cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Alfian Maarif
Contact Email
alfianmaarif@ee.uad.ac.id
Phone
-
Journal Mail Official
biste@ee.uad.ac.id
Editorial Address
-
Location
Kota yogyakarta,
Daerah istimewa yogyakarta
INDONESIA
Buletin Ilmiah Sarjana Teknik Elektro
Published by Universitas Ahmad Dahlan
ISSN : 26857936     EISSN : 26859572     DOI : 10.12928
Core Subject : Engineering,
Electrical & Electronics Engineering
Buletin Ilmiah Sarjana Teknik Elektro (BISTE) adalah jurnal terbuka dan merupakan jurnal nasional yang dikelola oleh Program Studi Teknik Elektro, Fakultas Teknologi Industri, Universitas Ahmad Dahlan. BISTE merupakan Jurnal yang diperuntukkan untuk mahasiswa sarjana Teknik Elektro. Ruang lingkup yang diterima adalah bidang teknik elektro dengan konsentrasi Otomasi Industri meliputi Internet of Things (IoT), PLC, Scada, DCS, Sistem Kendali, Robotika, Kecerdasan Buatan, Pengolahan Sinyal, Pengolahan Citra, Mikrokontroller, Sistem Embedded, Sistem Tenaga Listrik, dan Power Elektronik. Jurnal ini bertujuan untuk menerbitkan penelitian mahasiswa dan berkontribusi dalam pengembangan ilmu pengetahuan dan teknologi.
Arjuna Subject : -
Articles 351 Documents
 32   33   34   35   36 
A Portable Solar-Powered Wireless Charger: Design, Implementation, and Performance Analysis Alfarid Hendro Yuwono; Deshinta Arrova Dewi; Rajani Balakrishnan; Reza Rahmadian; Nafi Isbadrianingtyas; Widi Aribowo; Vugar Hacimahmud Abdullayev; Aliyu Sabo
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15226

Abstract

The increasing demand for portable and off-grid charging solutions has motivated the development of solar-powered wireless power transfer (WPT) systems for consumer electronics. This paper presents the design, implementation, and experimental performance evaluation of a portable solar-powered wireless charger that integrates a 3-Wp photovoltaic (PV) panel with a near-field inductive coupling WPT system operating at a resonant frequency of 90 kHz. The research contribution is a fully integrated, low-cost prototype that demonstrates the feasibility of combining solar energy harvesting with contactless inductive charging for mobile devices, addressing the gap in portable off-grid wireless charging solutions. The system comprises a solar panel connected to a powerbank serving as an energy buffer, a series-series (SS) compensated inductive coil pair, a high-frequency inverter, and an AC/DC rectifier stage. Experimental testing was conducted in Malang City, Indonesia, under natural sunlight conditions. Results showed that the solar panel output voltage ranged from 6.2 V to 6.8 V under direct sunlight, declining by more than 30% under cloudy conditions. Peak power transfer efficiency of 65.3% was achieved at the 90 kHz resonant frequency, and efficiency decreased inversely with coil separation distance, dropping from 65.3% at 0 cm to below 10% at 5 cm. The powerbank required approximately 460 minutes of solar charging to reach 4 V, and the mobile phone battery charged at an average rate of 8.5 minutes per 1% capacity increase, compared to approximately 4.2 minutes per 1% for a standard wired charger. The study demonstrates the practical feasibility of portable solar-WPT integration for outdoor and emergency charging applications, while identifying weather dependence and limited effective coil distance as primary constraints for future optimization. This research aligns with the United Nations Sustainable Development Goals (SDGs), particularly SDG 7 (Affordable and Clean Energy) by promoting renewable energy access and photovoltaic technology for off-grid communities, SDG 9 (Industry, Innovation and Infrastructure) through the development of innovative low-cost wireless charging infrastructure, SDG 11 (Sustainable Cities and Communities) by enabling resilient and portable energy solutions for underserved and emergency settings, and SDG 13 (Climate Action) by advancing clean energy alternatives that reduce dependence on fossil-fuel-based electricity.
Enhancing Federated Learning for Imbalanced Medical Image Classification through Adaptive Tuning and Autoencoder-Based Reconstruction Nadzurah Zainal Abidin; Amelia Ritahani Ismail; Cut Amalia Saffiera; Nurul A. Emran; Zammarah Nuha Abdullah
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15266

Abstract

Medical image classification has advanced significantly through deep learning techniques, yet its performance remains limited by class imbalance and decentralized data silos commonly found in healthcare settings. These issues reduce model sensitivity to rare but clinically important cases, and standard Federated Learning (FL) further struggles under non-independent and identically distributed (non-IID) data. To address this, an enhanced federated model integrating unsupervised autoencoder-based reconstruction and adaptive tuning is proposed. The research contribution is an enhanced FL model that improves minority-class detection and overall classification performance under imbalanced medical image distributions, while remaining applicable across decentralized healthcare data sources. The method incorporates an autoencoder to compute reconstruction error, enabling emphasis on underrepresented samples, while adaptive tuning dynamically adjusts local hyperparameters and global aggregation weights based on sample difficulty. This integration strengthens minority-class learning without requiring additional labels or altering the decentralized structure. Experimental evaluations were conducted using two benchmark medical image datasets across three induced imbalance ratios (1:10, 1:5, 1:2) for RetinaMNIST and naturally induced imbalance for PneumoniaMNIST dataset. Results show that under severe imbalance (1:10), the enhanced model improves minority-class recall by 59.6%, F1-score by 33.9%, and AUC-ROC by 13.3% compared to standard FL. At 1:5 imbalance, recall increases by 41.3% and F1-score by 26.5%, with accuracy gains up to 6.0%. Even under mild imbalance (1:2), the model maintains consistent improvements, achieving a 26.4% recall gain and 18.9% increase in F1-score. The performance of the enhanced model was further evaluated against three baseline FL models such as standard federated learning (FedAvg), FL with GAN augmentation, and FL with standalone autoencoder-based reconstruction. The results consistently confirm that federated learning integrating adaptive tuning and autoencoder-based reconstruction outperform the three baselines FL-models for accuracy, recall, and F1-score. These findings also demonstrate that the enhanced model provides scalable, coordination-free improvement in imbalanced federated medical image classification, offering stronger performance and stability across real-world heterogeneous settings.
Real-Time BISINDO Alphabet Recognition via Faster R-CNN Incorporating Skin Tone Diversity as a Classification Feature Lilis Nur Hayati; Anik Nur Handayani; Wahyu Sakti Gunawan Irianto; Rosa Andrie Asmara; Dolly Indra; Nor Salwa Damanhuri
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15587

Abstract

Indonesian Sign Language (Bahasa Isyarat Indonesia/BISINDO) enables communication for deaf individuals through hand gestures, yet limited public awareness creates significant barriers between deaf and hearing communities. Existing recognition systems often fail to generalize across diverse skin tones, reducing their effectiveness in inclusive real-world deployment. The contribution of this research is a BISINDO alphabet recognition system that integrates skin color features - extracted via HSV-based skin segmentation - as an additional preprocessing layer within the Faster R-CNN framework, explicitly improving detection robustness across varied skin tones. The dataset consists of 8,000 images from ten adult actors representing light, medium-brown, and dark skin tones, augmented through flipping and brightness variation, with a 90:10 training-to-testing ratio. The model was trained over 15,000 steps with a batch size of 24, selected through empirical validation to balance convergence stability and dataset size. Experimental results show that indoor conditions outperform outdoor settings due to controlled lighting. Light-skinned and dark-skinned participants achieved the highest accuracy of 87.5% and F1-score of 85.71%, while medium-brown-skinned participants showed slightly lower performance, likely attributed to greater variability in reflectance under mixed lighting. The system achieves 24 frames per second, demonstrating potential for real-time communication support. These findings confirm that Faster R-CNN with skin color feature integration is effective for BISINDO alphabet recognition, with skin tone diversity being a critical performance factor. Future work will explore larger participant pools and dynamic gesture recognition under varied real-world lighting scenarios.
Robust Speed Control of Permanent Magnet DC Motors Using an Arctic Puffin Optimized PI Controller and Nonlinear Disturbance Observer Ahmed Alkamachi
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15723

Abstract

Permanent magnet DC (PMDC) motors are widely used in many devices, such as in robotics, medical equipment, and industrial machinery, because they are small and easy to control. However, their operation can be affected by external disturbances such as load fluctuations. Conventional Proportional Integral (PI) controllers, although simple, are not sufficiently robust against such disturbances. This study proposes a novel control scheme for improving PMDC motor performance. It combines a simple PI controller with a Nonlinear Disturbance Observer (NDOB). A key advantage of the NDOB is its enhancement of robustness via actively estimating and compensating lumped disturbances. This makes the system more robust to disturbances and modelling errors while maintaining simplicity of structure and use. The controller parameters (PI gains and the NDOB low pass filter cutoff frequency) have been optimized using a custom algorithm called Arctic Puffin Optimization (APO) that ensure global optimal selection of the tuned parameters. The proposed combined weighted cost function allowed for the best balance between response speed, disturbance rejection, and control effort. The new controller has been tested in MATLAB/Simulink and compared with standard PI controllers. Under step load disturbance, the proposed controller achieves an 88.6% reduction in ITAE compared to conventional PI control. In the presence of sinusoidal load disturbance, the ITAE is further reduced by 94.9%, demonstrating strong disturbance rejection capability. Moreover, under parameter uncertainties, the settling time is improved by 36.8%, while the ITAE is reduced by 56.8%. The results demonstrate improved robustness and faster transient response compared to standard PI control making the proposed controller a superior solution for many applications such as robotic actuators and industrial positioning systems.
Developing a Hybrid Model for Malware Detection Using Artificial Intelligence and the Internet of Things Aseel Hamoud Hamza; Rusul H. Altaie
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15731

Abstract

The widespread adoption of Internet of Things (IoT) devices in smart homes, health care, and Industry 4.0 has brought new security challenges, especially given the growing complexity of malware and botnets threats. Conventional signature-based approaches are not always suitable for IoT device deployments due to device diversity, resource constraints, and the rise of zero-day attacks. The research introduces a multi-faceted approach to malware detection, combining signature-based methods, anomaly detection, and machine learning for better accuracy and timely detection. Data was captured from an IoT testbed comprising smart cameras, sensors and embedded devices. A dataset of 50,000 labeled network flow records was created with Wireshark and Snort, preprocessed, and then features were extracted. A Random Forest classifier was developed and combined with YARA-based signature matching and Z-score behavioral analysis, to create a hybrid detection system. The model was tested on a 7,500-sample test set, as well as in a 48-hour real-time IoT deployment trial. The testing results show that the hybrid system we propose has an accuracy of 97.4%, precision of 95.6%, recall of 96.8%, and an F1-score of 96.2%, with a false positive rate of 2.3%. The real-time test achieved 97% detection rate with an average decision time of 0.85 seconds. The system also achieved 92.1% accuracy with adversarial attacks using modified and new malicious samples. These results demonstrate that hybrid approaches using machine learning, signature analysis and behavioral analysis are effective in improving IoT malware detection. Our lightweight hybrid approach offers a lightweight, scalable and effective solution for IoT devices with limited computational power, and remains robust against emerging threats.
Transformer-Based Semantic Retrieval for Cultural Heritage Question Answering Tri Lathif Mardi Suryanto; Aji Prasetya Wibawa; Hariyono Hariyono; Andrew Nafalski
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15775

Abstract

Cultural heritage knowledge presents significant challenges for Question Answering (QA) systems due to their interpretive, context-dependent, and symbolically rich nature. While Transformer-based models have achieved strong performance in semantic representation, they remain prone to hallucination and contextual misalignment, particularly in culturally sensitive domains. This study proposes a Transformer-based cultural knowledge retrieval framework for domain-specific chatbots, combining a bi-encoder (MiniLM and MPNet) for efficient semantic retrieval and a cross-encoder (BERT-base) for fine-grained reranking. A curated dataset of 4,016 question–answer pairs in Indonesia is developed from cultural heritage sources and validated for contextual consistency. The proposed approach is evaluated using both quantitative and qualitative metrics, including accuracy, F1-score, Exact Match (EM), and semantic-based measures such as F1-BLEU, F1-EDIT, and F1-ANS. Experimental results show that while all models achieve high classification performance (accuracy up to 0.99), the BERT + MPNet configuration significantly outperforms others in answer quality metrics, indicating superior semantic fidelity. However, qualitative analysis reveals persistent issues of hallucination and contextual misalignment, highlighting the limitations of relying solely on statistical evaluation. These findings demonstrate that high numerical performance does not guarantee meaningful understanding in cultural domains. Therefore, this study emphasizes the need for hybrid evaluation frameworks and context-aware mechanisms to ensure epistemic fidelity. The proposed approach contributes to the development of more reliable and culturally grounded QA systems.
A Compact Wideband Microstrip-Fed Planar Slot Antenna with Partial Defected Ground Structure for 2.7–12 GHz Wireless Applications Rami Abduimawjood Mohammed; Uttam Laxman Bombale
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.15913

Abstract

This paper presents the design and numerical investigation of a compact wideband microstrip-fed planar slot antenna for broadband wireless communication applications. The antenna is designed on a low-cost FR-4 substrate with a compact planar configuration, making it suitable for space-constrained and economical wireless devices. In the initial design stage, a reference microstrip-fed antenna is developed and shown to operate from 3.25 to 11.6 GHz under the −10 dB impedance bandwidth criterion. The antenna is then optimized through a systematic trial-and-error parametric refinement process involving modifications of the radiator profile, feed-line geometry, slot configuration, and partial defected ground structure. To further enhance radiation performance, three compact rectangular parasitic directors are arranged in front of the main radiating structure with optimized spacing to improve forward radiation and gain. Full-wave electromagnetic simulations are carried out using Ansys High Frequency Structure Simulator (HFSS). The final optimized antenna achieves a continuous −10 dB impedance bandwidth from 2.7 to 12 GHz, covering the ultra-wideband (UWB) spectrum as well as several sub-6 GHz fourth-generation/fifth-generation (4G/5G) communication bands. The optimized antenna exhibits simulated gains of approximately 5.53 dB, 6.97 dB, and 6.19 dB at 6.7 GHz, 8.7 GHz, and 10.7 GHz, respectively, showing an overall improvement compared with the reference design. The radiation patterns remain reasonably stable across the investigated lower, middle, and upper frequency regions of the operating band, with quasi-omnidirectional characteristics and improved forward radiation due to the director elements. Compared with several reported wideband microstrip antenna designs, the proposed antenna offers a favorable combination of wide impedance bandwidth, compact structure, low-cost substrate realization, and enhanced gain. Therefore, the proposed antenna is a promising candidate for modern broadband wireless systems, including UWB and sub-6 GHz 4G/5G applications.
Grid-Aligned Patchification for Deep Learning-Based Macrophage Detection in Unstained Brightfield Haemocytometer Images Mohammad Ikhsan; Zino Ramdani Suharto; Rizal Azis; Basari Basari
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.16005

Abstract

Manual cell counting from haemocytometer images is slow, subjective, and operator-dependent, especially in unstained brightfield microscopy where cell boundaries and viability-related morphology are difficult to distinguish. Although prior cell detection models have mainly been evaluated on stained or fluorescence images, systematic comparisons between fine-tuned detectors and zero-shot cell segmentation models remain limited for unstained brightfield haemocytometer images. This study presents a controlled 2×2 factorial benchmark of patchification and augmentation across five detection approaches, with variance-decomposition analysis and comparison of fine-tuned versus zero-shot deployment modes. Using 24 unstained brightfield RAW 264.7 macrophage images with 6,307 polygon-level annotations, including 28.8% dead cells, we evaluated four preprocessing scenarios under six-fold stratified cross-validation. Faster R-CNN, Mask R-CNN, and YOLOv11n-Seg were fine-tuned within each fold, whereas Cellpose and StarDist were applied zero-shot. Grid-aligned patchification improved bounding-box mAP50 by 2.6–8.4× across all fine-tuned architectures (paired Wilcoxon p = 0.016, Cohen’s d > 3). A 2×2 ANOVA attributed 99.2–99.4% of explained variance to patchification, while augmentation and interaction effects each contributed less than 0.1%, suggesting that performance gains were driven mainly by scale rescaling rather than sample count. On patchified data, fine-tuned models converged to 85.5–86.4% mAP50. YOLOv11n-Seg achieved the highest mAP50-95 of 51.1%, with 6× faster inference and 17× fewer parameters. In contrast, zero-shot Cellpose and StarDist reached only 45.3–51.2% class-agnostic F1@0.5. These findings show that structure-aware patchification is critical for reliable cell detection in this modality.
Machine Learning Approaches for Binary Classification of Portion Size and Cooking Time in Indonesian Recipes Devi Dwi Purwanto; Aji Prasetya Wibawa; Mazarina Devi
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.16013

Abstract

Estimating portion sizes and cooking times are goals for smart kitchen assistants, enabling better meal planning and reducing food waste due to over-portioning. Existing approaches in computational gastronomy often struggle to provide estimates from prepared ingredient data. This study uses XGBoost to extract features from a dataset containing 1,400 Indonesian recipes to predict binary classification targets for portion sizes and required cooking times. The dataset used for the prediction includes information on ingredients and their quantities, as well as preparation steps. In addition to the recipe dataset, the TKPI dataset is also used to help determine the category of food ingredients, protein content, and cooking technique complexity. This dataset is then further optimized with hyperparameters to maximize model performance. This paper conducted trials with 6 models where the best model for portion size had an accuracy of 0.7821 with a balanced accuracy of 0.4929, and an F1 Score of 0.8763, while the accuracy for cooking time was 0.6929 with a balanced accuracy of 0.6445, and an F1 Score of 0.7737. From the best model, it was found that the quantity of weighted ingredients and the distribution of ingredients per step were among the most influential features, while step-based and technique-based features were the most important features for cooking time. The contribution of this research is the development of an interpretable model for meal planning efficiency in culinary applications. These results indicate that feature aggregation combined with XGBoost provides actionable insights for smart kitchen assistants and recommendation systems.
Intrusion Detection System: A Multimodal Analysis-based Machine Learning with Emphasis on Interpretability Tabark Nasser Abdul Hussein; Ayad Hameed Mousa
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 3 (2026): June
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i3.16024

Abstract

Detecting persistent threats, whether APTs or nonAPTs, is a constant challenge in the field of cybersecurity due to the multiplicity of attacks, their stealthy nature, and their multi-stage targeting of information systems over extended periods. The rigor of intrusion detection system selection is measured by its ability to detect these threats in their early stages and by the fundamental characteristics of network traffic. However, due to the large number of characteristics, some may be unrelated or of limited importance in determining the severity of malicious activity. Accordingly, selecting and defining relevant and influential characteristics for intrusion detection has become a necessity, especially in resource-constrained environments. In this paper, a set of machine learning algorithms (XGBoost, Random Forest, Support Vector Machine, Hybrid Decision Tree) was adopted in conjunction with artificial intelligence pre-interpretation (XAI) techniques to develop an intrusion detection model in a resource-constrained environment. The datasets CICAPT-IIoT, CICIoT2023, IoT-23 were used after preprocessing. XAI techniques were employed in two phases: first, during preprocessing to identify key features of the selected datasets, and second, during post-processing for interpretation. A real-world application based on the proposed model was developed to validate its accuracy and applicability in intrusion detection. Extensive testing demonstrated the superiority of the (XGBoost) algorithm with its accuracy (the CICIoT2023 dataset, achieving an F1-score of 0.9925, precision of 0.9933, recall of 0.9920, and an almost perfect ROC-AUC of 0.9999. the CICAPT-IIoT dataset scoring 99.16%, 0.9810 F1 score. And on the IoT-23 dataset accuracy 99.69% and achieved balanced precision, recall and F1-score of 0.9969). The study was distinguished by its reduction of complexity and improved performance of the proposed model.

Page 34 of 36 | Total Record : 351

 32   33   34   35   36 

Filter by Year

2019 2026


Reset
Filter By Issues
All Issue Vol. 8 No. 4 (2026): August Vol. 8 No. 3 (2026): June Vol. 8 No. 2 (2026): April Vol. 8 No. 1 (2026): February Vol. 7 No. 4 (2025): December Vol. 7 No. 3 (2025): September Vol. 7 No. 2 (2025): June Vol. 7 No. 1 (2025): March Vol. 6 No. 4 (2024): December Vol. 6 No. 3 (2024): September Vol. 6 No. 2 (2024): June Vol. 6 No. 1 (2024): March Vol. 5 No. 4 (2023): December Vol. 5 No. 3 (2023): September Vol. 5 No. 2 (2023): June Vol. 5 No. 1 (2023): March Vol. 4 No. 3 (2022): December Vol. 4 No. 2 (2022): Agustus Vol. 4 No. 1 (2022): April Vol. 3 No. 3 (2021): Desember Vol 3, No 2 (2021): Agustus Vol. 3 No. 2 (2021): Agustus Vol 3, No 1 (2021): April Vol. 3 No. 1 (2021): April Vol. 2 No. 3 (2020): Desember Vol 2, No 3 (2020): Desember Vol 2, No 2 (2020): Agustus Vol. 2 No. 2 (2020): Agustus Vol. 2 No. 1 (2020): April Vol 2, No 1 (2020): April Vol 1, No 3 (2019): Desember Vol. 1 No. 3 (2019): Desember Vol 1, No 2 (2019): Agustus Vol. 1 No. 2 (2019): Agustus Vol 1, No 1 (2019): April Vol. 1 No. 1 (2019): April More Issue