequences. To enhance the performance of the semantic segmentation model, image preprocessing techniques were applied before analyzing and comparing commonly used segmentation models. Additionally, a survival model was constructed using discrete genotype attributes of GBM patients. The results indicate that the DeepLabV3+ model achieved the highest accuracy for semantic segmentation, with an accuracy of 77.9% on T1WI image sequences, while the U-Net model achieved 80.1% accuracy on FLAIR image sequences. Furthermore, in constructing the survival model using a discrete attribute dataset, the dataset was divided into three subsets based on different missing value handling strategies. This study found that replacing missing values with 1 resulted in the highest accuracy, with the Bernoulli Bayesian model and the multinomial Bayesian model achieving an accuracy of 94.74%. This study integrates image preprocessing techniques and semantic segmentation models to improve the accuracy and efficiency of brain tumor segmentation while also developing a highly accurate survival model. The findings aim to assist physicians in saving time and facilitating preliminary diagnosis and analysis. Full article
(This article belongs to the Special Issue Image Segmentation, 2nd Edition)
Show Figures

Figure 1

19 pages, 1433 KiB  
Article
Cost-Optimised Machine Learning Model Comparison for Predictive Maintenance
by Yating Yang and Muhammad Zahid Iqbal
Electronics 2025, 14(12), 2497; https://doi.org/10.3390/electronics14122497 - 19 Jun 2025
Viewed by 236
Abstract
Predictive maintenance is essential for reducing industrial downtime and costs, yet real-world datasets frequently encounter class imbalance and require cost-sensitive evaluation due to costly misclassification errors. This study utilises the SCANIA Component X dataset to advance predictive maintenance through machine learning, employing seven [...] Read more.
Predictive maintenance is essential for reducing industrial downtime and costs, yet real-world datasets frequently encounter class imbalance and require cost-sensitive evaluation due to costly misclassification errors. This study utilises the SCANIA Component X dataset to advance predictive maintenance through machine learning, employing seven supervised algorithms, Support Vector Machine, Random Forest, Decision Tree, K-Nearest Neighbours, Multi-Layer Perceptron, XGBoost, and LightGBM, trained on time-series features extracted via a sliding window approach. A bespoke cost-sensitive metric, aligned with SCANIA’s misclassification cost matrix, assesses model performance. Three imbalance mitigation strategies, downsampling, downsampling with SMOTETomek, and manual class weighting, were explored, with downsampling proving most effective. Random Forest and Support Vector Machine models achieved high accuracy and low misclassification costs, whilst a voting ensemble further enhanced cost efficiency. This research emphasises the critical role of cost-aware evaluation and imbalance handling, proposing an ensemble-based framework to improve predictive maintenance in industrial applications Full article
Show Figures

Figure 1

29 pages, 5989 KiB  
Article
Risk Analysis Method of Aviation Critical System Based on Bayesian Networks and Empirical Information Fusion
by Xiangjun Dang, Yongxuan Shao, Haoming Liu, Zhe Yang, Mingwen Zhong, Maohua Sun and Wu Deng
Electronics 2025, 14(12), 2496; https://doi.org/10.3390/electronics14122496 - 19 Jun 2025
Viewed by 106
Abstract
The intrinsic hazards associated with high-pressure hydrogen, combined with electromechanical interactions in hybrid architectures, pose significant challenges in predicting potential system risks during the conceptual design phase. In this paper, a risk analysis methodology integrating systems theoretic process analysis (STPA), D-S evidence theory, [...] Read more.
The intrinsic hazards associated with high-pressure hydrogen, combined with electromechanical interactions in hybrid architectures, pose significant challenges in predicting potential system risks during the conceptual design phase. In this paper, a risk analysis methodology integrating systems theoretic process analysis (STPA), D-S evidence theory, and Bayesian networks (BN) is established. The approach employs STPA to identify unsafe control actions and analyze their loss scenarios. Subsequently, D-S evidence theory quantifies the likelihood of risk factors, while the BN model’s nodal uncertainties to construct a risk network identifying critical risk-inducing events. This methodology provides a comprehensive risk analysis process that identifies systemic risk elements, quantifies risk probabilities, and incorporates uncertainties for quantitative risk assessment. These insights inform risk-averse design decisions for hydrogen–electric hybrid powered aircraft. A case study demonstrates the framework’s effectiveness. The approach bridges theoretical risk analysis with early-stage engineering practice, delivering actionable guidance for advancing zero-emission aviation. Full article
Show Figures

Figure 1

54 pages, 2065 KiB  
Review
Edge Intelligence: A Review of Deep Neural Network Inference in Resource-Limited Environments
by Dat Ngo, Hyun-Cheol Park and Bongsoon Kang
Electronics 2025, 14(12), 2495; https://doi.org/10.3390/electronics14122495 - 19 Jun 2025
Viewed by 174
Abstract
Deploying deep neural networks (DNNs) in resource-limited environments—such as smartwatches, IoT nodes, and intelligent sensors—poses significant challenges due to constraints in memory, computing power, and energy budgets. This paper presents a comprehensive review of recent advances in accelerating DNN inference on edge platforms, [...] Read more.
Deploying deep neural networks (DNNs) in resource-limited environments—such as smartwatches, IoT nodes, and intelligent sensors—poses significant challenges due to constraints in memory, computing power, and energy budgets. This paper presents a comprehensive review of recent advances in accelerating DNN inference on edge platforms, with a focus on model compression, compiler optimizations, and hardware–software co-design. We analyze the trade-offs between latency, energy, and accuracy across various techniques, highlighting practical deployment strategies on real-world devices. In particular, we categorize existing frameworks based on their architectural targets and adaptation mechanisms and discuss open challenges such as runtime adaptability and hardware-aware scheduling. This review aims to guide the development of efficient and scalable edge intelligence solutions. Full article
Show Figures

Figure 1

15 pages, 634 KiB  
Article
Robust H Time-Varying Formation Tracking for Heterogeneous Multi-Agent Systems with Unknown Control Input
by Jichuan Liu, Song Yang, Chunxi Dong and Peng Song
Electronics 2025, 14(12), 2494; https://doi.org/10.3390/electronics14122494 - 19 Jun 2025
Viewed by 85
Abstract
This paper studies the robust H time-varying formation tracking (TVFT) problem for heterogeneous nonlinear multi-agent systems (MASs) with parameter uncertainties, external disturbances, and unknown leader inputs. The objective is to ensure that follower agents track the leader’s trajectory while achieving a desired [...] Read more.
This paper studies the robust H time-varying formation tracking (TVFT) problem for heterogeneous nonlinear multi-agent systems (MASs) with parameter uncertainties, external disturbances, and unknown leader inputs. The objective is to ensure that follower agents track the leader’s trajectory while achieving a desired time-varying formation, even under unmodeled dynamics and disturbances. Unlike existing methods that rely on global topology information or homogeneous system assumptions, an adaptive control protocol is proposed in full distribution, requiring no global topology information, and integrates nonlinear compensation terms to handle unknown leader inputs and parameter uncertainties. Based on the Lyapunov theory and laplacian matrix, a robust H TVFT criterion is developed. Finally, a numerical example is given to verify the theory. Full article
Show Figures

Figure 1

25 pages, 2361 KiB  
Article
Enhancing Bug Assignment with Developer-Specific Feature Extraction and Hybrid Deep Learning
by Geunseok Yang, Jinfeng Ji and Dongkyu Kim
Electronics 2025, 14(12), 2493; https://doi.org/10.3390/electronics14122493 - 19 Jun 2025
Viewed by 187
Abstract
The increasing reliance on software in diverse domains has led to a surge in user-reported functional enhancements and unexpected bugs. In large-scale open-source projects like Eclipse and Mozilla, initial bug assignment frequently faces challenges, with approximately 50% of bug reports being reassigned due [...] Read more.
The increasing reliance on software in diverse domains has led to a surge in user-reported functional enhancements and unexpected bugs. In large-scale open-source projects like Eclipse and Mozilla, initial bug assignment frequently faces challenges, with approximately 50% of bug reports being reassigned due to the inability of the initially assigned developer to resolve the issue effectively. This reassignment process contributes to elevated software maintenance costs and delays in bug resolution. To address this, we propose a developer recommendation model that assigns the most suitable developer for a given bug report at the outset, thereby minimizing reassignment rates. Our approach combines a top-K feature selection algorithm tailored for each developer with a hybrid Convolutional Neural Network–Long Short-Term Memory (CNN–LSTM) architecture to capture the nuanced patterns in bug reports and developer expertise. The model was evaluated on prominent open-source projects, including Google Chrome, Mozilla Core, and Mozilla Firefox. Experimental results show that the proposed model significantly outperforms baseline approaches, with an improvement in developer recommendation accuracy of approximately 0.3582 when comparing the best-performing configuration to the worst-performing configuration of our model. Furthermore, the baseline difference was reduced by approximately 0.1343. A statistical analysis confirms the significant performance improvement achieved by the proposed method over existing baselines. These findings underscore the potential of our model to enhance efficiency in bug resolution workflows, reduce maintenance costs, and improve overall software quality in open-source ecosystems. Full article
(This article belongs to the Special Issue Feature Papers in "Computer Science & Engineering", 2nd Edition)
Show Figures

Figure 1

35 pages, 2826 KiB  
Article
AI-Driven Anomaly Detection for Securing IoT Devices in 5G-Enabled Smart Cities
by Manuel J. C. S. Reis
Electronics 2025, 14(12), 2492; https://doi.org/10.3390/electronics14122492 - 19 Jun 2025
Viewed by 251
Abstract
This paper proposes a novel AI-driven anomaly detection framework designed to enhance cybersecurity in IoT-enabled smart cities operating over 5G networks. While prior research has explored deep learning for anomaly detection, most existing systems rely on single-model architectures, employ centralized training, or lack [...] Read more.
This paper proposes a novel AI-driven anomaly detection framework designed to enhance cybersecurity in IoT-enabled smart cities operating over 5G networks. While prior research has explored deep learning for anomaly detection, most existing systems rely on single-model architectures, employ centralized training, or lack support for real-time, privacy-preserving deployment—limiting their scalability and robustness. To address these gaps, our system integrates a hybrid deep learning model combining autoencoders, long short-term memory (LSTM) networks, and convolutional neural networks (CNNs) to detect spatial, temporal, and reconstruction-based anomalies. Additionally, we implement federated learning and edge AI to enable decentralized, privacy-preserving threat detection across distributed IoT nodes. The system is trained and evaluated using a combination of real-world (CICIDS2017, TON_IoT, UNSW-NB15) and synthetically generated attack data, including adversarial perturbations. Experimental results show our hybrid model achieves a precision of 97.5%, a recall of 96.2%, and an F1 score of 96.8%, significantly outperforming traditional IDS and standalone deep learning methods. These findings validate the framework’s effectiveness and scalability, making it suitable for real-time intrusion detection and autonomous threat mitigation in smart city environments. Full article
(This article belongs to the Special Issue New Insights of Internet of Things in Industry 4.0)
Show Figures

Figure 1

19 pages, 3066 KiB  
Article
A Convex Constraint Approach for High-Type Control Loop Design
by Chao Liu, Xiaoxia Qiu and Yao Mao
Electronics 2025, 14(12), 2491; https://doi.org/10.3390/electronics14122491 - 19 Jun 2025
Viewed by 153
Abstract
This paper proposes a high-type control loop design method for LQR-LMI based on Lyapunov and polyhedral model theory. The high-type control loop design problem is simplified into a convex constraint problem, which achieves superior tracking performance. In this framework, the input amplitude of [...] Read more.
This paper proposes a high-type control loop design method for LQR-LMI based on Lyapunov and polyhedral model theory. The high-type control loop design problem is simplified into a convex constraint problem, which achieves superior tracking performance. In this framework, the input amplitude of the control signal, the poles of the closed-loop system, the suppression of external interference and the perturbation of internal parameters are considered, and the linear matrix inequality (LMI) method is effectively used to solve the problems. In this paper, the polyhedral model control theory is introduced to characterize the uncertainty of the system for the change of model parameters of the controlled plant. Aiming at the problem of external disturbance suppression, the H2/H control method is introduced into the system. These control methods provide the basis for the design of the high-type control loop. Compared with the simulation results of other optimization algorithms, the effectiveness and superiority of the controller parameter tuning rules in the proposed high-type control loop are verified. Full article
Show Figures

Figure 1

12 pages, 2513 KiB  
Article
Optoelectronic Memristor Based on ZnO/Cu2O for Artificial Synapses and Visual System
by Chen Meng, Hongxin Liu, Tong Li, Jin Luo and Sijie Zhang
Electronics 2025, 14(12), 2490; https://doi.org/10.3390/electronics14122490 - 19 Jun 2025
Viewed by 189
Abstract
The development of artificial intelligence has resulted in significant challenges to conventional von Neumann architectures, including the separation of storage and computation, and power consumption bottlenecks. The new generation of brain-like devices is accelerating its evolution in the direction of high-density integration and [...] Read more.
The development of artificial intelligence has resulted in significant challenges to conventional von Neumann architectures, including the separation of storage and computation, and power consumption bottlenecks. The new generation of brain-like devices is accelerating its evolution in the direction of high-density integration and integrated sensing, storage, and computing. The structural and information transmission similarity between memristors and biological synapses signifies their unique potential in sensing and memory. Therefore, memristors have become potential candidates for neural devices. In this paper, we have designed an optoelectronic memristor based on a ZnO/Cu2O structure to achieve synaptic behavior through the modulation of electrical signals, demonstrating the recognition of a dataset by a neural network. Furthermore, the optical synaptic functions, such as short-term/long-term potentiation and learn-forget-relearn behavior, and advanced synaptic behavior of optoelectronic modulation, are successfully simulated. The mechanism of light-induced conductance enhancement is explained by the barrier change at the interface. This work explores a new pathway for constructing next-generation optoelectronic synaptic devices, which lays the foundation for future brain-like visual chips and intelligent perceptual devices. Full article
Show Figures

Figure 1

19 pages, 9631 KiB  
Article
Res2Former: Integrating Res2Net and Transformer for a Highly Efficient Speaker Verification System
by Defu Chen, Yunlong Zhou, Xianbao Wang, Sheng Xiang, Xiaohu Liu and Yijian Sang
Electronics 2025, 14(12), 2489; https://doi.org/10.3390/electronics14122489 - 19 Jun 2025
Viewed by 196
Abstract
Speaker verification (SV) is an exceptionally effective method of biometric authentication. However, its performance is heavily influenced by the effectiveness of the extracted speaker features and their suitability for use in resource-limited environments. Transformer models and convolutional neural networks (CNNs), leveraging self-attention mechanisms, [...] Read more.
Speaker verification (SV) is an exceptionally effective method of biometric authentication. However, its performance is heavily influenced by the effectiveness of the extracted speaker features and their suitability for use in resource-limited environments. Transformer models and convolutional neural networks (CNNs), leveraging self-attention mechanisms, have demonstrated state-of-the-art performance in most Natural Language Processing (NLP) and Image Recognition tasks. However, previous studies indicate that standalone Transformer and CNN architectures present distinct challenges in speaker verification. Specifically, while Transformer models deliver good results, they fail to meet the requirements of low-resource scenarios and computational efficiency. On the other hand, CNNs perform well in resource-constrained environments but suffer from significantly reduced recognition accuracy. Several existing approaches, such as Conformer, combine Transformers and CNNs but still face challenges related to high resource consumption and low computational efficiency. To address these issues, we propose a novel solution that enhances the Transformer model by introducing multi-scale convolutional attention and a Global Response Normalization (GRN)-based feed-forward network, resulting in a lightweight backbone architecture called the lightweight simple transformer (LST). We further improve LST by incorporating the Res2Net structure from CNN, yielding the Res2Former model—a low-parameter, high—precision SV model. In Res2Former, we design and implement a time-frequency adaptive feature fusion(TAFF) mechanism that enables fine-grained feature propagation by fusing features at different depths at the frame level. Additionally, holistic fusion is employed for global feature propagation across the model. To enhance performance, multiple convergence methods are introduced, improving the overall efficacy of the SV system. Experimental results on the VoxCeleb1-O, VoxCeleb1-E, VoxCeleb1-H, and Cn-Celeb(E) datasets demonstrate that Res2Former achieves excellent performance, with the Large configuration attaining Equal Error Rate (EER)/Minimum Detection Cost Function (minDCF) scores of 0.81%/0.08, 0.98%/0.11, 1.81%/0.17, and 8.39%/0.46, respectively. Notably, the Base configuration of Res2Former, with only 1.73M parameters, also delivers competitive results. Full article
(This article belongs to the Special Issue New Advances in Embedded Software and Applications)
Show Figures

Figure 1

21 pages, 4507 KiB  
Article
GSTD-DETR: A Detection Algorithm for Small Space Targets Based on RT-DETR
by Yijian Zhang, Huichao Guo, Yang Zhao, Laixian Zhang, Chenglong Luan, Yingchun Li and Xiaoyu Zhang
Electronics 2025, 14(12), 2488; https://doi.org/10.3390/electronics14122488 - 19 Jun 2025
Viewed by 248
Abstract
Ground-based optical equipment for detecting geostationary orbit space targets typically involves long-exposure imaging, facing challenges such as small and blurred target images, complex backgrounds, and star streaks obstructing the view. To address these issues, this study proposes a GSTD-DETR model based on Real-Time [...] Read more.
Ground-based optical equipment for detecting geostationary orbit space targets typically involves long-exposure imaging, facing challenges such as small and blurred target images, complex backgrounds, and star streaks obstructing the view. To address these issues, this study proposes a GSTD-DETR model based on Real-Time Detection Transformer (RT-DETR), which aims to balance model efficiency and detection accuracy. First, we introduce a Dynamic Cross-Stage Partial (DynCSP) backbone network for feature extraction and fusion, which enhances the network’s representational capability by reducing convolutional parameters and improving information exchange between channels. This effectively reduces the model’s parameter count and computational complexity. Second, we propose a ResFine model with a feature pyramid designed for small target detection, enhancing its ability to perceive small targets. Additionally, we improve the detection head and incorporate a Dynamic Multi-Channel Attention mechanism, which strengthens the focus on critical regions. Finally, we designed an Area-Weighted NWD loss function to improve detection accuracy. The experimental results show that compared to RT-DETR-r18, the GSTD-DETR model reduces the parameter count by 29.74% on the SpotGEO dataset. Its AP50 and AP50:95 improve by 1.3% and 4.9%, reaching 88.6% and 49.9%, respectively. The GSTD-DETR model demonstrates superior performance in the detection accuracy of faint and small space targets. Full article
Show Figures

Figure 1

21 pages, 7734 KiB  
Article
Thermal–Flow Coupling Simulation and Performance Analysis for Self-Starting Permanent Magnet Motors
by Jinhui Liu, Yunbo Shi, Yang Zheng and Minghui Wang
Electronics 2025, 14(12), 2487; https://doi.org/10.3390/electronics14122487 - 19 Jun 2025
Viewed by 1590
Abstract
In practical applications, the fully enclosed structure is always required by self-starting permanent magnet synchronous motors for safety. However, internal heat dissipation can be obstructed as a result, which affects operational reliability. To resolve the issue, this study takes a 3 kW self-starting [...] Read more.
In practical applications, the fully enclosed structure is always required by self-starting permanent magnet synchronous motors for safety. However, internal heat dissipation can be obstructed as a result, which affects operational reliability. To resolve the issue, this study takes a 3 kW self-starting permanent magnet synchronous motor as the research object. Based on fluid dynamics and fluid solid coupling heat transfer theory, the model is reasonably simplified according to the characteristics of the structure of motor cooling, and basic assumptions and boundary conditions are given to establish a three-dimensional, whole machine solution domain model. The finite element method is used to numerically analyze and calculate under rated conditions. The fluid flow characteristics, heat transfer characteristics, motion trajectories of the cooling medium on the surface of the external casing, fan, and internal stator and rotor domains, and winding ends are analyzed. Therefore, the internal rheological characteristics and temperature rise distribution law of the self-starting permanent magnet synchronous motor can be revealed. Based on the aforementioned research, a novel method to design the wind spur structure on the surface of the rotor end is proposed. By comparing the simulation results of the fluid field and temperature field of the motor under wind spur structures with different lengths and equidistant distributions in the circumferential direction of the rotor end, the influence of the convective heat characteristics can be systematically studied. Lastly, the accuracy of the calculation results and the rationality of the solution method are verified through experiments of temperature rise, and the flow temperature distribution characteristics of the motor can be optimized by the wind spur structure, which can be used in practical applications. Full article
Show Figures

Figure 1

14 pages, 3406 KiB  
Article
Development and Evaluation of a Novel Mixed Reality-Based Surgical Navigation System for Distal Locking of Intramedullary Nails
by Fei Lyu, Puxun Tu, Xingguang Tao and Huixiang Wang
Electronics 2025, 14(12), 2486; https://doi.org/10.3390/electronics14122486 - 19 Jun 2025
Viewed by 156
Abstract
Intramedullary nailing (IMN) is the gold standard for fixing mid-shaft fractures of long bones, but distal locking remains a challenging procedure. This study aims to develop and evaluate a novel mixed reality (MR)-based surgical navigation system for distal locking of IMN through phantom [...] Read more.
Intramedullary nailing (IMN) is the gold standard for fixing mid-shaft fractures of long bones, but distal locking remains a challenging procedure. This study aims to develop and evaluate a novel mixed reality (MR)-based surgical navigation system for distal locking of IMN through phantom experiments. Twelve bone models closely replicating the mechanical properties, anatomy, and density of human tibial bone were utilized. Six orthopedic surgeons participated in the phantom experiments using both MR and traditional electromagnetic (EM) navigation systems. Effectiveness was evaluated using postoperative fluoroscopic imaging and the time taken for distal locking. Compared to the EM navigation system, the MR system significantly reduced distal locking time (81.54 ± 6.06 vs. 132.67 ± 6.45 s per screw) and achieved a higher success rate (23/24 vs. 21/24 screws accurately placed), but the difference in terms of success rate is not statistically significant. The MR-based navigation system for distal locking of IMN is time-efficient, accurate, and shows high potential for enhancing surgical precision in orthopedic procedures. Full article
(This article belongs to the Special Issue Medical Robots: Safety, Performance and Improvement)
Show Figures

Figure 1

12 pages, 7323 KiB  
Article
WinEdge: Low-Power Winograd CNN Execution with Transposed MRAM for Edge Devices
by Milad Ashtari Gargari, Sepehr Tabrizchi and Arman Roohi
Electronics 2025, 14(12), 2485; https://doi.org/10.3390/electronics14122485 - 19 Jun 2025
Viewed by 225
Abstract
This paper presents a novel transposed MRAM architecture (WinEdge) specifically optimized for Winograd convolution acceleration in edge computing devices. Leveraging Magnetic Tunnel Junctions (MTJs) with Spin Hall Effect (SHE)-assisted Spin-Transfer Torque (STT) writing, the proposed design enables a single SHE current to simultaneously [...] Read more.
This paper presents a novel transposed MRAM architecture (WinEdge) specifically optimized for Winograd convolution acceleration in edge computing devices. Leveraging Magnetic Tunnel Junctions (MTJs) with Spin Hall Effect (SHE)-assisted Spin-Transfer Torque (STT) writing, the proposed design enables a single SHE current to simultaneously write data to four MTJs, substantially reducing power consumption. Additionally, the integration of stacked MTJs significantly improves storage density. The proposed WinEdge efficiently supports both standard and transposed data access modes regardless of bit-width, achieving up to 36% lower power, 47% reduced energy consumption, and 28% faster processing speed compared to existing designs. Simulations conducted in 45 nm CMOS technology validate its superiority over conventional SRAM-based solutions for convolutional neural network (CNN) acceleration in resource-constrained edge environments. Full article
(This article belongs to the Special Issue Emerging Computing Paradigms for Efficient Edge AI Acceleration)
Show Figures

Figure 1

22 pages, 2320 KiB  
Article
Two-Stage Coordinated Operation Mechanism for Virtual Power Plant Clusters Based on Energy Interaction
by Xingang Yang, Lei Qi, Di Wang and Qian Ai
Electronics 2025, 14(12), 2484; https://doi.org/10.3390/electronics14122484 - 18 Jun 2025
Viewed by 168
Abstract
As an essential platform for aggregating and coordinating distributed energy resources (DERs), the virtual power plant (VPP) has attracted widespread attention in recent years. With the increasing scale of VPPs, energy interaction and sharing among VPP clusters (VPPCs) have become key approaches to [...] Read more.
As an essential platform for aggregating and coordinating distributed energy resources (DERs), the virtual power plant (VPP) has attracted widespread attention in recent years. With the increasing scale of VPPs, energy interaction and sharing among VPP clusters (VPPCs) have become key approaches to improving energy utilization efficiency and reducing operational costs. Therefore, studying the coordinated operation mechanism of VPPCs is of great significance. This paper proposes a two-stage coordinated operation model for VPPCs based on energy interaction to enhance the overall economic performance and coordination of the cluster. In the day-ahead stage, a cooperative operation model based on Nash bargaining theory is constructed. The inherently non-convex and nonlinear problem is decomposed into a cluster-level benefit maximization subproblem and a benefit allocation subproblem. The Alternating Direction Method of Multipliers (ADMM) is employed to achieve distributed optimization, ensuring both the efficiency of coordination and the privacy and decision independence of each VPP. In the intra-day stage, to address the uncertainty in renewable generation and load demand, a real-time pricing mechanism based on the supply–demand ratio is designed. Each VPP performs short-term energy forecasting and submits real-time supply–demand information to the coordination center, which dynamically determines the price for the next trading interval according to the reported imbalance. This pricing mechanism facilitates real-time electricity sharing among VPPs. Finally, numerical case studies validate the effectiveness and practical value of the proposed model in improving both operational efficiency and fairness. Full article
Show Figures

Figure 1

23 pages, 2330 KiB  
Review
Radio Frequency Interference, Its Mitigation and Its Implications for the Civil Aviation Industry
by Adnan Malik and Muzaffar Rao
Electronics 2025, 14(12), 2483; https://doi.org/10.3390/electronics14122483 - 18 Jun 2025
Viewed by 174
Abstract
Radio Frequency Interference has emerged as a growing challenge for aviation safety and system integrity due to the increasing spectral overlap between communication technologies and aviation systems. This paper investigates the sources, types, and consequences of RFI in Global Navigation Satellite Systems, Instrument [...] Read more.
Radio Frequency Interference has emerged as a growing challenge for aviation safety and system integrity due to the increasing spectral overlap between communication technologies and aviation systems. This paper investigates the sources, types, and consequences of RFI in Global Navigation Satellite Systems, Instrument Landing Systems, and altimeters used in civil aviation. A detailed examination of both intentional and unintentional interference is presented, highlighting real-world incidents and simulated impact models. The study analyzes technical mechanisms such as receiver desensitization, intermodulation, and cross-modulation, and further explores UAV-based interference detection frameworks. Mitigation strategies are reviewed, including regulatory practices, spectrum filters, shielding architectures, and dynamic UAV sensing systems. Comparative insights into simulation results, shielding techniques, and regulatory gaps are discussed. The paper concludes with recommendations for enhancing current aviation standards and suggests a hybrid validation model combining in-flight measurements with simulation-based assessments. This research contributes to the understanding of electromagnetic vulnerabilities in aviation and provides a basis for future mitigation protocols. Full article
Show Figures

Figure 1

18 pages, 2712 KiB  
Article
Height Control and Experimental Study of Linear Motor-Based Active Suspension Systems
by Chao Jiang and Jialing Yao
Electronics 2025, 14(12), 2482; https://doi.org/10.3390/electronics14122482 - 18 Jun 2025
Viewed by 125
Abstract
This study addresses the challenge of ride height control in linear motor-based active suspension systems by proposing a control strategy based on linear active disturbance rejection control (LADRC). The effectiveness of the proposed approach is experimentally validated using a high-precision test platform built [...] Read more.
This study addresses the challenge of ride height control in linear motor-based active suspension systems by proposing a control strategy based on linear active disturbance rejection control (LADRC). The effectiveness of the proposed approach is experimentally validated using a high-precision test platform built on the NI cRIO-9014 real-time controller. The platform integrates a permanent magnet synchronous linear motor, a motor driver, acceleration sensors, and a vibration control system to realize closed-loop control of vehicle body height. Experimental results demonstrate that, compared with conventional PID control, LADRC achieves superior performance in height regulation accuracy, dynamic responsiveness, vertical acceleration suppression, and steady-state stability. In step response experiments, LADRC reduces the regulation time by 53.8% (from 1.3 s to 0.6 s) and lowers the steady-state error from 0.502 mm to 0.05 mm. In sinusoidal trajectory tracking tests, the LADRC approach reduces peak and RMS tracking errors by 81.5% and 80.3%, respectively. Moreover, under random road excitation, LADRC effectively attenuates high-frequency body vibrations, with reductions of 29.58% in peak vertical acceleration and 12.23% in RMS acceleration. Full article
Show Figures

Figure 1

20 pages, 2223 KiB  
Article
ChatGPT-Based Model for Controlling Active Assistive Devices Using Non-Invasive EEG Signals
by Tais da Silva Mota, Saket Sarkar, Rakshith Poojary and Redwan Alqasemi
Electronics 2025, 14(12), 2481; https://doi.org/10.3390/electronics14122481 - 18 Jun 2025
Viewed by 206
Abstract
With an anticipated 3.6 million Americans who will be living with limb loss by 2050, the demand for active assistive devices is rapidly increasing. This study investigates the feasibility of leveraging a ChatGPT-based (Version 4o) model to predict motion based on input electroencephalogram [...] Read more.
With an anticipated 3.6 million Americans who will be living with limb loss by 2050, the demand for active assistive devices is rapidly increasing. This study investigates the feasibility of leveraging a ChatGPT-based (Version 4o) model to predict motion based on input electroencephalogram (EEG) signals, enabling the non-invasive control of active assistive devices. To achieve this goal, three objectives were set. First, the model’s capability to derive accurate mathematical relationships from numerical datasets was validated to establish a foundational level of computational accuracy. Next, synchronized arm motion videos and EEG signals were introduced, which allowed the model to filter, normalize, and classify EEG data in relation to distinct text-based arm motions. Finally, the integration of marker-based motion capture data provided motion information, which is essential for inverse kinematics applications in robotic control. The combined findings highlight the potential of ChatGPT-generated machine learning systems to effectively correlate multimodal data streams and serve as a robust foundation for the intuitive, non-invasive control of assistive technologies using EEG signals. Future work will focus on applying the model to real-time control applications while expanding the dataset’s diversity to enhance the accuracy and performance of the model, with the ultimate aim of improving the independence and quality of life of individuals who rely on active assistive devices. Full article
(This article belongs to the Special Issue Advances in Intelligent Control Systems)
Show Figures

Figure 1

24 pages, 2653 KiB  
Article
DARC: Disturbance-Aware Redundant Control for Human–Robot Co-Transportation
by Al Jaber Mahmud, Amir Hossain Raj, Duc M. Nguyen, Xuesu Xiao and Xuan Wang
Electronics 2025, 14(12), 2480; https://doi.org/10.3390/electronics14122480 - 18 Jun 2025
Viewed by 133
Abstract
This paper introduces Disturbance-Aware Redundant Control (DARC), a control framework addressing the challenge of human–robot co-transportation under disturbances. Our method integrates a disturbance-aware Model Predictive Control (MPC) framework with a proactive pose optimization mechanism. The robotic system, comprising a mobile base and a [...] Read more.
This paper introduces Disturbance-Aware Redundant Control (DARC), a control framework addressing the challenge of human–robot co-transportation under disturbances. Our method integrates a disturbance-aware Model Predictive Control (MPC) framework with a proactive pose optimization mechanism. The robotic system, comprising a mobile base and a manipulator arm, compensates for uncertain human behaviors and internal actuation noise through a two-step iterative process. At each planning horizon, a candidate set of feasible joint configurations is generated using a Conditional Variational Autoencoder (CVAE). From this set, one configuration is selected by minimizing an estimated control cost computed via a disturbance-aware Discrete Algebraic Riccati Equation (DARE), which also provides the optimal control inputs for both the mobile base and the manipulator arm. We derive the disturbance-aware DARE and validate DARC with simulated experiments with a Fetch robot. Evaluations across various trajectories and disturbance levels demonstrate that our proposed DARC framework outperforms baseline algorithms that lack disturbance modeling, pose optimization, or both. Full article
(This article belongs to the Special Issue Advancements in Robotics: Perception, Manipulation, and Interaction)
Show Figures

Figure 1

23 pages, 5086 KiB  
Article
RMSENet: Multi-Scale Reverse Master–Slave Encoder Network for Remote Sensing Image Scene Classification
by Yongjun Wen, Jiake Zhou, Zhao Zhang and Lijun Tang
Electronics 2025, 14(12), 2479; https://doi.org/10.3390/electronics14122479 - 18 Jun 2025
Viewed by 192
Abstract
Aiming at the problems that the semantic representation of information extracted by the shallow layer of the current remote sensing image scene classification network is insufficient, and that the utilization rate of primary visual features decreases with the deepening of the network layers, [...] Read more.
Aiming at the problems that the semantic representation of information extracted by the shallow layer of the current remote sensing image scene classification network is insufficient, and that the utilization rate of primary visual features decreases with the deepening of the network layers, this paper designs a multi-scale reverse master–slave encoder network (RMSENet). It proposes a reverse cross-scale supplementation strategy for the slave encoder and a reverse cross-scale fusion strategy for the master encoder. This not only reversely supplements the high-level semantic information extracted by the slave encoder to the shallow layer of the master encoder network in a cross-scale manner but also realizes the cross-scale fusion of features at all stages of the master encoder. A multi-frequency coordinate channel attention mechanism is proposed, which captures the inter-channel interactions of input feature maps while embedding spatial position information and rich frequency information. A multi-scale wavelet self-attention mechanism is proposed, which completes lossless downsampling of input feature maps before self-attention operations. Experiments on open-source datasets RSSCN7, SIRI-WHU, and AID show that the classification accuracies of RMSENet reach 97.41%, 97.61%, and 95.9%, respectively. Compared with current mainstream deep learning models, RMSENet has lower network complexity and excellent classification accuracy. Full article
Show Figures

Figure 1

24 pages, 4466 KiB  
Article
Natural Interaction in Virtual Heritage: Enhancing User Experience with Large Language Models
by Isabel Sánchez-Berriel, Fernando Pérez-Nava and Lucas Pérez-Rosario
Electronics 2025, 14(12), 2478; https://doi.org/10.3390/electronics14122478 - 18 Jun 2025
Viewed by 113
Abstract
In recent years, Virtual Reality (VR) has emerged as a powerful tool for disseminating Cultural Heritage (CH), often incorporating Virtual Humans (VHs) to guide users through historical recreations. The advent of Large Language Models (LLMs) now enables natural, unscripted communication with these VHs, [...] Read more.
In recent years, Virtual Reality (VR) has emerged as a powerful tool for disseminating Cultural Heritage (CH), often incorporating Virtual Humans (VHs) to guide users through historical recreations. The advent of Large Language Models (LLMs) now enables natural, unscripted communication with these VHs, even on limited devices. This paper details a natural interaction system for VHs within a VR application of San Cristóbal de La Laguna, a UNESCO World Heritage Site. Our system integrates Speech-to-Text, LLM-based dialogue generation, and Text-to-Speech synthesis. Adhering to user-centered design (UCD) principles, we conducted two studies: a preliminary study revealing user interest in historically adapted language, and a qualitative test that identified key user experience improvements, such as incorporating feedback mechanisms and gender selection for VHs. The project successfully developed a prioritized user experience, focusing on usability evaluation, immersion, and dialogue quality. We propose a generalist methodology and recommendations for integrating unscripted VH dialogue in VR. However, limitations include dialogue generation latency and reduced quality in non-English languages. While a formative usability test evaluated the process, the small sample size restricts broad generalizations about user behavior. Full article
Show Figures

Figure 1

18 pages, 5314 KiB  
Article
Model-Free Predictive Current Control for an Improved Transverse-Flux Flux-Reversal Linear Motor
by Quanmao Li, Xin He and Xiaobao Yang
Electronics 2025, 14(12), 2477; https://doi.org/10.3390/electronics14122477 - 18 Jun 2025
Viewed by 133
Abstract
One of the significant features of the transverse flux linear motors (TFLMs) is the relatively higher thrust density, since the main flux loop of TFLM is located on a perpendicular plane to the motion direction. As one type of reluctance TFLM, the transverse-flux [...] Read more.
One of the significant features of the transverse flux linear motors (TFLMs) is the relatively higher thrust density, since the main flux loop of TFLM is located on a perpendicular plane to the motion direction. As one type of reluctance TFLM, the transverse-flux flux-reversal linear motor (TF-FRLM) is an interesting topology for the long stroke scene, which owns a passive reluctance type secondary, and the high-priced permanent magnets are only fixed on the short primary. To further enhance the practicality of the TF-FRLM, this paper focuses on the topology improvement and the control methods of TF-FRLM. Based on an improved TF-FRLM, a model-free predictive current control (MFPCC) method with suppressed sampling noise is proposed in this paper. Firstly, the details of structural improvements on the TF-FRLM topology are described, and some typical performances of TF-FRLMs are simulated by the three-dimensional finite element method and tested by a prototype. Then, based on the proposed basic principle of MFPCC, the reference current for inner-loop control is predicted. To ensure the prediction accuracy of the current in the MFPCC control method, the average filtering principle is used to suppress the impact of current sampling noise on performance. Finally, through comparative experiments on MFPCC schemes on the prototype platform, the effectiveness of the proposed control method is verified. Full article
Show Figures

Figure 1

43 pages, 2159 KiB  
Systematic Review
A Systematic Review and Classification of HPC-Related Emerging Computing Technologies
by Ehsan Arianyan, Niloofar Gholipour, Davood Maleki, Neda Ghorbani, Abdolah Sepahvand and Pejman Goudarzi
Electronics 2025, 14(12), 2476; https://doi.org/10.3390/electronics14122476 - 18 Jun 2025
Viewed by 277
Abstract
In recent decades, access to powerful computational resources has brought about a major transformation in science, with supercomputers drawing significant attention from academia, industry, and governments. Among these resources, high-performance computing (HPC) has emerged as one of the most critical processing infrastructures, providing [...] Read more.
In recent decades, access to powerful computational resources has brought about a major transformation in science, with supercomputers drawing significant attention from academia, industry, and governments. Among these resources, high-performance computing (HPC) has emerged as one of the most critical processing infrastructures, providing a suitable platform for evaluating and implementing novel technologies. In this context, the development of emerging computing technologies has opened up new horizons in information processing and the delivery of computing services. In this regard, this paper systematically reviews and classifies emerging HPC-related computing technologies, including quantum computing, nanocomputing, in-memory architectures, neuromorphic systems, serverless paradigms, adiabatic technology, and biological solutions. Within the scope of this research, 142 studies which were mostly published between 2018 and 2025 are analyzed, and relevant hardware solutions, domain-specific programming languages, frameworks, development tools, and simulation platforms are examined. The primary objective of this study is to identify the software and hardware dimensions of these technologies and analyze their roles in improving the performance, scalability, and efficiency of HPC systems. To this end, in addition to a literature review, statistical analysis methods are employed to assess the practical applicability and impact of these technologies across various domains, including scientific simulation, artificial intelligence, big data analytics, and cloud computing. The findings of this study indicate that emerging HPC-related computing technologies can serve as complements or alternatives to classical computing architectures, driving substantial transformations in the design, implementation, and operation of high-performance computing infrastructures. This article concludes by identifying existing challenges and future research directions in this rapidly evolving field. Full article
Show Figures

Figure 1

20 pages, 7927 KiB  
Article
Efficient License Plate Alignment and Recognition Using FPGA-Based Edge Computing
by Chao-Hsiang Hsiao, Hoi Lee, Yin-Tien Wang and Min-Jie Hsu
Electronics 2025, 14(12), 2475; https://doi.org/10.3390/electronics14122475 - 18 Jun 2025
Viewed by 268
Abstract
Efficient and accurate license plate recognition (LPR) in unconstrained environments remains a critical challenge, particularly when confronted with skewed imaging angles and the limited computational capabilities of edge devices. In this study, we propose a high-performance, FPGA-based license plate alignment and recognition (LPAR) [...] Read more.
Efficient and accurate license plate recognition (LPR) in unconstrained environments remains a critical challenge, particularly when confronted with skewed imaging angles and the limited computational capabilities of edge devices. In this study, we propose a high-performance, FPGA-based license plate alignment and recognition (LPAR) system to address these issues. Our LPAR system integrates lightweight deep learning models, including YOLOv4-tiny for license plate detection, a refined convolutional pose machine (CPM) for pose estimation and alignment, and a modified LPRNet for character recognition. By restructuring the pose estimation and alignment architectures to enhance the geometric correction of license plates and adding channel and spatial attention mechanisms to LPRNet for better character recognition, the proposed LPAR system improves recognition accuracy from 88.33% to 95.00%. The complete pipeline achieved a processing speed of 2.00 frames per second (FPS) on a resource-constrained FPGA platform, demonstrating its practical viability for real-time deployment in edge computing scenarios. Full article
Show Figures

Figure 1

1 pages, 122 KiB  
Correction
Correction: Pal et al. Fintech Agents: Technologies and Theories. Electronics 2023, 12, 3301
by Anagh Pal, Shreya Gopi and Kwan Min Lee
Electronics 2025, 14(12), 2474; https://doi.org/10.3390/electronics14122474 - 18 Jun 2025
Viewed by 66
Abstract
In the original publication [...] Full article
(This article belongs to the Section Computer Science & Engineering)
28 pages, 8777 KiB  
Article
Exploring Carbon-Fiber UAV Structures as Communication Antennas for Adaptive Relay Applications
by Cristian Vidan, Andrei Avram, Lucian Grigorie, Grigore Cican and Mihai Nacu
Electronics 2025, 14(12), 2473; https://doi.org/10.3390/electronics14122473 - 18 Jun 2025
Viewed by 220
Abstract
This study investigates the electromagnetic performance of two carbon fiber monopole antennas integrated into a UAV copter frame, with emphasis on design adaptation, impedance matching, and propagation behavior. A comprehensive experimental campaign was conducted to characterize key parameters such as center frequency, bandwidth, [...] Read more.
This study investigates the electromagnetic performance of two carbon fiber monopole antennas integrated into a UAV copter frame, with emphasis on design adaptation, impedance matching, and propagation behavior. A comprehensive experimental campaign was conducted to characterize key parameters such as center frequency, bandwidth, gain, VSWR, and S11. Both antennas exhibited dual-band resonance at approximately 381 MHz and 1.19 GHz, each achieving a 500 MHz bandwidth where VSWR ≤ 2. The modified antenna achieved a minimum reflection coefficient of –14.6 dB and a VSWR of 1.95 at 381.45 MHz, closely aligning with theoretical predictions. Gain deviations between measured (0.15–0.19 dBi) and calculated (0.19 dBi) values remained within 0.04 dB, while received power fluctuations did not exceed 1.3 dB under standard test conditions despite the composite material’s finite conductivity. Free-space link-budget tests at 0.5 m and 2 m of separation revealed received-power deviations of 0.9 dB and 1.3 dB, respectively, corroborating the Friis model. Radiation pattern measurements in both azimuth and elevation planes confirmed good directional behavior, with minor side lobe variations, where Antenna A displayed variations between 270° and 330° in azimuth, while Antenna B remained more uniform. A 90° polarization mismatch led to a 15 dBm signal drop, and environmental obstructions caused losses of 9.4 dB, 12.6 dB, and 18.3 dB, respectively, demonstrating the system’s sensitivity to alignment and surroundings. Additionally, signal strength changes observed in a Two-Ray propagation setup validated the importance of ground reflection effects. Small-scale fading analysis at 5 m LOS indicated a Rician-distributed envelope with mean attenuation of 53.96 dB, σdB = 5.57 dB, and a two-sigma interval spanning 42.82 dB to 65.11 dB; the fitted K-factor confirmed the dominance of the LOS component. The findings confirm that carbon fiber UAV frames can serve as effective directional antenna supports, providing proper alignment and tuning. These results support the future integration of lightweight, structure-embedded antennas in UAV systems, with potential benefits in communication efficiency, stealth, and design simplification. Full article
(This article belongs to the Special Issue Unmanned Aircraft Systems with Autonomous Navigation, 2nd Edition)
Show Figures

Figure 1

25 pages, 2296 KiB  
Article
Multimedia Graph Codes for Fast and Semantic Retrieval-Augmented Generation
by Stefan Wagenpfeil
Electronics 2025, 14(12), 2472; https://doi.org/10.3390/electronics14122472 - 18 Jun 2025
Viewed by 190
Abstract
Retrieval-Augmented Generation (RAG) has become a central approach to enhance the factual consistency and domain specificity of large language models (LLMs) by incorporating external context at inference time. However, most existing RAG systems rely on dense vector-based similarity, which fails to capture complex [...] Read more.
Retrieval-Augmented Generation (RAG) has become a central approach to enhance the factual consistency and domain specificity of large language models (LLMs) by incorporating external context at inference time. However, most existing RAG systems rely on dense vector-based similarity, which fails to capture complex semantic structures, relational dependencies, and multimodal content. In this paper, we introduce Graph Codes—a matrix-based encoding of Multimedia Feature Graphs—as an alternative retrieval paradigm. Graph Codes preserve semantic topology by explicitly encoding entities and their typed relationships from multimodal documents, enabling structure-aware and interpretable retrieval. We evaluate our system in two domains: multimodal scene understanding (200 annotated image-question pairs) and clinical question answering (150 real-world medical queries with 10,000 structured knowledge snippets). Results show that our method outperforms dense retrieval baselines in precision (+9–15%), reduces hallucination rates by over 30%, and yields higher expert-rated answer quality. Theoretically, this work demonstrates that symbolic similarity over typed semantic graphs provides a more faithful alignment mechanism than latent embeddings. Practically, it enables interpretable, modality-agnostic retrieval pipelines deployable in high-stakes domains such as medicine or law. We conclude that Graph Code-based RAG bridges the gap between structured knowledge representation and neural generation, offering a robust and explainable alternative to existing approaches. Full article
(This article belongs to the Special Issue AI Synergy: Vision, Language, and Modality)
Show Figures

Figure 1

Previous Issue
Next Issue
Back to TopTop