Was ist die Webseite?

https://www.mdpi.com/journal/electronics

Electronics MDPI, Basel (2026)

31/08/2026

🔥 in !
🧠 Electronic Artificial Intelligence–Digital Twin Model for Optimizing Electroencephalogram Signal Detection
🔗 Read at: https://www.mdpi.com/2079-9292/14/6/1122
Author: Alessandro Massaro from LUM Libera Università Mediterranea, Italy.

This study presents an electronic proof-of-concept Digital Twin (DT) model designed to support Electroencephalogram (EEG) signal detection and interpretation. The proposed model combines circuit modelling and simulation of EEG electrodes with an Artificial Intelligence (AI)-supervised algorithm to classify and adjust noisy EEG signals.
The study specifically investigates the use of Random Forest (RF) and Artificial Neural Network (ANN) algorithms for processing EEG signals affected by Flicker and white noise. Using a dataset of EEG signals associated with alcohol exposure, the researchers demonstrate how the RF-based approach can reduce noise ripple behaviour and facilitate the interpretation of time-domain peaks and waveform morphology.
One of the key contributions of the work is the integration of the circuit simulation and AI processing within a Digital Twin framework. The proposed DT can incorporate configurable physical and physiological parameters, support real-time checking of detected EEG signals, and potentially be adapted to different EEG-related pathologies by constructing specific training datasets.
The results indicate that the Random Forest approach is a promising alternative to conventional EEG filtering methods. For the investigated dataset, RF achieved a slightly lower probabilistic error than ANN and demonstrated a better cleaning action, while providing performance comparable to or slightly better than traditional approaches such as bandpass filtering, FIR filtering, and wavelet-based methods.
The study highlights the potential of combining Artificial Intelligence, Digital Twins, and biomedical electronics to improve the acquisition and interpretation of physiological signals, with possible future applications in wearable devices and other biomedical signal-processing systems.

26/08/2026

🚦 !
🤖 Meta-YOLOv8: Meta-Learning-Enhanced for Precise Color Detection in ADAS
🔗 Read at: https://www.mdpi.com/2079-9292/14/3/468
Authors: Vasu Tammisetti, Georg Stettinger, Manuel Pegalajar Cuéllar and Miguel Molina-Solana from Universidad de Granada

Accurate traffic light color detection is essential for the safety and reliability of Advanced Driver Assistance Systems (ADAS). This work introduces Meta-YOLOv8, a meta-learning-enhanced version of YOLOv8 specifically designed to improve traffic light color recognition.
The proposed approach is evaluated against established object-detection models, including SSD, Faster R-CNN, and Detection Transformers (DETR). The results demonstrate strong performance, achieving 93% F1 score and accuracy, along with 97% precision, highlighting the potential of meta-learning to enhance computer vision systems for intelligent vehicles.
By combining the capabilities of YOLOv8 with meta-learning, this research contributes to more accurate and adaptable traffic light detection—an important step toward safer and more reliable AI-powered ADAS and autonomous driving technologies.

25/08/2026

🎓 !
🤖 Explainable Approaches in Primary Education: A Review
🔗 Read at: https://www.mdpi.com/2079-9292/14/11/2279
Authors: Jim Prentzas and Ariadni Binopoulou from the Democritus University of Thrace

This review explores the growing use of Explainable Artificial Intelligence (XAI) in primary education, examining how transparent and interpretable AI approaches can support teaching, learning, decision-making, and educational administration. The authors conducted an extensive search using Google Scholar and Scopus and identified 23 relevant studies for review.
The paper proposes a categorization of XAI approaches in primary education into three main areas: support for teaching and learning, AI as a learning subject, and policymaking, decision support, and administrative tasks. The review also analyzes the main XAI tools and methods used, the educational subjects in which they are applied, and current research trends.
The findings highlight the potential of XAI to make AI-driven educational systems more transparent, understandable, accountable, and trustworthy, while also identifying areas for further research and development.

24/08/2026

🚀
⚡ Real-Time vs. -Accelerated Pipelines for Low-Cost Microscopy Applications
🔗 Read at: https://www.mdpi.com/2079-9292/14/5/930
Authors: Gloria Bueno, Lucia Sanchez-Vargas, Alberto Diaz-Maroto, Jesus Ruiz-Santaquiteria, Maria Blanco, Jesus Salido and Gabriel Cristobal from Universidad de Castilla-La Mancha and Instituto de Óptica "Daza de Valdés" (IO-CSIC) Váldes

This paper explores the potential of hashtag and hashtag for low-cost microscopy applications, comparing real-time edge-based processing with GPU-accelerated computational pipelines.
The study highlights how modern computational approaches can make advanced microscopy more accessible by enabling efficient , , and real-time analysis on affordable hardware. By examining the trade-offs between processing performance, computational resources, and cost, the work provides valuable insights into designing practical microscopy systems for applications where affordability and real-time operation are essential.
The findings demonstrate the importance of selecting an appropriate computing architecture based on application requirements, balancing processing speed, computational efficiency, cost, and real-time performance.
A valuable contribution to the intersection of , , , , , and .

21/08/2026

🚀
📍 A Survey on the Main Techniques Adopted in Indoor and Outdoor Localization
🔗 Read at: https://www.mdpi.com/2079-9292/14/10/2069
Authors: Massimo Stefanoni, Imre Kovács, Peter Sarcevic, and Ákos Odry from Obuda University and University of Szeged

This survey provides a comprehensive overview of the main techniques used for indoor and outdoor localization, addressing one of the key challenges in modern positioning and navigation systems: accurately determining the location of people, devices, robots, and other objects across diverse environments.
The paper reviews a broad range of localization technologies and approaches, highlighting how different sensing and positioning techniques can be selected depending on environmental conditions, required accuracy, system complexity, and application scenarios.
Particular attention is given to the evolution of localization methods across indoor and outdoor environments, where factors such as signal propagation, obstacles, multipath effects, infrastructure availability, and dynamic conditions can significantly influence positioning performance.
By bringing together established and emerging localization techniques, the survey offers a useful perspective for researchers and practitioners working on robotics, autonomous systems, wireless technologies, navigation, IoT, and smart environments.
The review also highlights the importance of combining complementary technologies to achieve more robust and accurate localization, especially in complex real-world scenarios where a single positioning method may not be sufficient.

20/08/2026

🚀 !
⚡ Remote Vibration Monitoring of Combustion Engines Utilising
🔗 Read at: https://www.mdpi.com/2079-9292/14/11/2118
Authors: Rafał Kociszewski and Wojciech Wojtkowski from Bialystok University of Technology
This paper presents a remote vibration monitoring system for combustion engines that combines vibration sensing, edge computing, and wireless communication to support real-time condition monitoring and early fault detection.
The proposed system uses an MPU-6050 accelerometer together with AVR and ESP32 microcontrollers to acquire and process vibration signals close to the data source. By calculating statistical parameters directly at the edge—including RMS, variance, standard deviation, and impulse factor—the system can reduce data transmission requirements while maintaining useful diagnostic information.
Experimental tests demonstrated that changes in selected vibration measures can be associated with deteriorating engine conditions, highlighting their potential as early indicators of mechanical faults. The system also enables processed data to be transmitted to a central server via mobile communication and supports OTA firmware updates, providing a flexible platform for future diagnostic developments.
The study highlights the potential of for scalable and distributed vibration diagnostics, particularly in mobile and resource-constrained environments where low latency, reduced bandwidth requirements, and reliable operation are essential.

19/08/2026

🔥 !
⚡Developing a Novel Muscle Fatigue Index for Wireless sEMG Sensors: Metrics and Regression Models for Real-Time Monitoring
🔗Read at: https://shorturl.at/Gqasn
Authors: Dimitrios Miaoulis, Ioannis Stivaros and Stavros Koubias

This article presents a novel approach for real-time muscle fatigue monitoring using wireless hashtag sensors. The study investigates time-domain, frequency-domain, and hybrid-domain metrics—including RMS, ARV, MNF, and the MNF/ARV ratio—to identify reliable indicators of muscle fatigue.
The proposed methodology combines dynamic standardization with regression models to improve the robustness and comparability of fatigue measurements across users. The results demonstrate the potential of wireless sEMG systems and data-driven models to provide objective and real-time assessment of muscle fatigue, with promising applications in hashtag , sports, rehabilitation, and human activity monitoring.
This work contributes to the development of practical and non-invasive solutions for continuous muscle-state monitoring. Future research can further investigate the approach across larger and more diverse populations, different muscle groups, and real-world conditions to strengthen its applicability in wearable and healthcare technologies.

18/08/2026

🔥 !
🚗🔐 of Automotive Wired Networking Systems: Evolution, Challenges, and Countermeasures
🔗 Read at: https://www.mdpi.com/2079-9292/14/3/471
Authors: Nicasio Canino, Pierpaolo Dini, Stefano Mazzetti, Daniele Rossi, Sergio Saponara and Ettore Soldaini from the Università di Pisa and Embedded Software Systems (ESWS) S.r.l.
As vehicles evolve from traditional mechanical systems into highly connected, software-defined machines, cybersecurity has become a critical part of automotive development. This review provides a comprehensive overview of the cybersecurity challenges affecting modern automotive wired networking systems, with a particular focus on the vulnerabilities of the Controller Area Network (CAN) protocol.
The paper examines the evolution of automotive Electrical and Electronic (E/E) architectures and the expanding attack surface created by increasingly interconnected vehicles. It discusses key automotive cybersecurity standards, including ISO 26262 and ISO/SAE 21434, and introduces a dual taxonomy for classifying automotive attack surfaces according to the proximity of the attacker.
The authors also review known CAN-related cyberattacks and cybersecurity vulnerabilities, as well as publicly available CAN datasets used to develop and evaluate Intrusion Detection Systems (IDSs). Particular attention is given to anomaly-based, rule-based and hybrid IDS approaches, including the growing role of artificial intelligence and machine learning in real-time attack detection.
The findings highlight the need for comprehensive cybersecurity strategies capable of protecting increasingly connected and automated vehicles while supporting the transition toward software-defined automotive architectures.

17/08/2026

🚀 !
🤖 in Practice: A Survey and Deployment Framework for Neural Networks on
🔗 Read at: https://www.mdpi.com/2079-9292/14/24/4877
Authors: Ruth Córdova Cárdenas, Daniel Amor and Álvaro Gutiérrez from Universidad Politécnica de Madrid and RBZ Robot Design

Edge Artificial Intelligence (Edge AI) is transforming how intelligent systems process data by bringing neural network inference closer to where data is generated. This paper provides a comprehensive survey of neural networks on embedded systems, focusing on the practical challenges and opportunities involved in deploying AI models at the edge.
The study examines key aspects of Edge AI deployment, including model optimization, hardware constraints, computational efficiency, energy consumption, latency, and real-time inference. It also proposes a structured deployment framework to support the transition from AI models developed in the cloud or on high-performance platforms to resource-constrained embedded devices.
By bridging the gap between AI algorithms and embedded hardware, Edge AI enables intelligent applications that are faster, more energy-efficient, privacy-aware, and capable of operating with limited connectivity.
📌 An insightful contribution for researchers and practitioners working at the intersection of Artificial Intelligence, embedded systems, neural networks, and edge computing.

14/08/2026

🚀 !
🤖 A Comparative Evaluation of Transformer-Based Language Models for Topic-Based Sentiment Analysis
🔗 Read at: https://www.mdpi.com/2079-9292/14/4/758
Authors: Spyridon Tzimiris, Stefanos Nikiforos, Maria Nefeli Nikiforos, Despoina Mouratidis & Katia Lida Kermanidis

This study presents a comparative evaluation of transformer-based language models for topic-based sentiment analysis in the Greek educational context. The research compares four models—GreekBERT, XLM-R-Greek, mBERT, and Palobert—using three original sentiment-annotated datasets covering perspectives from parents, teachers, and school directors.
The findings demonstrate the strong potential of language-specific transformer models for capturing sentiment in domain-specific Greek-language data. GreekBERT achieved the best overall performance, reaching an F1 score of 0.91, with particularly strong performance in identifying negative sentiment (F1 = 0.95).
The study also highlights differences in classification difficulty across educational topics, providing valuable insights into the application of NLP and transformer architectures for analyzing complex educational and social data.

Electronics MDPI

31/08/2026

26/08/2026

25/08/2026

24/08/2026

21/08/2026

20/08/2026

19/08/2026

18/08/2026

17/08/2026

14/08/2026

Adresse

Webseite

Benachrichtigungen

Service Kontaktieren

Verknüpfungen

Teilen

Kategorie

Antrag auf Entfernung des Eintrags