Electronics

14 pages, 4914 KB

Open AccessArticle

Demonstration of 2D Optoelectronic THz-Wave Beam Steering

by Bo Li, Hussein Ssali, Yuanhao Li, Ming Che, Shenghong Ye, Yuya Mikami and Kazutoshi Kato

Electronics 2025, 14(24), 4980; https://doi.org/10.3390/electronics14244980 - 18 Dec 2025

Viewed by 299

Advanced two-dimensional (2D) beam steering is essential for unlocking the full potential of terahertz (THz) systems in future 6G communications and high-resolution imaging. However, achieving wide-angle, high-speed, and high-precision 2D beam control within a compact THz platform remains a significant challenge. In this [...] Read more.

Advanced two-dimensional (2D) beam steering is essential for unlocking the full potential of terahertz (THz) systems in future 6G communications and high-resolution imaging. However, achieving wide-angle, high-speed, and high-precision 2D beam control within a compact THz platform remains a significant challenge. In this work, we experimentally demonstrate an optoelectronic

2 \times 2

THz antenna array that enables flexible 2D beam steering, beam hopping, and beam scanning around the 300 GHz band. This work employs a

2 \times 2

microstrip patch antenna (MPA) array directly driven by InGaAs/InP UTC-PDs on a silicon carbide (SiC) substrate. The relative phases of the four radiating elements are precisely programmed using an optical phased array (OPA), which provides fully decoupled and low-latency phase control in the optical domain. Experimentally, we demonstrate 2D beam steering and 2D beam hopping among three representative directions at a polar angle of

25^{\circ}

and azimuth angles of

60^{\circ}

,

180^{\circ}

, and

300^{\circ}

. Furthermore, continuous 2D beam scanning at a fixed polar angle of

25^{\circ}

is achieved, enabling a full

360^{\circ}

azimuth sweep within 0.43 s while maintaining high beam quality. These results confirm that the proposed UTC-PD based

2 \times 2

MPA array provides a practical and robust approach for 2D THz beam manipulation, and offers strong potential for future 6G wireless links and THz imaging applications. Full article

(This article belongs to the Special Issue Smart Antennas and Systems for 5G and Beyond: Latest Advances and Prospects, 2nd Edition)

► Show Figures

Figure 1

17 pages, 3010 KB

Open AccessArticle

Research on Transient Stability Optimization Control of Photovoltaic–Storage Virtual Synchronous Generators

by Fen Gong, Xiangyang Xia, Xianliang Luo, Wei Hu and Yijie Zhu

Electronics 2025, 14(24), 4979; https://doi.org/10.3390/electronics14244979 - 18 Dec 2025

Viewed by 240

Abstract

In the case of small disturbances in the power grid, virtual synchronous generators (VSGs) often exhibit active power steady-state errors and significant frequency overshoot, and it is difficult to balance the reduction of active power steady-state errors and the mitigation of frequency overshoot. [...] Read more.

In the case of small disturbances in the power grid, virtual synchronous generators (VSGs) often exhibit active power steady-state errors and significant frequency overshoot, and it is difficult to balance the reduction of active power steady-state errors and the mitigation of frequency overshoot. This paper proposes an improved control method based on active power differential compensation (APDC). First, an active power differential compensation loop is introduced, effectively addressing the issues of active power steady-state deviation and frequency overshoot caused by fixed parameters in the traditional VSG. Secondly, by incorporating a fuzzy logic control (FLC) algorithm, an adaptive PID tuning strategy is proposed as a replacement for the traditional fixed virtual inertia; the PID parameters are dynamically adjusted in real time according to the power–angle deviation and its rate of change, thereby enhancing the small-disturbance dynamic performance of the VSG. Finally, MATLAB R2020b/Simulink simulations and StarSim hardware-in-the-loop simulations validate the effectiveness and accuracy of the proposed control strategy. Simulation results indicate that, compared to traditional control strategies, under peak regulation conditions, the frequency overshoot is reduced by approximately 4.4%, and the active power overshoot is reduced by approximately 5%; under frequency regulation conditions, the frequency overshoot is reduced by approximately 0.26%, and the power overshoot is reduced by approximately 12%. Full article

(This article belongs to the Special Issue Stability Analysis and Optimal Operation Strategy for Power Systems with High Renewable Energy Penetration)

► Show Figures

Figure 1

25 pages, 21508 KB

Open AccessArticle

PVConv: Enhancing Depthwise Separable Convolution via Preference-Value Learning for Similar-Feature Discrimination

by Weixiong Peng, Bingyan Li, Ping Wang, Huiping Huang, Yangyang Zou and Xiaoli Qiao

Electronics 2025, 14(24), 4978; https://doi.org/10.3390/electronics14244978 - 18 Dec 2025

Viewed by 311

Abstract

Depthwise Separable Convolution (DSC) is widely used due to its significant reduction in parameters and computational cost. However, the depthwise convolution process leads to a decrease in spatial information integration, limiting the network’s expressive power. To address this, we propose a novel Preference-Value [...] Read more.

Depthwise Separable Convolution (DSC) is widely used due to its significant reduction in parameters and computational cost. However, the depthwise convolution process leads to a decrease in spatial information integration, limiting the network’s expressive power. To address this, we propose a novel Preference-Value Convolution (PVConv) to enhance DSC’s expressiveness. By integrating PVConv into DSC, we introduce the Preference-Value Depthwise Separable Convolution (PVDSC) structure. We integrate both DSC and PVDSC into the YOLOv8 framework and conduct experiments on a beverage container dataset containing visually similar object categories and background interference. Results show that, with minimal increase in parameters and computational cost, introducing preference values significantly improves detection accuracy, F1 score, and attention consistency, especially at high IoU thresholds (mAP@50:95), where object localization is greatly enhanced and certain metrics even surpass complex baseline models. Overall, PVConv significantly enhances the expressiveness of DSC-based networks while maintaining low computational overhead, with promising applications. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

24 pages, 9315 KB

Open AccessArticle

Secure LoRa-Based Transmission System: An IoT Solution for Smart Homes and Industries

by Sebastian Ryczek and Maciej Sobieraj

Electronics 2025, 14(24), 4977; https://doi.org/10.3390/electronics14244977 - 18 Dec 2025

Viewed by 512

Abstract

This article addresses the lack of low-cost, secure image-transmission solutions for IoT systems in remote environments. The design and implementation of a complete LoRa-based transmission system using ESP32 microcontrollers and Ebyte E220 modules, featuring AES-CBC encryption, HMAC integrity protection, and a custom retransmission [...] Read more.

This article addresses the lack of low-cost, secure image-transmission solutions for IoT systems in remote environments. The design and implementation of a complete LoRa-based transmission system using ESP32 microcontrollers and Ebyte E220 modules, featuring AES-CBC encryption, HMAC integrity protection, and a custom retransmission protocol, are presented. The system achieves 100% packet delivery ratio (PDR) for 20 kB images over distances exceeding 2 km under line-of-sight conditions, with functional transmission up to 4.1 km. Image transmission time ranges from 35 s (0.1 m) to 110 s (600 m), while energy consumption increases from 4.95 mWh to 15.18 mWh. Critically, encryption imposes less than 1% overhead on total energy consumption. Unlike prior work focusing on isolated components, this article provides a complete, deployable architecture combining (i) low-cost hardware (<USD 50 total), (ii) long-range LoRa communication, (iii) custom reliability mechanisms for fragmenting 20 kB images into 100 packets, and (iv) end-to-end cryptographic protection, all evaluated experimentally across multi-kilometer distances. These findings demonstrate that secure long-range image transmission using commodity hardware is feasible and scalable for smart home and industrial monitoring applications. Full article

(This article belongs to the Special Issue The Future of Unmanned Aerial Vehicles (UAV) and IoT Technologies: Challenges, and Innovations)

► Show Figures

Figure 1

16 pages, 3581 KB

Open AccessArticle

Enabling Fast Frequency Response with Adaptive Demand-Side Resource Control: Strategy and Field-Testing Validation

by Shunxin Wei, Yingqi Liang, Zhendong Zhao, Yan Guo, Jiyu Huang, Ying Xue and Yiping Chen

Electronics 2025, 14(24), 4976; https://doi.org/10.3390/electronics14244976 - 18 Dec 2025

Viewed by 249

Abstract

With the large-scale integration of new energy and power electronic devices into power systems, frequency stability has become an increasingly critical concern. To maintain frequency stability while mitigating the high capital expenditure of energy storage systems (ESSs), this paper develops a control framework [...] Read more.

With the large-scale integration of new energy and power electronic devices into power systems, frequency stability has become an increasingly critical concern. To maintain frequency stability while mitigating the high capital expenditure of energy storage systems (ESSs), this paper develops a control framework centered on edge energy management terminals (EEMTs). The design is based on a demonstration project in which distributed energy resources (DERs) and flexible loads collaboratively provide frequency regulation. A monitoring station is implemented to make fast frequency response (FFR) resources dispatchable, detectable, measurable, and tradable. Furthermore, a control strategy tailored for building- and factory-level applications is proposed. This strategy enables real-time optimal scheduling of DERs and flexible loads through coordinated communication between EEMTs and net load units (NLUs). Two field tests further demonstrate the effectiveness and advantages of the proposed approach. In addition, this paper proposes a coordinated scheme in which wind farms and NLUs jointly participate in frequency regulation, aiming to mitigate the response delay of NLUs and the secondary frequency drop observed in wind farms. The feasibility and benefits of this scheme are validated through experimental tests. Full article

(This article belongs to the Section Systems & Control Engineering)

► Show Figures

Figure 1

22 pages, 2574 KB

Open AccessArticle

FedTULGAC: A Federated Learning Method for Trajectory User Linking Based on Graph Attention and Clustering

by Haitao Zhang, Yang Xu, Huixiang Jiang, Yuanjian Liu, Weigang Wang, Yi Li, Yuhao Luo and Yuxuan Ge

Electronics 2025, 14(24), 4975; https://doi.org/10.3390/electronics14244975 - 18 Dec 2025

Viewed by 185

Abstract

Trajectory User Linking (TUL) is a pivotal technology for identifying and associating the trajectory information from the same user across various data sources. To address the privacy leakage challenges inherent in traditional TUL methods, this study introduces a novel federated learning-based TUL method: [...] Read more.

Trajectory User Linking (TUL) is a pivotal technology for identifying and associating the trajectory information from the same user across various data sources. To address the privacy leakage challenges inherent in traditional TUL methods, this study introduces a novel federated learning-based TUL method: FedTULGAC. This approach utilizes a federated learning framework to aggregate model parameters, thereby avoiding the sharing of local data. Within this framework, a Graph Attention-based Trajectory User Linking and Embedding Regression (GATULER) model and an FL-DBSCAN clustering algorithm are designed and integrated to capture short-term temporal dependencies in user movement trajectories and to handle the non-independent and identically distributed (Non-IID) characteristics of client-side data. Experimental results on the synthesized datasets demonstrate that the proposed method achieves the highest prediction accuracy compared to the baseline models and maintains stable performance with minimal sensitivity to variations in client selection ratios, which reveals its effectiveness in bandwidth-constrained real-world applications. Full article

(This article belongs to the Special Issue Advances in Deep Learning for Graph Neural Networks)

► Show Figures

Figure 1

21 pages, 2686 KB

Open AccessArticle

A Deep Learning Approach to Classifying User Performance in BCI Gaming

by Aimilia Ntetska, Anastasia Mimou, Katerina D. Tzimourta, Pantelis Angelidis and Markos G. Tsipouras

Electronics 2025, 14(24), 4974; https://doi.org/10.3390/electronics14244974 - 18 Dec 2025

Viewed by 414

Abstract

Brain–Computer Interface (BCI) systems are rapidly evolving and increasingly integrated into interactive environments such as gaming and Virtual/Augmented Reality. In such applications, user adaptability and engagement are critical. This study applies deep learning to predict user performance in a 3D BCI-controlled game using [...] Read more.

Brain–Computer Interface (BCI) systems are rapidly evolving and increasingly integrated into interactive environments such as gaming and Virtual/Augmented Reality. In such applications, user adaptability and engagement are critical. This study applies deep learning to predict user performance in a 3D BCI-controlled game using pre-game Motor Imagery (MI) electroencephalographic (EEG) recordings. A total of 72 EEG recordings were collected from 36 participants, 17 using the Muse 2 headset and 19 using the Emotiv Insight device, during left and right hand MI tasks. The signals were preprocessed and transformed into time–frequency spectrograms, which served as inputs to a custom convolutional neural network (CNN) designed to classify users into three performance levels: low, medium, and high. The model achieved classification accuracies of 83% and 95% on Muse 2 and Emotiv Insight data, respectively, at the epoch level, and 75% and 84% at the subject level, using LOSO-CV. These findings demonstrate the feasibility of using deep learning on MI EEG data to forecast user performance in BCI gaming, enabling adaptive systems that enhance both usability and user experience. Full article

(This article belongs to the Special Issue Research and Application of Neural Networks in Human-Computer Interaction)

► Show Figures

Figure 1

21 pages, 2001 KB

Open AccessArticle

A Unified Fault-Tolerant Batch Authentication Scheme for Vehicular Networks

by Yifan Zhao, Hu Liu, Xinghua Li, Yunwei Wang, Zhe Ren and Peiyao Wang

Electronics 2025, 14(24), 4973; https://doi.org/10.3390/electronics14244973 - 18 Dec 2025

Viewed by 297

Abstract

This paper proposes a unified fault-tolerant batch authentication scheme for vehicular networks, designed to address key limitations in existing approaches, namely the segregation between in-vehicle and V2I authentication scenarios and the lack of fault tolerance in traditional batch authentication methods. Based on a [...] Read more.

This paper proposes a unified fault-tolerant batch authentication scheme for vehicular networks, designed to address key limitations in existing approaches, namely the segregation between in-vehicle and V2I authentication scenarios and the lack of fault tolerance in traditional batch authentication methods. Based on a hardware–software co-design philosophy, the scheme deeply integrates the security features of hardware such as Tamper-Proof Devices (TPDs) and Physical Unclonable Functions (PUFs) with the efficiency of cryptographic primitives like Aggregate Message Authentication Codes (MACs) and the Chinese Remainder Theorem (CRT). It establishes an end-to-end, integrated authentication framework spanning from in-vehicle electronic control units (ECUs) to external roadside units (RSUs), effectively meeting the diverse requirements for secure and efficient authentication among the three core entities involved in Internet of Vehicles (IoV) data collection: in-vehicle ECUs, vehicle gateways, and RSUs. Security analysis demonstrates that the proposed scheme fulfills the necessary security requirements. And extensive experimental results confirm its high efficiency and practical utility. Full article

(This article belongs to the Special Issue Cryptography and Computer Security)

► Show Figures

Figure 1

16 pages, 1381 KB

Open AccessArticle

Dual Routing Mixture-of-Experts for Multi-Scale Representation Learning in Multimodal Emotion Recognition

by Da-Eun Chae and Seok-Pil Lee

Electronics 2025, 14(24), 4972; https://doi.org/10.3390/electronics14244972 - 18 Dec 2025

Viewed by 300

Abstract

Multimodal emotion recognition (MER) often relies on single-scale representations that fail to capture the hierarchical structure of emotional signals. This paper proposes a Dual Routing Mixture-of-Experts (MoE) model that dynamically selects between local (fine-grained) and global (contextual) representations extracted from speech and text [...] Read more.

Multimodal emotion recognition (MER) often relies on single-scale representations that fail to capture the hierarchical structure of emotional signals. This paper proposes a Dual Routing Mixture-of-Experts (MoE) model that dynamically selects between local (fine-grained) and global (contextual) representations extracted from speech and text encoders. The framework first obtains local–global embeddings using WavLM and RoBERTa, then employs a scale-aware routing mechanism to activate the most informative expert before bidirectional cross-attention fusion. Experiments on the IEMOCAP dataset show that the proposed model achieves stable performance across all folds, reaching an average unweighted accuracy (UA) of 75.27% and weighted accuracy (WA) of 74.09%. The model consistently outperforms single-scale baselines and simple concatenation methods, confirming the importance of dynamic multi-scale cue selection. Ablation studies highlight that neither local-only nor global-only representations are sufficient, while routing behavior analysis reveals emotion-dependent scale preferences—such as strong reliance on local acoustic cues for anger and global contextual cues for low-arousal emotions. These findings demonstrate that emotional expressions are inherently multi-scale and that scale-aware expert activation provides a principled approach beyond conventional single-scale fusion. Full article

(This article belongs to the Special Issue Artificial Intelligence and Advanced Signal Processing Techniques and Their Applications)

► Show Figures

Figure 1

22 pages, 1046 KB

Open AccessArticle

What if, Behind the Curtain, There Is Only an LLM? A Holistic Evaluation of TinyLlama-Generated Synthetic Cyber Threat Intelligence

by Zuzanna Pietrzak, Krzysztof Mączka and Marcin Niemiec

Electronics 2025, 14(24), 4971; https://doi.org/10.3390/electronics14244971 - 18 Dec 2025

Viewed by 477

Abstract

The generation of synthetic cyber threat intelligence (CTI) has emerged as a significant area of research, particularly regarding the capacity of large language models (LLMs) to produce realistic yet deceptive security content. This study explores both the generative and evaluative aspects of CTI [...] Read more.

The generation of synthetic cyber threat intelligence (CTI) has emerged as a significant area of research, particularly regarding the capacity of large language models (LLMs) to produce realistic yet deceptive security content. This study explores both the generative and evaluative aspects of CTI synthesis by employing a custom-developed detection system and publicly accessible LLMs. The evaluation combined automated analysis with a human study involving cybersecurity professionals. The results indicate that even a compact, resource-efficient fine-tuned model can generate highly convincing CTI misinformation capable of deceiving experts and AI-based classifiers. Human participants achieved an average accuracy around 50% in distinguishing between authentic and generated CTI reports. However, the proposed hybrid detection model achieved 98.5% accuracy on the test set and maintained strong generalization with 88.5% accuracy on unseen data. These findings demonstrate both the potential of lightweight models to generate credible CTI narratives and the effectiveness of specialized detection systems in mitigating such threats. The study underscores the growing risk of harmful misinformation in AI-driven CTI and highlights the importance of incorporating robust validation mechanisms within cybersecurity infrastructures to enhance defense resilience. Full article

(This article belongs to the Special Issue Enhancing Cybersecurity: Advanced Attack Detection and Defense Techniques)

► Show Figures

Figure 1

19 pages, 5899 KB

Open AccessArticle

Small-Signal Modeling of Asymmetric PWM Control-Based Parallel Resonant Converter

by Na-Yeon Kim and Kui-Jun Lee

Electronics 2025, 14(24), 4970; https://doi.org/10.3390/electronics14244970 - 18 Dec 2025

Viewed by 255

Abstract

This paper proposes a small-signal model of a DC–DC parallel resonant converter operating in continuous conduction mode based on asymmetric pulse-width modulation (APWM) under light-load conditions. The parallel resonant converter enables soft switching and no-load control over a wide load range because the [...] Read more.

This paper proposes a small-signal model of a DC–DC parallel resonant converter operating in continuous conduction mode based on asymmetric pulse-width modulation (APWM) under light-load conditions. The parallel resonant converter enables soft switching and no-load control over a wide load range because the resonant capacitor is connected in parallel with the load. However, the resonant energy required for soft switching is already sufficient, and the current flowing through the resonant tank is independent of the load magnitude; therefore, as the load decreases, the energy that is not delivered to the load and instead circulates meaninglessly inside the resonant tank increases. This results in conduction loss and reduced efficiency. To address this issue, APWM with a fixed switching frequency is required, which reduces circulating energy and improves efficiency under light-load conditions. Precise small-signal modeling is required to optimize the APWM controller. Unlike PFM or PSFB, APWM includes not only sine components but also DC and cosine components in the control signal due to its asymmetric switching characteristics, and this study proposes a small-signal model that can relatively accurately reflect these multi-harmonic characteristics. The proposed model is derived based on the Extended Describing Function (EDF) concept, and the derived transfer function is useful for systematically analyzing the dynamic characteristics of the APWM-based parallel resonant converter. In addition, it provides information that can systematically analyze the dynamic characteristics of various APWM-based resonant converters and control signals that reflect various harmonic characteristics, and it can be widely applied to future control design and analysis studies. The validity of the model is verified through MATLAB (R2025b) and PLECS (4.7.5) switching-model simulations and experimental results, confirming its high accuracy and practicality. Full article

(This article belongs to the Special Issue New Insights in Power Electronics: Prospects and Challenges)

► Show Figures

Figure 1

27 pages, 2804 KB

Open AccessArticle

Intelligent Cooperative Perception Technology for Vehicles and Experiments Based on V2V/V2I Semantic Communication

by Cheng Li, Huiping Liu, Qiqi Jia, Lei Xiong and Hao Wu

Electronics 2025, 14(24), 4969; https://doi.org/10.3390/electronics14244969 - 18 Dec 2025

Viewed by 423

Abstract

In recent years, intelligent driving has attracted more and more attention due to its potential to revolutionize transportation safety and efficiency, emerging as a disruptive technology that reshapes the future landscape of transportation. Environmental perception serves as the primary and fundamental cornerstone of [...] Read more.

In recent years, intelligent driving has attracted more and more attention due to its potential to revolutionize transportation safety and efficiency, emerging as a disruptive technology that reshapes the future landscape of transportation. Environmental perception serves as the primary and fundamental cornerstone of intelligent driving systems. To address the intrinsic blind spots in environmental perception, this paper presents a vehicle collaborative perception approach based on Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) semantic communication. Specifically, a Transformer-based semantic segmentation technique is proposed for application to images acquired from surrounding vehicles and ground-based cameras. Subsequently, the generated semantic segmentation maps are transmitted via V2V/V2I communication. In the receiver, a semantic-guided image reconstruction technique based on Generative Adversarial Networks (GANs) is developed to generate images with high realism. The generated Image images can be further fused with locally perceived data, facilitating intelligent collaborative perception. This method achieves effective elimination of blind spots. Furthermore, as only semantic segmentation maps—with a data size significantly smaller than that of raw images—are transmitted instead of the latter, it exhibits excellent adaptability to the dynamically time-varying characteristics of V2V/V2I channels. Even in poor channel condition, the proposed method maintains high reliability and real-time performance. Full article

(This article belongs to the Special Issue Next-Generation Wireless Networks: Advances in 5G, 6G, and Broadband Communication)

► Show Figures

Figure 1

25 pages, 9939 KB

Open AccessArticle

RAC-RTDETR: A Lightweight, Efficient Real-Time Small-Object Detection Algorithm for Steel Surface Defect Detection

by Zhenping Xu and Nengxi Wang

Electronics 2025, 14(24), 4968; https://doi.org/10.3390/electronics14244968 - 18 Dec 2025

Viewed by 479

Abstract

Steel, a fundamental material in modern industry, is widely used across manufacturing, construction, and energy sectors. Steel surface defects exhibit characteristics such as multiple classes, multi-scale features, small detection targets, and low-contrast backgrounds, making detection difficult. We propose RAC-RTDETR, a lightweight real-time detection [...] Read more.

Steel, a fundamental material in modern industry, is widely used across manufacturing, construction, and energy sectors. Steel surface defects exhibit characteristics such as multiple classes, multi-scale features, small detection targets, and low-contrast backgrounds, making detection difficult. We propose RAC-RTDETR, a lightweight real-time detection algorithm designed for accurately identifying small surface defects on steel. Key improvements include: (1) The ARNet network, combining the ADown module and the RepNCSPELAN4-CAA module with a CAA-based attention mechanism, results in a lighter backbone network with better feature extraction and enhanced small-object detection by integrating contextual information; (2) The novel AIFI-ASMD module, composed of Adaptive Sparse Self-Attention (ASSA), Spatially Enhanced Feedforward Network (SEFN), Multi-Cognitive Visual Adapter (Mona), and Dynamic Tanh (DyT), optimizes feature interactions at different scales, reduces noise interference, and improves spatial awareness and long-range dependency modeling for better detection of multi-scale objects; (3) The Converse2D upsampling module replaces traditional upsampling methods, preserving details and enhancing small-object recognition in low-contrast, sparse feature scenarios. Experimental results on the NEU-DET and GC10-DET datasets show that RAC-RTDETR outperforms baseline models with MAP improvements of 3.56% and 3.47%, a 36.18% reduction in Parameters, a 40.70% decrease in GFLOPs, and a 7.96% increase in FPS. Full article

(This article belongs to the Special Issue Advances in Real-Time Object Detection and Tracking)

► Show Figures

Figure 1

22 pages, 2236 KB

Open AccessArticle

An AI-Driven System for Learning MQTT Communication Protocols with Python Programming

by Zihao Zhu, Nobuo Funabiki, Htoo Htoo Sandi Kyaw, I Nyoman Darma Kotama, Anak Agung Surya Pradhana, Alfiandi Aulia Rahmadani and Noprianto

Electronics 2025, 14(24), 4967; https://doi.org/10.3390/electronics14244967 - 18 Dec 2025

Viewed by 425

Abstract

With rapid developments of wireless communication and Internet of Things (IoT) technologies, an increasing number of devices and sensors are interconnected, generating massive amounts of data in real time. Among the underlying protocols, Message Queuing Telemetry Transport (MQTT) has become a widely adopted [...] Read more.

With rapid developments of wireless communication and Internet of Things (IoT) technologies, an increasing number of devices and sensors are interconnected, generating massive amounts of data in real time. Among the underlying protocols, Message Queuing Telemetry Transport (MQTT) has become a widely adopted lightweight publish–subscribe standard due to its simplicity, minimal overhead, and scalability. Then, understanding such protocols is essential for students and engineers engaging in IoT application system designs. However, teaching and learning MQTT remains challenging for them. Its asynchronous architecture, hierarchical topic structure, and constituting concepts such as retained messages, Quality of Service (QoS) levels, and wildcard subscriptions are often difficult for beginners. Moreover, traditional learning resources emphasize theory and provide limited hands-on guidance, leading to a steep learning curve. To address these challenges, we propose an AI-assisted, exercise-based learning platform for MQTT. This platform provides interactive exercises with intelligent feedback to bridge the gap between theory and practice. To lower the barrier for learners, all code examples for executing MQTT communication are implemented in Python for readability, and Docker is used to ensure portable deployments of the MQTT broker and AI assistant. For evaluations, we conducted a usability study using two groups. The first group, who has no prior experience, focused on fundamental concepts with AI-guided exercises. The second group, who has relevant background, engaged in advanced projects to apply and reinforce their knowledge. The results show that the proposed platform supports learners at different levels, reduces frustrations, and improves both engagement and efficiency. Full article

(This article belongs to the Special Issue Advanced System Architectures and AI-Driven Innovations for Next-Generation Computing)

► Show Figures

Figure 1

25 pages, 1338 KB

Open AccessArticle

Aquaculture Real Estate Price Prediction System: Machine Learning Approach for PropTech

by Ngoc-Bao-Van Le, Jun-Ho Huh, Han-Jong Ko and Yuanyuan Liu

Electronics 2025, 14(24), 4966; https://doi.org/10.3390/electronics14244966 - 18 Dec 2025

Viewed by 459

Abstract

The aquaculture industry is growing strongly, especially in facilities and technical infrastructure. Aquaculture real estate serves as an important component for these developments. Although the aquaculture real estate sector represents a dynamic market with substantial transaction volumes and high economic value, this sector [...] Read more.

The aquaculture industry is growing strongly, especially in facilities and technical infrastructure. Aquaculture real estate serves as an important component for these developments. Although the aquaculture real estate sector represents a dynamic market with substantial transaction volumes and high economic value, this sector has received less attention. This paper proposes an aquaculture real estate price prediction system that utilizes a machine learning approach. First, this paper focuses on collecting and analyzing data from aquaculture real estate trading websites and platforms. The dataset encompasses prices and features of aquaculture real estate from 2021 to the beginning of 2025, specifically in Southern Vietnam, obtained from various Vietnam real estate websites. We conduct experiments and assess six different models, including Random Forest, CatBoost, Hybrid, XGBoost, LightGBM and Linear Regression, to predict aquaculture real estate prices. Results from testing revealed that the Random Forest is the best model, due to the nature of aquaculture real estate data, which includes many categories associated with the location of land, the type of land, and the legal status, and has many complicated nonlinear relationships among its attributes. Following this, hyperparameter tuning using Optuna (50 trials, 5-fold cross-validation) can further enhance the model’s performance. The experimental results demonstrate that the Stacking Ensemble model achieved the best performance after hyperparameter tuning, with RMSE of 4394.77 million, MAE of 2114.19 million, and R² of 0.5738, representing improvements of 3.3%, 6.0%, and 5.4%. The paper will provide a more comprehensive view of the aquaculture real estate market and support sustainable development for this industry. Full article

(This article belongs to the Special Issue Artificial Intelligence-Driven Emerging Applications)

► Show Figures

Figure 1

14 pages, 13781 KB

Open AccessArticle

Neurosynaptic Core Prototype for Memristor Crossbar Arrays Diagnostics

by Ivan V. Alyaev, Igor A. Surazhevsky, Dmitry V. Ichyotkin, Vladimir V. Rylkov and Vyacheslav A. Demin

Electronics 2025, 14(24), 4965; https://doi.org/10.3390/electronics14244965 - 18 Dec 2025

Viewed by 594

Abstract

The use of neural network technologies is becoming more widespread today, from automating routine office tasks to developing new medicines. However, at the same time, the load on power grids and generation systems increases significantly, which, alongside the desire to increase equipment performance, [...] Read more.

The use of neural network technologies is becoming more widespread today, from automating routine office tasks to developing new medicines. However, at the same time, the load on power grids and generation systems increases significantly, which, alongside the desire to increase equipment performance, further motivates the development of specialized architectures for hardware implementation and training of neural networks. Memristor-based systems are considered one of the promising areas for creating energy-efficient platforms for artificial intelligence (AI) due to their ability to implement in-memory computing at the hardware level. A crucial step towards the realization of such systems is the comprehensive characterization of memristive devices. This work presents the implementation of a hardware platform for the automated measurement of key memristor characteristics, including current-voltage (I-V) curves, retention time, and endurance. The developed device features a modular architecture for validating the functionality of individual subsystems and incorporates a unipolar pulse switching scheme to mitigate the risk of gate-oxide breakdown in 1T1R active arrays that can occur when applying negative voltages during synaptic weight programming. Full article

(This article belongs to the Section Artificial Intelligence)

► Show Figures

Figure 1

15 pages, 2369 KB

Open AccessArticle

The Effect of Tactile Feedback on the Manipulation of a Remote Robotic Arm via a Haptic Glove

by Christos Papakonstantinou, Konstantinos Giannakos, George Kokkonis and Maria S. Papadopoulou

Electronics 2025, 14(24), 4964; https://doi.org/10.3390/electronics14244964 - 18 Dec 2025

Viewed by 640

Abstract

This paper investigates the effect of tactile feedback on the power efficiency and timing of controlling a remote robotic arm using a custom-built haptic glove. The glove integrates flex sensors to monitor finger movements and vibration motors to provide tactile feedback to the [...] Read more.

This paper investigates the effect of tactile feedback on the power efficiency and timing of controlling a remote robotic arm using a custom-built haptic glove. The glove integrates flex sensors to monitor finger movements and vibration motors to provide tactile feedback to the user. Communication with the robotic arm is established via the ESP-NOW protocol using an Arduino Nano ESP32 microcontroller (Arduino, Turin, Italy). This study examines the impact of tactile feedback on task performance by comparing precision, completion time, and power efficiency in object manipulation tasks with and without feedback. Experimental results demonstrate that tactile feedback significantly enhances the user’s control accuracy, reduces task execution time, and enables the user to control hand movement during object grasping scenarios precisely. It also highlights its importance in teleoperation systems. These findings have implications for improving human–robot interaction in remote manipulation scenarios, such as assistive robotics, remote surgery, and hazardous environment operations. Full article

(This article belongs to the Special Issue Advanced Research in Technology and Information Systems, 2nd Edition)

► Show Figures

Figure 1

19 pages, 1802 KB

Open AccessArticle

Technique for Sub-mHz Low-Frequency Corner in Capacitively Coupled Instrumentation Amplifiers

by Miguel Barrales-Romero, José Luis Valtierra, Esteban Tlelo-Cuautle and Alejandro Díaz-Sánchez

Electronics 2025, 14(24), 4963; https://doi.org/10.3390/electronics14244963 - 18 Dec 2025

Viewed by 416

Abstract

This work introduces a tunable technique to push the low-frequency corner (

f_{L}

) of capacitively coupled instrumentation amplifiers (CCIAs) to the sub-mHz range for emerging biosensing applications. The proposed approach combines Complementary Transimpedance Boosting (CTB) to limit the DC feedback current [...] Read more.

This work introduces a tunable technique to push the low-frequency corner (

f_{L}

) of capacitively coupled instrumentation amplifiers (CCIAs) to the sub-mHz range for emerging biosensing applications. The proposed approach combines Complementary Transimpedance Boosting (CTB) to limit the DC feedback current and segmented duty-cycled resistors (SDR) for tunable resistance. The CTB-SDR technique achieves a stable effective post-layout pseudo-resistance of

535.8 T Ω

, equivalent to

f_{L} = 660 μ Hz

while occupying

0.062 {mm}^{2}

in a 180 nm process. According to JESD91 standards, it shows a standard deviation of

0.19 mHz

under post-layout Monte Carlo + process analysis,

1.1 %

spread under voltage variations

(\pm 5.56 % V_{D D})

and

6.2 %

under temperature variations

(- 20 ° C, 27 ° C,

and

60 ° C)

. In addition, duty-cycling calibration can compensate for worst-case process corner variations and mismatch-induced feedback instability. Full article

(This article belongs to the Special Issue Design and Applications of Nonlinear Circuits and Systems, 2nd Edition)

► Show Figures

Figure 1

18 pages, 1196 KB

Open AccessArticle

Automatic Metadata Extraction Leveraging Large Language Models in Digital Humanities

by Adriana Morejón, Borja Navarro-Colorado, Carmen García-Barceló, Alberto Berenguer, David Tomás and Jose-Norberto Mazón

Electronics 2025, 14(24), 4962; https://doi.org/10.3390/electronics14244962 - 18 Dec 2025

Viewed by 612

Abstract

DCAT-based data ecosystems, such as open data portals and data spaces, have shown their potential to foster data economy by supporting the FAIR (Findability, Accessibility, Interoperability, Reusability) principles. Nevertheless, there are domains where metadata are tailored to specific semantics of the domain, resulting [...] Read more.

DCAT-based data ecosystems, such as open data portals and data spaces, have shown their potential to foster data economy by supporting the FAIR (Findability, Accessibility, Interoperability, Reusability) principles. Nevertheless, there are domains where metadata are tailored to specific semantics of the domain, resulting in the absence of DCAT-based catalogs that adhere to FAIR principles. A particularly relevant case is that of the digital humanities, where texts encoded in TEI (Text Encoding Initiative) constitute a consolidated standard in the field of literature. However, TEI metadata are not always well aligned with the FAIR principles, nor easily integrated into interoperable catalogs that enable seamless combination with external datasets. To address this gap, our approach aims to (i) generate DCAT catalogs derived from TEI by identifying which metadata can be mapped and how, and (ii) explore the use of Large Language Models (LLMs) to assist in the generation and enrichment of metadata when transforming TEI to DCAT. Our approach contributes to catalog-level harmonization, enabling domain-specific standards such as TEI to be aligned with cross-domain standards like DCAT, thus facilitating adherence to FAIR principles. Full article

(This article belongs to the Special Issue Artificial Intelligence and Pattern Recognition for Intelligent Systems)

► Show Figures

Figure 1

1 pages, 125 KB

Open AccessCorrection

Correction: Zhuang et al. A Generalized Optimization Scheme for Memory-Side Prefetching to Enhance System Performance. Electronics 2025, 14, 2811

by Yuzhi Zhuang, Ming Zhang and Binghao Wang

Electronics 2025, 14(24), 4961; https://doi.org/10.3390/electronics14244961 - 18 Dec 2025

Viewed by 148

Abstract

In the original publication [...] Full article

21 pages, 1991 KB

Open AccessArticle

Zero-Shot Resume–Job Matching with LLMs via Structured Prompting and Semantic Embeddings

by Panagiotis Skondras, Panagiotis Zervas and Giannis Tzimas

Electronics 2025, 14(24), 4960; https://doi.org/10.3390/electronics14244960 - 17 Dec 2025

Viewed by 770

Abstract

In this article, we present a tool for matching resumes to job posts and vice versa (job post to resumes). With minor modifications, it may also be adapted to other domains where text matching is necessary. This tool may help organizations save time [...] Read more.

In this article, we present a tool for matching resumes to job posts and vice versa (job post to resumes). With minor modifications, it may also be adapted to other domains where text matching is necessary. This tool may help organizations save time during the hiring process, as well as assist applicants by allowing them to match their resumes to job posts they have selected. To achieve text matching without any model training (zero-shot matching), we constructed dynamic structured prompts that consisted of unstructured and semi-structured job posts and resumes based on specific criteria, and we utilized the Chain of Thought (CoT) technique on the Mistral model (open-mistral-7b). In response, the model generated structured (segmented) job posts and resumes. Then, the job posts and resumes were cleaned and preprocessed. We utilized state-of-the-art sentence similarity models hosted on Hugging face (nomic-embed-text-v1-5 and google-embedding-gemma-300m) through inference endpoints to create sentence embeddings for each resume and job post segment. We used the cosine similarity metric to determine the optimal matching, and the matching operation was applied to eleven different occupations. The results we achieved reached up to 87% accuracy for some of the occupations and underscore the potential of zero-shot techniques in text matching utilizing LLMs. The dataset we used was from indeed.com, and the Spring AI framework was used for the implementation of the tool. Full article

(This article belongs to the Special Issue Advances in Text Mining and Analytics)

► Show Figures

Figure 1

17 pages, 9727 KB

Open AccessArticle

An Energy-Efficient Neuromorphic Processor Using Unified Refractory Control-Based NoC for Edge AI

by Su-Hwan Na and Dong-Sun Kim

Electronics 2025, 14(24), 4959; https://doi.org/10.3390/electronics14244959 - 17 Dec 2025

Viewed by 400

Abstract

Neuromorphic computing has emerged as a promising paradigm for edge AI systems owing to its event-driven operation and high energy efficiency. However, conventional spiking neural network (SNN) architectures often suffer from redundant computation and inefficient power control, particularly during on-chip learning. This paper [...] Read more.

Neuromorphic computing has emerged as a promising paradigm for edge AI systems owing to its event-driven operation and high energy efficiency. However, conventional spiking neural network (SNN) architectures often suffer from redundant computation and inefficient power control, particularly during on-chip learning. This paper proposes a network-on-chip (NoC) architecture featuring a unified refractory-enabled neuron (UREN)-based router that globally coordinates spike-driven computation across multiple neuron cores. The router applies a unified refractory time to all neurons following a winner spike event, effectively enabling clock gating and suppressing redundant activity. The proposed design adopts a star-routing topology with multicasting support and integrates nearest-neighbor spike-timing-dependent plasticity (STDP) for local online learning. FPGA-based experiments demonstrate a 30% reduction in computation and 86.1% online classification accuracy on the MNIST dataset compared with baseline SNN implementations. These results confirm that the UREN-based router provides a scalable and power-efficient neuromorphic processor architecture, well suited for energy-constrained edge AI applications. Full article

► Show Figures

Figure 1

27 pages, 1242 KB

Open AccessArticle

Dual-Tower TTP Semantic Matching Method Based on Soft–Hard Label Supervision and Gated Binary Interaction

by Zhenghao Qian, Fengzheng Liu, Mingdong He, Bo Li and Yinghai Zhou

Electronics 2025, 14(24), 4958; https://doi.org/10.3390/electronics14244958 - 17 Dec 2025

Viewed by 283

Abstract

Existing methods for identifying Tactics, Techniques, and Procedures (TTPs) from complex cyber-attack descriptions face three core challenges: (1) severe semantic asymmetry between unstructured attack narratives and standardized TTP definitions; (2) continuously distributed semantic relations that cannot be fully captured by hard-label supervision; and [...] Read more.

Existing methods for identifying Tactics, Techniques, and Procedures (TTPs) from complex cyber-attack descriptions face three core challenges: (1) severe semantic asymmetry between unstructured attack narratives and standardized TTP definitions; (2) continuously distributed semantic relations that cannot be fully captured by hard-label supervision; and (3) an open, long-tailed TTP taxonomy that impairs model generalization. To address these limitations, we introduce DTGBI-TM, a lightweight dual-tower semantic matching framework that integrates soft-label supervision, hierarchical hard-negative sampling, and gated binary interaction modeling. The model separately encodes attack descriptions and TTP definitions and employs a gated interaction module to adaptively fuse shared and divergent semantics, enabling fine-grained asymmetric alignment. A confidence-guided soft–hard collaborative supervision mechanism unifies weighted classification, semantic regression, and contrastive consistency into a multi-objective loss, dynamically rebalancing gradients to mitigate long-tail effects. Leveraging ATT & CK hierarchical priors, the model further performs in-tactic and cross-tactic hard-negative sampling to enhance semantic discrimination. Experiments on a real-world corpus demonstrate that DTGBI-TM achieves 98.53% F1 in semantic modeling and 79.77% Top-1 accuracy in open-set TTP prediction, while maintaining high inference efficiency and scalability in deployment. Full article

► Show Figures

Figure 1

18 pages, 1616 KB

Open AccessArticle

Efficient Failure Prediction: A Transfer Learning-Based Solution for Imbalanced Data Classification

by Abdullah Caliskan, Hasan Badem, Joseph Walsh and Daniel Riordan

Electronics 2025, 14(24), 4957; https://doi.org/10.3390/electronics14244957 - 17 Dec 2025

Viewed by 527

Abstract

Industrial predictive maintenance at the edge faces persistent challenges such as extreme class imbalance, limited labeled failure data, and the need for efficient yet scalable AI models. This paper proposes a transfer learning-based edge AI framework that addresses these challenges through a signal-to-image [...] Read more.

Industrial predictive maintenance at the edge faces persistent challenges such as extreme class imbalance, limited labeled failure data, and the need for efficient yet scalable AI models. This paper proposes a transfer learning-based edge AI framework that addresses these challenges through a signal-to-image transformation and fine-tuning of deep residual networks (ResNet). One-dimensional sensor signals are converted into two-dimensional RGB images, enabling the use of powerful convolutional architectures originally trained on large-scale datasets. The approach emulates an edge–cloud synergy, where knowledge distilled from large pre-trained models is efficiently adapted and executed on resource-constrained edge environments. Trained on less than 5% of the original dataset, the model achieves a negative predictive value of 96.53%, significantly reducing classification cost and outperforming both conventional deep learning and traditional machine learning methods. The results demonstrate that transfer learning-driven edge intelligence offers a cost-effective, scalable, and generalizable solution for predictive maintenance and industrial automation under data scarcity. Full article

(This article belongs to the Special Issue Exploring Edge AI: Architectures, Algorithms, and the Role of Edge–Cloud Cooperation for Scalable AI Systems)

► Show Figures

Figure 1

12 pages, 2116 KB

Open AccessArticle

A Design of High-Precision and Low-Noise High-Current Power Amplifier

by Meng Li, Zishu He, Yu Cao, Binghui He, Bin Liu and Jian Ren

Electronics 2025, 14(24), 4956; https://doi.org/10.3390/electronics14244956 - 17 Dec 2025

Viewed by 385

Abstract

Addressing the limitations of existing power amplifiers, particularly in terms of accuracy and noise performance, a high-voltage and high-current power amplifier has been developed. The input stage utilizes a rail-to-rail circuit structure, allowing the amplifier to deal with the full swing of input [...] Read more.

Addressing the limitations of existing power amplifiers, particularly in terms of accuracy and noise performance, a high-voltage and high-current power amplifier has been developed. The input stage utilizes a rail-to-rail circuit structure, allowing the amplifier to deal with the full swing of input signals from the negative to the positive power supply. The output stage features an innovative class AB configuration with a bias structure, effectively reducing the crossover distortion typically associated with traditional circuits. This design improves linearity, achieving an output range that extends to the rails, while also enhancing the power supply rejection ratio and optimizing noise performance. Furthermore, over-temperature protection and current limiting circuits have been integrated to safeguard the system against permanent damage under extreme conditions. The power amplifier circuit was simulated and validated using Cadence 61 Spectre software. With a power supply of ±30 V, the amplifier achieved an output current of 560 mA, a low-frequency gain of 138 dB, a bandwidth of 24 MHz, and a noise level of 4.8 nV/

\sqrt{H z}

. The slew rate was measured at 14.2 V/μs. Compared to existing literature, significant advancements have been achieved in terms of gain, bandwidth, and noise performance. Full article

(This article belongs to the Section Circuit and Signal Processing)

► Show Figures

Figure 1

27 pages, 3666 KB

Open AccessArticle

Curriculum to Immersion: A Conceptual Framework of Artificial Intelligence-Assisted Scenario Generation in Extended Reality for Primary and Secondary Education

by Tudor-Mihai Ursachi and Maria-Iuliana Dascalu

Electronics 2025, 14(24), 4955; https://doi.org/10.3390/electronics14244955 - 17 Dec 2025

Viewed by 473

Abstract

In this paper, we present a conceptual design framework for developing immersive learning experiences at scale with generative AI and extended reality (XR) for primary and secondary education. Based on the synthesis of current literature, our framework asserts a practical five-step pipeline: curriculum [...] Read more.

In this paper, we present a conceptual design framework for developing immersive learning experiences at scale with generative AI and extended reality (XR) for primary and secondary education. Based on the synthesis of current literature, our framework asserts a practical five-step pipeline: curriculum ingestion, AI-powered blueprinting, asset assembly, educator review, and classroom deployment with formative assessment. The model is designed to be flexible, focusing on narrative and gamification for primary students, moving on to sophisticated simulations and analytical activities for secondary students. We place this framework into the context of recent developments in generative 3D models, bridging fundamental technical and ethical gaps between concept and classroom practice. Finally, we summarize a prioritized research agenda around evaluation, access, and teacher workflows to enable near-term pilot studies. This work is intended to inform educators, researchers, and stakeholders who are interested in implementing effective AI-XR solutions in schools in a pedagogically sound way. Full article

(This article belongs to the Special Issue Artificial Intelligence for Extended Reality (AI4XR): Advancing Immersive Technologies)

► Show Figures

Graphical abstract

36 pages, 7057 KB

Open AccessArticle

Design and Application of a Nurse-Following Medical Bed Robot with a Negative Pressure Chamber for Patient Transportation in the Hospital: A Korean Case of Federated Digital Twins

by Jiyoung Woo, Hyojin Shin, Changhoon Jeon and Sangchan Park

Electronics 2025, 14(24), 4954; https://doi.org/10.3390/electronics14244954 - 17 Dec 2025

Viewed by 429

Abstract

Robots and artificial intelligence have revolutionized the healthcare sector. Transporting patients within hospitals is critical; however, reducing errors and inefficiencies caused by human intervention and increasing task efficiency are necessary. Therefore, there is a clear need to reduce these interventions and increase overall [...] Read more.

Robots and artificial intelligence have revolutionized the healthcare sector. Transporting patients within hospitals is critical; however, reducing errors and inefficiencies caused by human intervention and increasing task efficiency are necessary. Therefore, there is a clear need to reduce these interventions and increase overall task efficiency. We implemented a digital twin of the situation in which a nurse-following patient transport bed robot (in short, nurse-following bed robot or medical bed robot) transports patients in an infectious disease situation. To operate multiple bed robots, a federated digital twin was implemented, and all processes that occur in a hospital when an infectious disease patient arrives were defined, and scenarios for various situations were constructed. These scenarios were then simulated to validate system performance and preparedness for real-world situations. This study investigates and provides a detailed explanation of the core technologies required for this digital implementation process. Full article

(This article belongs to the Special Issue Digital Twins in Industry 4.0, 2nd Edition)

► Show Figures

Figure 1

14 pages, 638 KB

Open AccessArticle

A Low-Cost Head-Controlled and Sip-and-Puff Mouse: System Design and Preliminary Findings

by Rodrigo Duarte, Nuno Vieira Lopes and Paulo Jorge Coelho

Electronics 2025, 14(24), 4953; https://doi.org/10.3390/electronics14244953 - 17 Dec 2025

Viewed by 346

Abstract

This work introduces a low-cost, wearable assistive mouse designed to support digital interaction for individuals with motor impairments. The system combines inertial sensing for head-movement tracking and a pressure-based interface for simulating mouse clicks via “sip-and-puff” actions. The device enables full mouse control [...] Read more.

This work introduces a low-cost, wearable assistive mouse designed to support digital interaction for individuals with motor impairments. The system combines inertial sensing for head-movement tracking and a pressure-based interface for simulating mouse clicks via “sip-and-puff” actions. The device enables full mouse control (pointer movement, clicks, and double-clicks) without relying on hand mobility. Preliminary evaluations, conducted with input from occupational therapy professionals, demonstrated promising usability and functionality comparable to commercial devices. The proposed solution offers a cost-effective, open-source alternative to existing adaptive technologies, with future development aimed at broader testing and integration in rehabilitation settings. Future work will include usability testing with individuals presenting real motor impairments to validate clinical applicability. Full article

(This article belongs to the Special Issue Assistive Technology: Advances, Applications and Challenges)

► Show Figures

Graphical abstract

29 pages, 777 KB

Open AccessReview

Blockchain-Based Fraud Detection: A Comparative Systematic Literature Review of Federated Learning and Machine Learning Approaches

by Halima Farrukh, Sidra Zafar, Zia Ul Rehman, Asghar Ali Shah and Nizal Alshammry

Electronics 2025, 14(24), 4952; https://doi.org/10.3390/electronics14244952 - 17 Dec 2025

Viewed by 1480

Abstract

This systematic literature review uses the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology to assess progress in blockchain-based Federated learning (FL) and Machine Learning (ML) for detecting financial fraud over the last five years (2020–2025). An initial pool of 29,274 [...] Read more.

This systematic literature review uses the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology to assess progress in blockchain-based Federated learning (FL) and Machine Learning (ML) for detecting financial fraud over the last five years (2020–2025). An initial pool of 29,274 records identified across IEEE Xplore, ACM Digital Library, and ScienceDirect yielded 1585 peer-reviewed studies that met the inclusion criteria. Both qualitative and quantitative approaches were used. The examined papers were classified according to algorithm type, fraud types, and evaluation measures. Credit card fraud and cryptocurrency fraud dominated the literature, with supervised learning (e.g., XGBoost, 95% accuracy) and federated learning (e.g., FedAvg, 91% accuracy) emerging as dominant methodologies. Centralized ML outperforms FL in latency but poses privacy risks. FL–blockchain hybrids reduce false positives. While precision, recall, and F1-score are commonly used, few studies use cost-sensitive criteria. Future research should prioritize adaptive FL aggregation, privacy-preserving ML, and cross-industry collaboration. Full article

(This article belongs to the Special Issue Machine Learning: Applications for Cybersecurity)

► Show Figures

Figure 1

15 pages, 11922 KB

Open AccessArticle

Construction Method of Knowledge Graph of Chain Disaster in Alpine Gorge Area, China

by Haixing Shang, Lanling Jia, Jiahuan Xu, Jiangbo Xi and Chaofeng Ren

Electronics 2025, 14(24), 4951; https://doi.org/10.3390/electronics14244951 - 17 Dec 2025

Viewed by 365

Abstract

In high-mountain canyon areas, complex geological environments lead to frequent cascading disasters with unclear triggering mechanisms, posing severe threats to human life and property. Existing knowledge graph research in geology predominantly focuses on single-hazard types or general geological entities, lacking structured modeling and [...] Read more.

In high-mountain canyon areas, complex geological environments lead to frequent cascading disasters with unclear triggering mechanisms, posing severe threats to human life and property. Existing knowledge graph research in geology predominantly focuses on single-hazard types or general geological entities, lacking structured modeling and specialized datasets for cascading disaster processes, particularly the evolutionary chains in high-mountain canyon settings. To address this gap, this study proposes a method for constructing a knowledge graph tailored to cascading disasters in high-mountain canyon regions. First, a three-layer schema framework—comprising concept, relation, and instance layers—was designed to systematically characterize the knowledge elements and evolutionary relationships of disaster chains. To address the lack of a knowledge dataset for cascade disasters, this paper integrates multi-source heterogeneous data to construct a high-mountain canyon cascading disasters entity–relation dataset (DCER-MC), providing a reliable benchmark for related tasks. Based on this dataset, we implemented the knowledge graph and conducted disaster chain analysis. Experiments and applications demonstrate that the constructed knowledge graph effectively supports structured storage, centralized management, and scenario-based application of regional cascading disaster information. The main contributions of this work are (1) proposing a targeted schema framework for cascading-disaster knowledge graphs; (2) releasing a specialized dataset for cascading disasters in high-mountain canyon regions; and (3) establishing a complete pipeline from data to knowledge to scenario-based services, offering a novel knowledge-driven paradigm for disaster chain risk identification, inference prediction, and emergency decision-making in these areas. Full article

► Show Figures

Figure 1

Journal Menu

Journal Browser

Electronics, Volume 14, Issue 24 (December-2 2025) – 200 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI