Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

31,377 Article Results

Hardware-aware comparative study of lightweight convolutional neural networks for Raspberry Pi-based autonomous driving

10.11591/ijece.v16i3.pp1493-1507
Hyung In Kim , Youngmin Park
Deploying deep learning models for autonomous driving on resource-constrained edge devices, such as the Raspberry Pi, presents significant challenges due to strict limitations on inference latency and memory capacity. To address these constraints, this study conducts a comprehensive comparative evaluation of lightweight convolutional neural networks (CNNs) optimized for dual-output regression of steering angle and driving speed. We benchmark a task-specific end-to-end baseline (NVIDIA CNN) against representative classification-oriented architectures—including MobileNet, ShuffleNet, EfficientNet, GhostNet, and SqueezeNet—all reformulated for this regression task. Experiments were conducted on a physical Raspberry Pi-based autonomous RC car platform to assess prediction accuracy, inference speed, and real-world closed-loop driving stability using quantitative metrics such as the normalized jerk ratio. Experimental results demonstrate a clear trade-off: while GhostNetV1 0.5x achieved the highest regression accuracy with a Total R2 score of 95.8% and MobileNetV1 recorded a competitive MAE of 1.95, they failed to provide stable control due to severe high-frequency steering jitter. Conversely, the NVIDIA CNN proved to be the most practical solution for general edge deployment, achieving the lowest inference latency of 61.1 ms (16.4 FPS) and a minimal memory footprint of 2.78 MB, ensuring stable autonomous navigation (1.50xjerk ratio). Furthermore, ShuffleNetV2 0.5x emerged as the superior architecture for trajectory precision, recording the lowest weighted MAE of 1.60. These findings underscore that theoretical accuracy does not guarantee real-world drivability on embedded systems, providing practical guidelines for hardware-aware model selection in edge-based autonomous driving.
Volume: 16
Issue: 3
Page: 1493-1507
Publish at: 2026-06-01

Comparative analysis of PM6:L8-BO organic and inverted organic solar cell

10.11591/ijape.v15.i2.pp770-780
Karthika Krishnakumar , Ashish Grover , Pardeep Kumar
Advancements in solar technologies are driven by the pursuit of higher efficiency and reduced environmental impact. This study presents a comprehensive and comparative analysis of organic and inverted organic solar cells (OSC and IOSC), using the OghmaNano software for simulations and analysis. This work is specifically designed to compare conventional and inverted structures and understand how device engineering impacts performance metrics. When OSCs are characterized by a low work-function cathode on top, IOSCs feature a clear conductive oxide cathode at the bottom. The study focusses on extracting key electrical output, including short circuit current density (JSC), open-circuit voltage (VOC), fill factor (FF) and power conversion efficiency (PCE), through the calculated current-voltage characteristic (J-V). Various physical characteristics, such as thickness of different layers and materials deployed as electron transport layer (ETL) and hole transport layer (HTL), are systematically investigated. Diverse top and bottom electrodes, encompassing monothin and multithin layer configurations, are proposed. The study shows that IOSC achieves higher efficiency than OSC, reaching 21.60%, while using a multithin layer ZTZ (ZnO/TiOx/ZnO) as the bottom contact, demonstrating improved charge transport and overall efficiency.
Volume: 15
Issue: 2
Page: 770-780
Publish at: 2026-06-01

Comparison of differential evolution optimization technique with other techniques in solving multi-objective optimal power flow

10.11591/ijape.v15.i2.pp663-673
Vineeta S. Chauhan , Jaydeep Chakravorty , Siddharthsingh K. Chauhan
Optimal power flow (OPF) is a complex, non-linear optimization problem focused on determining the steady-state operating parameters of power systems for economic and secure operation. The challenge intensifies due to numerous system constraints that must be satisfied simultaneously. Although various evolutionary algorithms (EAs) have been applied to OPF in recent decades, these algorithms often use unconstrained search strategies. A common approach to handle constraint violations is the static penalty function, which penalizes infeasible solutions. However, selecting suitable penalty coefficients typically involves time-consuming trial and error, affecting overall performance. This study explores the integration of advanced constraint handling (CH) techniques within the differential evolution (DE) framework to enhance the performance of optimal power flow (OPF) solutions. In particular, it looks at three approaches: a hybrid ensemble of two CH techniques (ECHT), a self-adaptive penalty method (SP), and superiority of viable solutions (SF). The IEEE 30-bus and IEEE-57 bus benchmark systems are used to evaluate the efficacy of these techniques under a variety of OPF goals, including lowering emissions and generation costs, cutting power losses, and enhancing voltage stability. We took into consideration both weighted-sum multi-objective and single-objective formulations. The simulation outcomes indicate that the proposed CH-DE approaches deliver robust and competitive optimization results, demonstrating improved constraint handling capabilities when compared to contemporary methods in the literature.
Volume: 15
Issue: 2
Page: 663-673
Publish at: 2026-06-01

Enhancing sEMG finger gesture recognition using optimized 1D-convolutional neural network

10.11591/ijece.v16i3.pp1576-1587
Daniel Sutopo Pamungkas , Sumantri K. Risandriya
Robust and precise finger gesture recognition using surface electromyography (sEMG) is essential for developing intuitive prosthetic control systems. However, sEMG signals are inherently stochastic and non-stationary, posing significant challenges for high-accuracy classification in fine-grained movements. This study proposes an optimized 1D convolutional neural network (1D-CNN) framework for classifying 20 distinct fine-grained finger gestures using raw sEMG data from an 8-channel wearable Myo Armband sensor. Unlike traditional methods that rely on manual feature engineering, the proposed 1D-CNN performs end-to-end learning to automatically extract temporal features. The research specifically investigates the impact of temporal windowing strategies, ranging from 400 to 750 ms, on model performance. Experimental results demonstrate that the optimized 1D-CNN achieves a peak test accuracy of 94.4% with a 550 ms window size, demonstrating the model’s robustness across complex gesture classes and significantly outperforming the baseline principal component analysis- support vector machine (PCA-SVM) method which only attained 73.0% accuracy. While the model achieved perfect classification (100%) for index, middle, and little finger movements, a performance drop was observed in thumb recognition (50%) due to muscular crosstalk from deeper anatomical layers. These findings indicate that the integration of optimized windowing and 1D-CNN architectures provides a highly reliable solution for complex large-scale gesture recognition, offering a robust foundation for the next generation of multi-functional prosthetic hands.
Volume: 16
Issue: 3
Page: 1576-1587
Publish at: 2026-06-01

Hybrid deep learning for sentiment analysis of online student experiences

10.11591/ijai.v15.i3.pp2736-2749
Raja Ouadad , Hicham Mouncif
The COVID-19 pandemic disrupted millions of lives worldwide, and social media platforms became a significant outlet for people to share their emotions and experiences, providing valuable insights into the challenges and opportunities of remote education. This paper analyzes student sentiments about online learning during the pandemic using Twitter data. An experimental approach is developed to analyze public comments, focusing on the sentiment expressed in tweets related to online education. A hybrid deep learning model, based on the logistic regression (LR) sentiment model, is used to predict sentiment from a large dataset of online learning-related tweets. After performing n-gram analysis to extract key topics, tweets are classified into sentiment classes. The proposed convolutional long short term memory (Conv-LSTM) and convolutional bidirectional long short-term memory (Conv-BiLSTM) models are trained on tweets annotated with granular sentiment classifications, achieving validation accuracies of 93% and 95%, respectively. This work provides meaningful insights into the emotional effects of online learning during the pandemic, contributing to the understanding of students' experiences and challenges in remote education.
Volume: 15
Issue: 3
Page: 2736-2749
Publish at: 2026-06-01

Machine learning-enabled joint antenna selection and precoding

10.11591/ijai.v15.i3.pp2369-2376
Monica Nilesh Kalbande , Kanala Sai Madhuri , M. Venkateswara Rao , Saradha Rani Sabbavarapu , Rajyalakshmi Uppada , Lakshmi Durga Rajamahendravarapu
Joint antenna selection (AS) and precoding design is essential for improving spectral efficiency and energy efficiency in multi-antenna wireless communication systems. However, conventional optimization-based solutions rely on exhaustive search and iterative processing, leading to high computational complexity that limits real-time applicability. This work proposes a machine learning-enabled framework that shifts the computational burden from online operation to offline training. Optimal AS and precoding decisions are first generated offline using model-based optimization under diverse channel conditions. A supervised machine learning model is then trained to learn the relationship between channel state information (CSI) and optimal transmission configurations. During online operation, the trained model enables fast and efficient AS with significantly reduced processing time. Numerical results demonstrate that the proposed approach achieves near-optimal system performance while substantially lowering computational complexity, making it well suited for real-time and next-generation wireless communication systems.
Volume: 15
Issue: 3
Page: 2369-2376
Publish at: 2026-06-01

Parkinsonian gait classification in older adults using time–frequency spectrograms and 2D convolutional neural network

10.11591/ijai.v15.i3.pp2698-2708
Kazi Ashikur Rahman , Ezreen Farina Shair , Nur Zawani Saharuddin , Muhammad Hazwan Adlin Jumaris
Gait disorders in adults aged 50 years and above are a common concern and are often linked to reduced mobility, a higher risk of falls, and a lower quality of life. This study presents a deep learning-based approach to detect gait disorders using vertical ground reaction force (vGRF) signals. The data were collected from older adults, including individuals with Parkinson’s disease (PD) and healthy controls, using force-sensitive resistor sensors. The raw signals were first processed using band-pass filtering and wavelet denoising to remove noise and unwanted variations. After that, the signals were converted into time–frequency representations using the continuous wavelet transform (CWT). These representations were then used as input to a convolutional neural network (CNN) for classification. The model achieved a validation accuracy of 93.48%, with precision, recall, and F1-score all above 92% for both groups. The results show that combining CWT with CNN provides a reliable and efficient way to detect gait disorders. This approach can support clinical evaluation by offering a practical and scalable method for analyzing gait patterns in older adults.
Volume: 15
Issue: 3
Page: 2698-2708
Publish at: 2026-06-01

Transforming electric vehicle charging through solar integration and high-frequency magnetic induction for seamless wireless power transfer

10.11591/ijece.v16i3.pp1118-1131
Selvan Chinnaiyan , Prabhakar Manickam , Madhu Chandra G. , Aarthi V. , Narendra Babu C. R.
The rapid adoption of electric vehicles (EVs) is constrained by limited charging infrastructure, prolonged charging duration, grid dependency, and inefficiencies in conventional wireless charging systems. To address these challenges, this paper proposes a solar-integrated high-frequency inductive wireless charging framework that enables efficient, contactless, and partially dynamic EV charging. The proposed system combines a photovoltaic (PV) energy harvesting subsystem with maximum power point tracking (MPPT), a high-frequency resonant inductive coupling mechanism using a series–series (SS) topology, and an intelligent solar inductive synergy optimization algorithm (SISOA) for adaptive power and energy storage management. The theoretical foundation of the system is based on Faraday’s law of electromagnetic induction and resonant magnetic coupling to enhance mutual inductance and power transfer efficiency. Simulation studies conducted in MATLAB/Simulink demonstrate that the proposed approach achieves a mutual inductance of 82.5, an output voltage of 500 V, and an output power of 4,800 W, while reducing overall power losses to 21.18% and improving system efficiency to 94.5%. The results further reveal that vehicle speed and the number of receiver coils significantly influence charging effectiveness and state-of-charge performance.
Volume: 16
Issue: 3
Page: 1118-1131
Publish at: 2026-06-01

Hybrid convolutional neural network-bidirectional long short-term memory model for Arabic sentence readability assessment

10.11591/ijai.v15.i3.pp2849-2862
Mohamed Amine Ouassil , Mohammed Jebbari , Rabia Rachidi , Mouaad Errami , Soufiane Hamida , Bouchaib Cherradi , Abdelhadi Raihani
In the current educational landscape, a large number of educators prefer using generative artificial intelligence techniques to produce textual content to be presented for learning. However, these generated texts may not meet the specific needs of learners or align with their abilities. Many traditional methods and techniques can be employed to assess the complexity of a text, such as traditional readability formulas, but these techniques are time consuming and labor-intensive. In this paper, we introduce a deep learning approach for automatically evaluating the readability of Arabic texts by analyzing and classifying sentences into different difficulty levels within educational content. The initial stage of the proposed approach is preprocessing textual content and leveraging natural language processing (NLP) methodologies for feature extraction, such as Word2Vec. The approach then concentrates on refining and evaluating a deep learning model to classify text into different readability levels. This paper introduces a hybrid classification model that combines convolutional neural networks (CNNs) and bidirectional long short-term memory (BiLSTM) layers, attaining an accuracy of 96.68%. This model demonstrates the significance of applying hybrid deep learning models in analyzing educational materials and establishes a foundation for subsequent progress in the field of automated Arabic readability assessment.
Volume: 15
Issue: 3
Page: 2849-2862
Publish at: 2026-06-01

Applications of machine intelligence and decision analytics in hospitality: a systematic review

10.11591/ijai.v15.i3.pp2024-2040
Omar Jawabreh , Ehab Abdul Raheem Alshatnawi , Rami Mahmoud
This systematic review synthesizes how machine intelligence (MI) and decision analytics (DA) are deployed across hospitality functions (customer relationship management (CRM)/personalization, revenue management (RM) and pricing, and operations) and clarifies barriers and research frontiers. Following a PRISMA-oriented protocol for the 2008–2025 period, the study searched Scopus, Web of Science, IEEE/Elsevier, and SpringerLink using combined terms for MI/artificial intelligence (AI)/machine learning (ML) and hospitality/tourism with CRM, revenue/pricing, forecasting, and operations. The inclusion criteria focused on peer-reviewed English studies reporting concrete models/applications or empirical evidence. Studies were thematically coded into four streams: CRM and personalization; RM and dynamic pricing; operations and scheduling; and governance (privacy/ethics/skills). The findings consistently show that MI/DA improves customer segmentation, demand forecasting, dynamic pricing (revenue per available room (RevPAR)), and workforce/maintenance efficiency. Yet, adoption is slowed by data governance (privacy/general data protection regulation (GDPR)) and siloed systems. Unlike prior broad IT/e Tourism reviews, this study integrates recent MI advances (spatiotemporal deep learning, real-time decisioning, and human–AI teaming) into a hospitality-specific decision pipeline and outlines a testable research agenda around scalability/drift, privacy-by-design, and controlled field experiments.
Volume: 15
Issue: 3
Page: 2024-2040
Publish at: 2026-06-01

Designing self-healing database fabrics for real-time payment rails

10.11591/ijece.v16i3.pp1360-1368
Raghu Gollapudi
Real-time payment platforms operating at scale face an unforgiving operational reality: even brief outages translate directly into failed transactions, regulatory exposure, and eroded customer trust. Database replication and failover automation have matured considerably over the past two decades, yet a troubling blind spot remains. Recovery frameworks built for general-purpose distributed systems were never designed with settlement finality in mind, and that design omission leaves payment operators exposed to split-brain scenarios that generic high-availability tooling cannot reliably prevent. This paper addresses that omission head-on through a self-healing database fabric purpose-built for payment rail environments. The proposed autonomous resilience fabric architecture (ARFA) operates across three coordinated layers: a continuous monitoring layer that harvests telemetry from compute, storage, and network subsystems; a decision layer that fuses rule-based heuristics with an ensemble of isolation forests, recurrent neural networks, and gradient boosting classifiers to separate genuine fault conditions from transient noise; and a deterministic action layer that executes recovery procedures anchored to explicit settlement finality constraints. In fault injection trials covering node crashes, network partitions, replication lag, and performance degradation, the architecture cut average recovery times by 88% against manual baselines, restoring service in roughly 8 seconds rather than the 180 seconds that human-driven remediation typically requires. False positive rates held below 2% across all failure categories, and the system achieved a 98% recovery success rate. Taken together, these results make a practical case that autonomous resilience and regulatory compliance reinforce rather than conflict with each other when the regulatory constraints are designed in from the start.
Volume: 16
Issue: 3
Page: 1360-1368
Publish at: 2026-06-01

A multi-modal framework for improving the accuracy of phishing email detection

10.11591/ijece.v16i3.pp1608-1625
Lamees Mohamed Faraj , Sayed Abdel-Gaber , Hanan Fahmy
Phishing emails continue to pose a significant cybersecurity threat, particularly through the increasing use of malicious attachments to evade traditional text-based detection systems. Most existing approaches focus primarily on email content, creating a blind spot in attachment-aware phishing detection. This paper proposes a multi-modal phishing email classification model that integrates email header features, body text analysis, and attachment inspection within an ensemble learning framework. Independent machine learning classifiers are employed for each email component, and a majority voting mechanism is used to determine the final classification decision. The proposed model is evaluated using publicly available email and attachment datasets that are combined to simulate attachment-bearing phishing emails. Experimental results demonstrate strong detection performance across multiple evaluation metrics. Nevertheless, the study acknowledges the limitation of using synthetically paired email bodies and attachments, which may not fully capture real-world semantic relationships. The findings highlight the importance of incorporating attachment-aware analysis into phishing detection systems and provide a foundation for future research on semantic consistency modeling and transformer-based architectures.
Volume: 16
Issue: 3
Page: 1608-1625
Publish at: 2026-06-01

Effects of sparse datasets on time interval-aware self-attention sequential recommendation models

10.11591/ijai.v15.i3.pp2761-2773
Weishan Ooi , Lee-Yeng Ong , Meng-Chew Leow
Recommendation models serve as crucial filters in managing information, yet they face a few crucial challenges, such as capturing user-item interaction behaviors in sparse datasets. Data sparsity refers to an issue where there is a lack of interactions or missing values in the recommendation dataset. A sparse dataset with a massive number of missing values and interactions leads to more dynamic user behaviors, which suffers a poor recommendation quality. The self-attention mechanism from Transformer can alleviate the effects of data sparsity in datasets by assigning weights to items of interaction behaviors. This allows the model to capture the user dependencies in complex user behavior, which is beneficial for sparse datasets with patterns that are not immediately apparent. This approach has shown its capability to handle large and sparse datasets, as seen in time interval-aware self-attention sequential recommendation model (TiSASRec). It utilized the self-attention mechanism, considering the timestamp and absolute positions of items to estimate the higher attention weights to show the importance of recent items. Thus, this study aims to investigate the effects of sparse datasets by comparing the performance of TiSASRec model with self-attention based sequential recommendation model (SASRec), which excludes time interval-awareness.
Volume: 15
Issue: 3
Page: 2761-2773
Publish at: 2026-06-01

AI-driven log reduction and storage optimization for security operations

10.11591/ijece.v16i3.pp1417-1424
Nutthakorn Chalaemwongwan
In this study, we present an AI-driven framework that integrates semantic log reduction with compliance-aware storage optimization, specifically designed for security operations center (SOC) and managed security service provider (MSSP) environments. Traditional approaches such as uniform compression, keyword filtering, and static tiering often either miss critical anomalies or preserve redundant noise, leading to excessive storage use, slower search performance, and analyst fatigue. The proposed framework addresses these challenges by combining three components: semantic reduction of repetitive entries, anomaly-focused retention supported by self-supervised models, and adaptive tiering aligned with regulatory requirements. Evaluations on HDFS, BGL, CICIDS2017, and Suricata datasets achieved 70%–80% log reduction, 55%–65% storage savings, recall rates above 95%, and a one-third reduction in query latency. These results demonstrate that pre-index reduction, together with anomaly- and compliance-aware retention, offers a scalable and regulator-ready solution for operational security environments.
Volume: 16
Issue: 3
Page: 1417-1424
Publish at: 2026-06-01

Hybrid systems modelling and control using multiple mixed logical dynamical predictive model control: Application to a three-tank spherical system

10.11591/ijece.v16i3.pp1148-1158
Tahar Benaissa , Mohamed Fouzi Belazreg , Khaled Halbaoui , Belaid Djaroum , Djamel Boukhetala
This study employs the mixed logical dynamical (MLD) framework for modelling, simulating, and controlling hybrid dynamical systems. Hybrid systems, which combine continuous-time dynamics and discrete logical events, pose significant challenges for conventional control strategies, such as proportional-integral-derivative (PID) controllers, particularly under complex operational constraints. To address these challenges, the MLD formalism provides a unified representation that integrates differential equations, logical rules, and inequality constraints. Based on the MLD model, a multivariable hybrid model predictive control (HMPC) approach is designed to optimize control system performance and operational efficiency over a prediction time horizon. At each sampling time step, a mixed quadratic programming (MIQP) optimization problem is solved online to determine the control law. The proposed control approach is applied to a three-spherical tank system, where simulation and experimental results demonstrate its effectiveness in ensuring stability, minimizing tracking errors, and satisfying physical constraints. These results underscore the relevance of MLD-based predictive control approaches for the optimization and advanced control of complex multivariable hybrid dynamical systems in industrial fields.
Volume: 16
Issue: 3
Page: 1148-1158
Publish at: 2026-06-01
1 2 3
...
2091 2092
Show 72 of 2092

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration