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

30,185 Article Results

Optimizing planar micro-transformer performance

10.12928/telkomnika.v24i1.27276
Tahar; University of Science and Technology of Oran USTO-MB Alili , Fatima Zohra; University of Science and Technology of Oran USTO-MB Medjaoui , Azzedine; Nour El Bachir University Center Hamid , Abderahim; National Polytechnic School of Oran Maurice Audin Mokhefi , Yacine; Nour El Bachir University Center Guettaf , Hocine; Nour El Bachir University Center Guentri
Faced with new requirements for isolated switching power supplies with high efficiency and power density, planar transformer technology has emerged as a serious alternative to wound components. The work presented in this paper addresses the issue of developing planar transformers in the context of low-power electronics, where volume and weight constraints are paramount. The flat shape of the coils and the interlacing of the windings do not allow for control of magneto-thermal phenomena. Although scientific literature offers numerous simulation tools to aid in the design of such transformers, it must be noted that they do not allow for a rigorous account of these phenomena. In this paper, methods and a geometric and electrical sizing tool in planar technology are used for the design of flyback direct current to direct current (DC/DC) converters. Methods for dimensioning and estimating temperature rise are presented and compared in order to develop calculation tools for design purposes. This study enabled us to observe the distribution of the magnetic field, the role of ferrite, the distribution of currents and voltages in the coils, and the distribution of temperature in our device. It should be noted that conductive and convective heat transfer processes were considered in steady state.
Volume: 24
Issue: 1
Page: 313-328
Publish at: 2026-02-01

Optimized edge-aware frequency-guided filtering for robust image denoising

10.12928/telkomnika.v24i1.27338
Iman; Karabuk University Elawady , Ahmet Reşit; Karabuk University Kavsaoğlu , İsmail Rakıp; Karabuk University Karaş
The problem of denoising intrusion is still of great concern in computational imaging because of the trade-off between noise reduction and image structure and details recovery. This paper proposes an optimized edge-aware fast adaptive guided filter (E-FAGF) combining wavelet-domain decomposition, edge-awareness, and lightweight deep learning for efficient and effective denoising. The biorthogonal wavelet transform is employed to decompose noisy images into low- and high-frequency sub bands and an improved edge-attention map for selective high-frequency denoising. Regularization parameters are estimated pixel-wise by a compact convolutional neural network (CNN), allowing spatial-varying filtering to be done with multi-scale processing. The resultant E-FAGF consistently outperforms the state of the art on this dataset: on BSD500 for speckle and Gaussian noise (peak signal-to-noise ratio (PSNR) of 39.63 dB and 33.97 dB, respectively), and competitive performance for Poisson noise (30.84 dB) a large margin compared to the reference bilateral and non-local means. Our method maintains high structural similarity (up to 0.97 in structural similarity index measure (SSIM)), runs at 0.015 seconds per 512×512 image on graphics processing unit (GPU), and can be applied without dataset specific training. These results suggest the possibility of E-FAGF to achieve a balance between classical efficiency and learning-based adaptability, thereby forming a new scenario to combine fast and reliable image restoration for actual scenarios.
Volume: 24
Issue: 1
Page: 219-227
Publish at: 2026-02-01

Advanced microwave imaging and artificial neural networks for early detection and localization of breast tumors

10.12928/telkomnika.v24i1.27126
Abdelfettah; University of Mustapha Stambouli Miraoui , Lotfi; School of Applied Sciences Tlemcen Merad , Djalal; LARATIC Laboratory at National Institute of Telecommunications and Information Technology and Communication (ENSTTIC) Ziani-Kerarti
This study investigates the detection and localization of breast tumors based on dielectric property differences between cancerous and normal tissues. A microwave imaging technique integrated with artificial neural networks (ANNs) is proposed as a noninvasive alternative to conventional screening methods such as mammography and magnetic resonance imaging (MRI). A breast model with a 2.5 mm spherical tumor was designed using CST Microwave Studio. Simulation results show that the ANN achieves a detection rate close to 100%, providing negative outputs for tumor-free cases and positive outputs for cases with tumors. Additionally, ANN outputs strongly correlate with the actual tumor positions in the simulated environment. These findings suggest that microwave imaging combined with ANNs offers a cost effective, radiation-free, and patient-friendly solution for the early detection and localization of breast cancer, with promising potential for clinical translation.
Volume: 24
Issue: 1
Page: 240-248
Publish at: 2026-02-01

Development of generalized principal component analysis using multiple imputation genetic algorithm

10.11591/ijai.v15.i1.pp454-468
Fahrezal Zubedi , I Made Sumertajaya , Khairil Anwar Notodiputro , Utami Dyah Syafitri
In this study, we propose an innovative method called the integrated GPCA MIGA, which integrates the multiple imputation genetic algorithm (MIGA) and generalized principal component analysis (GPCA) to perform missing value imputation and data dimensionality reduction simultaneously. The approximated original data produced by GPCA serves as the basis for MIGA to update missing values in the next iteration. At the same time, GPCA refines the low-dimensional representation using the latest imputation results from MIGA, thereby balancing the accuracy of missing value imputation and the stability of dimensionality reduction. The objective of this study is to evaluate the performance of the integrated GPCA-MIGA and analyze trends in human development at the district/city level in Indonesia. The findings of this study show that the integrated GPCA-MIGA effectively reduces the dimensionality of data containing missing values compared to other methods. The integrated GPCA-MIGA method was applied to human development data. The results were then visualized using a biplot, which revealed that human development trends in Jayawijaya from 2019 to 2022 indicate progress in school enrollment rates for ages 16–18 years.
Volume: 15
Issue: 1
Page: 454-468
Publish at: 2026-02-01

Need analysis of holistic reading literacy teaching module integrating critical thinking and character

10.11591/ijere.v15i1.35171
Rahma Dona , Siti Rahaimah Ali , Florence Yulisinta , Restiana Ertika Latifah
The low levels of reading literacy and character development remain significant educational challenges among Indonesian students in the 21st century. Despite emphasizing higher-order thinking skills (HOTS) and character education in the national curriculum, elementary teachers struggle to effectively integrate these aspects into reading literacy teaching in the Bahasa Indonesia subject. This study identifies teachers’ challenges, difficult literacy topics, and the need for a structured module integrating HOTS and character. Using a design and development research (DDR) approach, an online survey of 287 grade 4 teachers from 171 elementary schools in Depok, West Java, was conducted. This study is the first to develop a grade-4-tailored holistic-HOTS-character (H2C) framework for reading literacy instruction, integrating cognitive, social, emotional, psychomotor, and moral elements. Additionally, teachers found independently retrieving and synthesizing information from multiple sources (22.30%) and teaching main ideas and supporting details (21.95%) to be the most challenging topics. Teachers strongly supported the development of an H2C teaching module (M=4.28 out of 5). The study also presents a visual H2C conceptual model to guide classroom implementation. This needs-analysis study confirms that teachers perceive a structured, holistic reading literacy teaching module as essential for enhancing student engagement, HOTS, and character development.
Volume: 15
Issue: 1
Page: 760-770
Publish at: 2026-02-01

Role of empowering leadership in improving faculty engagement and motivation in higher education institutions

10.11591/ijere.v15i1.32492
Sharish Abdullah , Riffat-un-Nisa Awan , Abida Parveen , Zunaira Fatima , Ghazanfar Ali
Higher education institutions face challenges in maintaining faculty motivation and engagement. Leadership empowerment is usually perceived as an important factor in fostering employee outcomes. This study intended to examine the effect of leadership empowerment on employee engagement and motivation at higher education levels. It was a quantitative study and a survey was used to collect data. The questionnaire consisted of 59 items, regarding three variables i.e., leadership empowerment (leading by example, participative decision-making, coaching, informing, and interaction with team), motivation (intrinsic motivation and extrinsic motivation), and employee engagement (affective engagement, intellectual engagement). Data were collected from a sample of 200 teachers (130 males, 70 females) from the University of Sargodha. The data was analyzed using Pearson correlation and regression analysis. The study concluded that leadership empowerment significantly affected employee engagement and motivation. Leading by example significantly affected intellectual and affective engagement and both intrinsic and extrinsic motivation, whereas only participative decision-making affected extrinsic motivation. It is suggested that leaders should create a supportive and empowering environment by setting examples, recognizing efforts, and providing growth opportunities. They should empower employees to feel valued and motivated to contribute to the organization’s success.
Volume: 15
Issue: 1
Page: 40-52
Publish at: 2026-02-01

Enhancing reflective elements of intelligent reflective surfaces in 6G communications using artificial intelligence

10.12928/telkomnika.v24i1.27307
Jehan Kadhim Shareef; University of Thi-Qar Al-Safi , Abbas Thajeel Rhaif; University of Thi-Qar Alsahlanee
The dynamic landscape of 6G communication networks necessitates innovative strategies to address energy inefficiency and signal degradation in densely populated regions with limited line-of-sight (LOS) coverage. A novel technology known as an intelligent reflecting surface (IRS) has emerged; it can dynamically modify the characteristics of electromagnetic waves to enhance signal propagation. Unfortunately, current IRS models frequently neglect the balance between energy efficiency (EE) and the quantity of reflective elements (N) in Rayleigh fading scenarios. This study introduces an algorithm called dynamic-static particle swarm optimization (DS-PSO) aimed at improving EE and decreasing the quantity of reflective components in the performance optimization of IRS. The research assesses the proposed model in comparison to single-input single-output (SISO) systems, conventional IRS models, and IRS models from prior studies within a realistic urban framework. The optimized IRS, which only uses seven reflective elements, has a peak EE of 366 Mbit/Joule. This is a big improvement over IRS models from earlier research, as shown by the numbers. The findings indicate that artificial intelligence (AI)-driven optimization can enhance IRS technology for sustainable and efficient 6G networks.
Volume: 24
Issue: 1
Page: 22-33
Publish at: 2026-02-01

A decoupling-based multivariable H∞ controller for PMSM speed and current regulation

10.12928/telkomnika.v24i1.27515
Farid; Higher School of Applied Sciences of Tlemcen Oudjama , Mohammed; University of Tlemcen Messirdi , Mokhtar; University Centre of Maghnia Bourdim , Abelmadjid; University of Tlemcen Boumédiène
High precision speed regulation of the permanent magnet synchronous motor (PMSM) is a critical challenge in modern industrial applications, including electric vehicles and traction systems. This task is significantly affected by external disturbances, such as variable load torque, as well as physical phenomena often neglected in analytical models, such as magnetic circuit saturation or thermal variations in electrical parameters. In this context, conventional control methods often fail to ensure both dynamic performance and robustness. This paper proposes a multivariable H∞ control strategy based on field-oriented control (FOC) and d/q decoupling to design a robust and high-performance controller. The diagonal multiple-input multiple-output (MIMO) model, linking the direct-axis voltage𝑣𝑑to the current 𝑖𝑑and the quadrature-axis voltage 𝑣𝑞to the rotational speed 𝜔𝑟, is derived directly from the decoupling principles of FOC, without relying on linearization around an operating point or modeling of parametric uncertainties. The H∞ controller is synthesized using the standard configuration, with carefully selected weighting functions to ensure dynamic performance, closed-loop stability, and effective disturbance rejection. Numerical simulations demonstrate that the proposed controller achieves accurate speed reference tracking, fine current regulation, and fast load disturbance rejection, confirming its effectiveness and robustness. This approach provides an advanced alternative to conventional control methods by fully exploiting the multivariable structure of the system.
Volume: 24
Issue: 1
Page: 293-301
Publish at: 2026-02-01

A practical approach to Candi Siwa 3D reconstruction with COLMAP and Nerfstudio

10.12928/telkomnika.v24i1.27212
Helena; Prasetiya Mulya University Widiarti , Rokhmat; Prasetiya Mulya University Febrianto , Agung; Prasetiya Mulya University Alfiansyah
We demonstrate a practical approach for large-scale object three-dimensional (3D) reconstruction with freely available frameworks, COLMAP and Nerfstudio. We performed the reconstruction of a temple named Candi Siwa, located at Prambanan Site, which is situated between Central Java and Yogyakarta Province, Indonesia. We utilized COLMAP and Nerfstudio as platforms for 3D reconstruction from images captured by an everyday smartphone. In the 3D model construction process, COLMAP generates a dense point cloud, whereas Nerfstudio generates a scene from source images. We selected 96 images of Candi Siwa to perform reconstruction using COLMAP. As a result, a 3D model for the temple with a clear structure and color was observed. A scene rendered in MP4 format was also generated using Nerfstudio. Additionally, we performed the 3D reconstruction from 150 images taken by the public and found them insufficient for constructing the object. This occurred despite the number of images being larger than those used in the previous reconstruction. The results indicate that the success of a crowdsourcing project for reconstructing a large-scale object should consider not only the number of images but also the variation in point of view and the completeness of the whole structure.
Volume: 24
Issue: 1
Page: 249-257
Publish at: 2026-02-01

Energy-efficient AI-enhanced secure routing for protecting IoT networks from advanced attacks

10.11591/ijeecs.v41.i2.pp731-739
Leelavathi R. , Vidya A.
This paper proposes artificial intelligence-enhanced secure routing (AIRS), a lightweight AI-enhanced secure routing protocol for internet of things (IoT) networks operating under advanced routing attacks. Unlike existing approaches that treat intrusion detection and routing separately, AIRS tightly integrates anomaly scoring into trust-aware routing decisions using a compact random forest model designed for constrained nodes. The anomaly detector is trained offline on simulated IoT traffic features and deployed for real-time inference during routing. Extensive Cooja simulations demonstrate that AIRS improves intrusion detection accuracy and packet delivery while reducing energy consumption compared to secure-RPL and trust-LEACH. The current validation is limited to simulation environments, and real-world testbed evaluation is left for future work.
Volume: 41
Issue: 2
Page: 731-739
Publish at: 2026-02-01

Enhancing medical language models with big data technologies

10.11591/ijai.v15.i1.pp289-299
Ayoub Allali , Ibtihal Abouchabaka , Najat Rafalia
In this study, we present an end-to-end, big-data–driven framework for continuously enriching and fine-tuning large language models (LLMs) with the latest professional and scientific medical knowledge. Streaming updates from premier sources such as The New England Journal of Medicine (NEJM) are ingested via an Apache Kafka cluster for low-latency delivery and durably archived in a three-node Apache Hadoop (Hadoop distributed file system (HDFS)) system. Each new article is preprocessed into high dimensional embeddings and indexed in a Milvus vector database to enable sub-second semantic retrieval over millions of records. At query or batch time, our retrieval-augmented generation (RAG) module retrieves the top-k relevant embeddings from Milvus and injects them into prompts for DeepSeek-R1, GPT-4o-mini, and Llama 3, models which are hosted, fine tuned, and served via Ollama on an NVIDIA GeForce RTX 3050 Ti GPU for efficient inference and continual learning. The enriched outputs are seamlessly delivered to end users through a Telegram bot programmed in Python using the Telebot library, linking the RAG-enhanced LLMs to an intuitive chat interface. Our Kafka, HDFS, Milvus, RAG, LLM, or Telegram bot pipeline demonstrably improves factual accuracy and topical currency of AI-generated medical insights across clinical decision support, patient engagement and education, drug discovery and development, virtual health assistants, and mental health support, laying the groundwork for truly intelligent, responsive, and data-driven healthcare solutions.
Volume: 15
Issue: 1
Page: 289-299
Publish at: 2026-02-01

Optimized IMC with GWO algorithm and variable switching function for voltage regulation of SEPIC converter

10.12928/telkomnika.v24i1.27330
Reza; Shahrood University of Technology Fazeli , Mohammad; Shahrood University of Technology Haddad Zarif , Mahmoud; Islamic Azad University Zadehbagheri , Tole; Universitas Ahmad Dahlan Embedded System and Power Electronics Research Group Sutikno
With the growing application of single-ended primary-inductor converter (SEPIC) converters in power electronic systems, precise output voltage regulation under uncertainties and nonlinear conditions remains a significant challenge. Although internal model control (IMC) effectively addresses issues arising from unstable zeros and fixed time delays in non-minimum phase systems, its performance can degrade under large transient errors or sudden disturbances, leading to control signal saturation and instability. In this study, a modified IMC scheme is proposed, which integrates a variable switching function into the control structure. This addition enhances the robustness of the system by dynamically adapting the control effort to mitigate abrupt changes in the control signal and stabilize the output voltage. Furthermore, it prevents controller saturation during large-signal deviations, thereby improving transient response and maintaining system stability. The design parameters of the controller are optimized using the gray wolf algorithm to achieve an optimal balance between voltage overshoot, settling time, and closed-loop stability. Simulation results under various operating conditions confirm the superior performance of the proposed control method compared to conventional IMC.
Volume: 24
Issue: 1
Page: 258-270
Publish at: 2026-02-01

Adaptive deformable feature augmentation and refinement network for scene text detection and recognition

10.11591/ijai.v15.i1.pp831-840
Ratnamala S. Patil , Geeta Hanji , Rakesh Hudud
Scene text recognition (STR) is the task of detecting and identifying text within images captured from natural scenes, a challenging process due to variations in text appearance, orientation, and background complexity. The proposed methodology, adaptive deformable feature augmentation and refinement network (ADFARN), is designed to address these challenges by combining deformable convolutional networks for robust enhanced feature extraction with a novel deep feature refinement (FRE) that leverages refinement for precise text localization. This approach enhances the differentiation between text and background, significantly improving recognition accuracy. The ADFARN methodology includes a comprehensive process of feature extraction, deep feature augmentation module (DFAM), and the generation of score and threshold maps through differentiable binarization. The adaptive nature of the model allows it to handle low resolution and partially occluded text effectively, further increasing its robustness. Additionally, the proposed method aligns visual and textual features seamlessly. Extensive performance evaluation on the common objects in context (COCO)-Text dataset demonstrates that ADFARN outperforms existing state-of-the-art methods in terms of precision, recall, and F1-scores, establishing it as a highly effective solution for STR in real world applications.
Volume: 15
Issue: 1
Page: 831-840
Publish at: 2026-02-01

SELLA: An IoT-based smart shopping trolley with real-time RFID tracking and automated checkout

10.12928/telkomnika.v24i1.27464
Hadj; Djillali Liabes University of Sidi Bel Abbes Zerrouki , Salima; Djillali Liabes University of Sidi Bel Abbes Azzaz-Rahmani
The contemporary retail sector faces a persistent challenge in enhancing in store customer experience, primarily due to inefficiencies at checkout. This paper presents smart e-cart for lean logistics application (SELLA), a smart shopping trolley system engineered to eliminate this bottleneck. The system’s architecture is centered on a Raspberry Pi 4 microcontroller, orchestrating an ultra-high frequency (UHF) radio frequency identification (RFID) subsystem for instantaneous, non-line-of-sight product identification, and a responsive 7-inch touchscreen graphical user interface (GUI) developed in PyQt. The core contribution lies in developing a self contained shopping solution with integrated payment processing, supported by comprehensive performance validation. We present a detailed methodology, including the system’s multi-threaded software architecture and core operational algorithm. Experimental evaluation demonstrates a mean tag detection accuracy of 98.2% under optimal conditions, a robust user interface (UI) latency of under 500 ms, and an average central processing unit (CPU) utilization of 28%, proving system efficiency. Comparative analysis confirms that SELLA’s integration of on-trolley automated payment and detailed performance metrics represents a significant advancement over existing prototypes. The system provides a validated, high-performance solution for next-generation smart retail environments.
Volume: 24
Issue: 1
Page: 184-195
Publish at: 2026-02-01

Human activity recognition using selective kernel network-2D convolutional neural network with ArcFace loss

10.11591/ijai.v15.i1.pp350-360
Banushri Srinivasaiah , Jagadeesha Ramegowda
Human activity recognition (HAR) is a widely adopted technique in applications requiring accurate identification of human actions. However, HAR approaches often face challenges in generalizing across complex datasets with multi-view variations, resulting in reduced classification accuracy. Existing classifiers face shortcomings in predicting human activities due to the presence of irrelevant video frames, leading to frequent misclassifications. This research proposes a selective kernel network-2D convolutional neural network with additive angular margin loss for deep face recognition (SKN-2D-CNN with ArcFace loss) to recognize human activity effectively. SKN dynamically adapts the receptive field for learning multi scale spatial features, enhancing the recognition of intricate human activities with varying motion scales. In the embedding space, ArcFace loss introduces an angular margin penalty that improves inter-class separability and intra class compactness for recognition. Together, the proposed method minimizes misclassification in human activity by improving the robustness of feature representation. Feature extraction using visual geometry group 19 (VGG19) captures spatial features like edges, textures and shapes from video frames, enhancing the model’s ability to differentiate between complex human activities. The proposed method achieves high accuracy of 99.16 and 98.75% on the UCF101 and HMDB-51 datasets, respectively, when compared with existing methods such as CNN and bidirectional gated recurrent unit (BiGRU).
Volume: 15
Issue: 1
Page: 350-360
Publish at: 2026-02-01
1 2 3
...
2012 2013
Show 65 of 2013

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