This work aims to improve the performance of electric vehicles (EVs) based on linear induction mo tors (LIM). The Particle Swarm Optimization (PSO) method is proposed to tune the PID regulator of the Field-Oriented Control (FOC) technique. The main objective of this study is to develop innovative solutions that maximize the efficiency and precision of electric vehicles on various paths. The LIM model is imple mented using the d-q synchronous reference frame and takes into account the end-effect phenomenon. This phenomenon occurs due to the termination of the mo tor’s physical structure, which leads to distortion in the magnetic field at the ends of the motor’s primary (sta tor). It is also highly nonlinear, which increases its complexity and makes control difficult. To overcome this issue, the Field-Oriented Control (FOC) technique is suggested to achieve better efficiency, dynamic per formance, and greater control flexibility of the motor. Furthermore, the use of the (PSO) optimization tech nique enables the determination of optimal control pa rameters to maximize the performance of the (FOC LIM) system under different operating conditions, such as speed variation and disturbance load. A compari son between the PSO-PID and conventional methods in terms of response stability, steady-state error, and rise time is conducted using MATLAB/Simulink. The results demonstrate a more efficient, precise, and high performing electric vehicle system.

This work aims to improve the performance of electric vehicles (EVs) based on linear induction mo tors (LIM). The Particle Swarm Optimization (PSO) method is proposed to tune the PID regulator of the Field-Oriented Control (FOC) technique. The main objective of this study is to develop innovative solutions that maximize the efficiency and precision of electric vehicles on various paths. The LIM model is imple mented using the d-q synchronous reference frame and takes into account the end-effect phenomenon. This phenomenon occurs due to the termination of the mo tor’s physical structure, which leads to distortion in the magnetic field at the ends of the motor’s primary (sta tor). It is also highly nonlinear, which increases its complexity and makes control difficult. To overcome this issue, the Field-Oriented Control (FOC) technique is suggested to achieve better efficiency, dynamic per formance, and greater control flexibility of the motor. Furthermore, the use of the (PSO) optimization tech nique enables the determination of optimal control pa rameters to maximize the performance of the (FOC LIM) system under different operating conditions, such as speed variation and disturbance load. A compari son between the PSO-PID and conventional methods in terms of response stability, steady-state error, and rise time is conducted using MATLAB/Simulink. The results demonstrate a more efficient, precise, and high performing electric vehicle system.

This work aims to improve the performance of electric vehicles (EVs) based on linear induction mo tors (LIM). The Particle Swarm Optimization (PSO) method is proposed to tune the PID regulator of the Field-Oriented Control (FOC) technique. The main objective of this study is to develop innovative solutions that maximize the efficiency and precision of electric vehicles on various paths. The LIM model is imple mented using the d-q synchronous reference frame and takes into account the end-effect phenomenon. This phenomenon occurs due to the termination of the mo tor’s physical structure, which leads to distortion in the magnetic field at the ends of the motor’s primary (sta tor). It is also highly nonlinear, which increases its complexity and makes control difficult. To overcome this issue, the Field-Oriented Control (FOC) technique is suggested to achieve better efficiency, dynamic per formance, and greater control flexibility of the motor. Furthermore, the use of the (PSO) optimization tech nique enables the determination of optimal control pa rameters to maximize the performance of the (FOC LIM) system under different operating conditions, such as speed variation and disturbance load. A compari son between the PSO-PID and conventional methods in terms of response stability, steady-state error, and rise time is conducted using MATLAB/Simulink. The results demonstrate a more efficient, precise, and high performing electric vehicle system.

Résumé. Le phénomène d’accréditation des facultés de médecine a vu une progression remarquable. En quête de qualité, la faculté de médecine engage un processus d’accréditation auprès d’un organisme indépendant afin de s'assurer qu'elle dispense un enseignement de haute qualité et qu’elle forme des médecins compétents et répondent aux exigences professionnelles universelles. Elle doit répondre aux critères établis par l’organisme accréditeur. Certaines facultés se trouvent confrontées à des incertitudes qu'elles sont censées résoudre, ce qui soulève des questions quant à leur capacité à y parvenir. La littérature scientifique a rapporté des solutions inscrites dans le cadre d'isomorphisme institutionnel. Cette forme d'adaptation institutionnelle ne se limite pas à une simple fonction d'accréditation, mais elle constitue en soi un indicateur de qualité. L’adoption d’une stratégie proactive et l’appropriation de comportements mimétiques, peut se traduire par une homogénéisation du champ organisationnel. Les facultés de médecine Algériennes sont engagées dans le processus d’accréditation depuis octobre 2024. Il est mis en avant les stratégies que la faculté de médecine adopte afin de réduire ses incertitudes ainsi que les opportunités que celle-ci peut saisir afin de revoir son organisation tout en adoptant une stratégie décanale proactive structurée (SDPS). C’est dans le cadre d’isomorphisme institutionnel mimétique que la faculté peut transcrire sa démarche en quête d’assurance qualité.
Mots clés. Accréditation ; assurance qualité ; stratégie décanale structurée ; missions ; isomorphisme institutionnel

This study aims to highlight the importance of evaluating explosive atmosphere (ATEX) risks and to
propose a novel, more accurate and comprehensive risk assessment method compared to traditional existing
approaches. To address the limitations of traditional ATEX risk assessment methods, this study introduces a new
approach based on three key parameters: occurrence, detection, and severity. These factors are combined using
a weighted sum (WS) formula to calculate a modified risk priority number (RPN), enhanced by the analytic
hierarchy process.
This hybrid methodology provides a more precise and reliable assessment of ATEX risks. Moreover, the
findings of the proposed WS-RPN method (and its ATEX-specific version, WS-RATEX) demonstrate superior
accuracy and lower duplication rates compared to conventional risk assessment techniques. A sensitivity
analysis confirms the robustness, stability, and reliability of this approach, making it a valuable tool for
quantifying ATEX risks.
It is evident to say that WS-RPN represents a significant advancement in ATEX risk assessments. By
integrating a refined calculation model, and factoring in occurrence, detection, and severity, the latter ensures
a more systematic and reliable evaluation. Thus, with improved risk prioritization, this approach not only enhances
industrial installations’ safety but also serves as a more effective alternative to traditional assessment methods.

Introduction: A psychometric analysis of multiple-choice questions (MCQs) is an important step in the assessment process. It makes sure that MCQ items are valid and reliable. The residency competition is an important part of medical training. It is how we choose future doctors who can meet the clinical and scientific requirements. Our aim is to find out what the medical residency competition's MCQs are like.

 

Methods: 150 multiple-choice questions (MCQs) were analysed from the October 2025 medical residency exam. There were 410 candidates in the exams. This analysis looked at the difficulty index (DIF I), the discrimination index (DI), and Cronbach's alpha coefficient.

 

Results: 78 (52%) of the multiple-choice questions were at a recommended/acceptable level of difficulty, with an average DIF I of 43± 2,03. 45 (30%) were too difficult. 97 items (64%) had a DI between excellent and acceptable, and 40 (26%) were rejected, with an overall average DI of 0,325± 0,017. There was a relationship between DIF I and DI (r=0.36, p<0.01). The reliability of the items was satisfactory at 0.84, as measured by Cronbach's alpha coefficient.

 

Conclusion: Our study allowed us to find good MCQs to add to the question bank for future tests, as well as MCQs that needed to be changed. This thorough analysis makes sure that the assessments are valid and reliable. It makes sure that the selection process is tough and that it meets the necessary academic and practical standards.

 

Keywords: MCQ, residency exam, psychometrics, difficulty index, discrimination index

Objectives: Traffic accidents cause severe social and economic impacts, demanding fast and reliable detection to minimize secondary collisions and improve emergency response. However, existing cloud-dependent detection systems often suffer from high latency and limited scalability, motivating the need for an edge-centric and consensus-free accident detection framework in IoV environments. 

Methods: This study presents a real-time accident detection framework tailored for Internet of Vehicles (IoV) environments. The proposed system forms an integrated IoV architecture combining on-vehicle inference, RSU-based validation, and asynchronous cloud reporting. The system integrates a lightweight ensemble of Vision Transformer (ViT) and EfficientNet models deployed on vehicle nodes to classify video frames. Accident alerts are generated only when both models agree (vehicle-level ensemble consensus), ensuring high precision. These alerts are transmitted to nearby Road Side Units (RSUs), which validate the events and broadcast safety messages without requiring inter-vehicle or inter-RSU consensus. Structured reports are also forwarded asynchronously to the cloud for long-term model retraining and risk analysis. 

Results: Evaluated on the CarCrash and CADP datasets, the framework achieves an F1-score of 0.96 with average decision latency below 60 ms, corresponding to an overall accuracy of 98.65% and demonstrating measurable improvement over single-model baselines. 

Conclusions: By combining on-vehicle inference, edge-based validation, and optional cloud integration, the proposed architecture offers both immediate responsiveness and adaptability, contrasting with traditional cloud-dependent approaches.

first_page

settings

Order Article Reprints

Open AccessArticle

A Real-Time Consensus-Free Accident Detection Framework for Internet of Vehicles Using Vision Transformer and EfficientNet

by 

Zineb Seghir

 ¹,

Lyamine Guezouli

 ^2,*,

Kamel Barka

 ¹,

Djallel Eddine Boubiche

 ^2,*,

Homero Toral-Cruz

 ^3,* and

Rafael Martínez-Peláez

 ^4,5

¹

LaSTIC Laboratory, Computer Science Department, University of Batna 2, Batna 05000, Algeria

²

Laboratory of Renewable Energy, Energy Efficiency and Smart Systems (LEREESI), Higher National School of Renewable Energies, Environment and Sustainable Development (HNS-RE2SD), Batna 05000, Algeria

³

Departamento de Ingeniería y Tecnología, Universidad Autónoma del Estado de Quintana Roo, Chetumal 77019, Mexico

⁴

Unidad Académica de Computación, Universidad Politécnica de Sinaloa, Mazatlán 82199, Mexico

⁵

Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte, Antofagasta 1270709, Chile

^*

Authors to whom correspondence should be addressed.

AI 2026, 7(1), 4; https://doi.org/10.3390/ai7010004

Submission received: 12 November 2025 / Revised: 14 December 2025 / Accepted: 16 December 2025 / Published: 22 December 2025

Downloadkeyboard_arrow_down 

Browse Figures

 Versions Notes

 

Abstract

Objectives: Traffic accidents cause severe social and economic impacts, demanding fast and reliable detection to minimize secondary collisions and improve emergency response. However, existing cloud-dependent detection systems often suffer from high latency and limited scalability, motivating the need for an edge-centric and consensus-free accident detection framework in IoV environments. Methods: This study presents a real-time accident detection framework tailored for Internet of Vehicles (IoV) environments. The proposed system forms an integrated IoV architecture combining on-vehicle inference, RSU-based validation, and asynchronous cloud reporting. The system integrates a lightweight ensemble of Vision Transformer (ViT) and EfficientNet models deployed on vehicle nodes to classify video frames. Accident alerts are generated only when both models agree (vehicle-level ensemble consensus), ensuring high precision. These alerts are transmitted to nearby Road Side Units (RSUs), which validate the events and broadcast safety messages without requiring inter-vehicle or inter-RSU consensus. Structured reports are also forwarded asynchronously to the cloud for long-term model retraining and risk analysis. Results: Evaluated on the CarCrash and CADP datasets, the framework achieves an F1-score of 0.96 with average decision latency below 60 ms, corresponding to an overall accuracy of 98.65% and demonstrating measurable improvement over single-model baselines. Conclusions: By combining on-vehicle inference, edge-based validation, and optional cloud integration, the proposed architecture offers both immediate responsiveness and adaptability, contrasting with traditional cloud-dependent approaches.

Oil well drilling is an expensive process that needs a particular focus. For this reason, Rate Of Penetration (ROP) has been widely approved as a measure of drilling efficiency and adequate configuration parameters. Our aim in this work consists in the elaboration of a smart system using Hoeffding trees for predicting the Rate of Penetration (ROP) in oilfield drilling. The choice of Hoeffding trees to build our model is motivated by their adaptive learning capability and drift detection. They offer continuous, fast, and efficient learning both online on data streams and offline on batch data. To validate our approach, we used real drilling data from the “Hassi-Terfa” oilfield located in Southeast Algeria. The obtained results show in comparison to the eXtreme Gradient Boosting (XGBoost) algorithm that Hoeffding trees maintain their learning capacity and produce more accurate predictions even in the presence of drifts. This is thanks to the combination of the Adaptive Windowing (ADWIN) algorithm to manage drifts and least mean squares (LMS) filters to reduce noise. This observation highlights the effectiveness of our approach to predict the ROP while oil-well drilling. The proposed smart system offers more efficient solution to predict the ROP, whether in real-time or offline. By leveraging its adaptability to changes in data distribution, our approach ensures more accurate and adaptive predictions, facilitating drilling operations optimization and boosting the overall efficiency of the process.

This research conducts a comprehensive numerical evaluation into an advanced heat dissipation system for low-concentrated photovoltaic systems, addressing the limitations of conventional minichannel heat sink designs. To overcome their inherent inefficiencies, a novel minichannel configuration with wavy surfaces and a trapezoidal inlet section (TWMC) is proposed, aiming to enhance convective heat transfer through increased surface area and induced flow turbulence. Three configurations wavy minichannel (TWMC), trapezoidal minichannel (TMC), and rectangular minichannel (RMC) are systematically compared in terms of key performance metrics, including thermal resistance, Nusselt number, pressure loss, and friction index. Water serves as the coolant, operating in a laminar flow regime (Re = 200–900) and absorbing a uniform heat flux of 100 kW/m² applied to the channel base. Results demonstrate that the TWMC configuration outperforms conventional designs, achieving a 30.82 % decline in heat resistance and a 9.2 % surge in Nusselt number at peak Reynolds numbers. The TWMC design improves the performance evaluation criterion (PEC) to 1.06, with exceptional overall thermohydraulic performance PEC(R) ranging from 1.078 to 1.271, despite higher pressure drop. These findings offer insights into optimizing CPV system performance, emphasizing the potential of innovative wavy-channel geometries to revolutionize thermal management and energy efficiency in advanced photovoltaic applications.

This article presents a quantitative and qualitative analysis of causal markers in scientific articles published in France and Algeria. Based on two corpora drawn from Synergie France and Synergie Algérie, the study examines the frequency, distribution, and functions of connectors such as car, donc, puisque, and parce que. The results reveal a common core of markers but distinct preferences:  French authors favor a structured and diversified argumentative style, while Algerian writers adopt a more explicit and pedagogical approach. These differences reflect contrasting academic traditions and highlight the didactic importance of causal markers in the teaching of scientific writing.

Wild chicory, or Hyoseris radiata L., is indigenous to the Mediterranean region, is a plant used in traditional medicine as a diuretic, blood depurative, and against kidney stones. The present study aimed to assess for the first time the acute and subacute toxicity, to quantify the total amount of polyphenols and flavonoids, and to assess the antioxidant activity of H. radiata collected from Setif, Algeria. The overall amount of flavonoids and polyphenols was quantified spectrophotometrically. The antioxidant activity of the extract was evaluated according to two methods, DPPH and FRAP. The acute toxicity of H. radiata was carried out according to the OECD guideline 423 to determine the median lethal dose LD50 and the subacute toxicity was evaluated according to OECD guideline 407 to assess the possible pathological effects of the extract administered for 28 days by oral route. The results show that the total amount of polyphenols and flavonoids was 132.53 ± 2 µg of GAE·1 mg-1 and 96.11 ± 3.65 µg of QE·1 mg-1 of extract, respectively. The extract shows a good antioxidant potential in both tests. The administered dose (2 g·kg-1 of BW) didn’t produce any changes in general behaviors or mortality, so the LD50 is greater than 2 g·kg-1 of BW. Moreover, the daily administration of the extract with 2 doses, 100 mg·kg-1 and 200 mg·kg-1 didn’t cause any changes in body weight, behavior test, hematological parameters, and organ relative weight. A significant decrease in triglyceride was recorded in both concentrations. Based on the present findings, the extract of H. radiata has no significant toxicity. These findings offer valuable information about the toxicity profile of the traditional medicine plant Hyoseris radiata L.

The undifferentiated nasopharyngeal cancer (NPC) is a multifactorial disease mainly due to Epstein-Barr Virus (EBV) infection. The transmembrane tyrosine kinase 'EphA2' and the protease 'Furin' are implicated in the EBV entry into epithelial cells and other physiological processes. To gain insights into the association of single-nucleotide polymorphisms (SNPs) rs4702 and rs6603883 (FURIN and EPHA2 genes, respectively) with the risk and prognosis of the NPC, the genotypes of 471 individuals (228 cases and 243 controls) were assessed alongside risk cofactors (sex, tobacco, alcohol, occupation, and recurrent Ear, Nose and Throat infections) and prognosis cofactors (Tumor stage, local invasion, lymph node involvement, and metastasis) using multivariable logistic regression. We found that only the rs4702 AG/GG genotypes were statistically significantly associated with a reduced risk of cancer, both in the overall population and in men (approximately 50% reduction). The rs4702 GG genotype was also associated with a low frequency of local tumor invasion in the whole population (OR = 0.382, p = 0.017, co-dominant model, and OR = 0.409, p = 0.02, recessive model), but heterozygous women were associated with a higher lymph node involvement (OR = 3.53, p = 0.031, co-dominant model, and OR = 3.62, p = 0.02, overdominant model). The rs6603883 GG genotype was associated, in the dominant model, with distant metastasis in the whole population (OR = 2.5, p = 0.024), with advanced clinical stage in men (OR = 2.22, p = 0.034), and with advanced clinical stage and distant metastasis in patients under 49 years (OR = 3.13, p = 0.009, and OR = 5.15, p = 0.011, respectively). Additionally, men having the rs6603883 GA genotype were associated with lymph node invasion (OR = 2.22, p = 0.027, overdominant model). Our study is the first to demonstrate that FURIN and EPHA2 germline gene polymorphisms are associated with NPC risk (for rs4702) and prognosis (for both rs4702 and rs6603883), with sex-specific differences. These results need to be replicated and further investigated in other populations.

Wireless sensor networks (WSNs) have recently emerged as a critical technology in various applications, including industrial automation, building monitoring, and military. However, the data generated by these networks are often prone to outliers, which can compromise sensor data quality and reliability. Detecting outliers is paramount to ensure proper network functioning. Traditional detection techniques pose several challenges, such as weak adaptability to the increasing complexity and dynamic environmental changes, limited accuracy, and higher computation costs. To address these challenges, this paper proposes an optimized one-dimensional convolutional neural networks (1D CNN)-based outlier detection approach for WSNs. This approach comprises two key modules: a predictor module and an outlier detector. The predictor module employs a 1D CNN model to forecast forthcoming sensor measurements based on historical data. Bayesian optimization is used to enhance the 1D CNN model’s accuracy. The outlier detector identifies outliers based on the Euclidean distance between the predicted measurements and their corresponding actual values. Experiments on synthetic and real-world datasets reveal that our proposed approach outperforms other existing deep learning-based frameworks in terms of accuracy, F1 score, and false alarm rates.