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

30,907 Article Results

Heading stabilization of a mecanum wheel mobile robot using Kalman filter and SMC under variation condition

10.12928/telkomnika.v24i4.27654
Ardianto; Politeknik Negeri Jember Syaifur Rohman , Tunjung; Politeknik Negeri Jember Genarsih , Sihmaulana; Politeknik Negeri Jember Dwianto , Nuzula; Politeknik Negeri Jember Afianah , Salsabila; Politeknik Negeri Jember Liandra Putri , Nurul; Politeknik Negeri Jember Zainal Fanani , Mochamad; Politeknik Negeri Jember Irwan Nari , Ahmad; Politeknik Negeri Jember Rofi'i , Fendik; Politeknik Negeri Jember Eko Purnomo , Syamsiar; Politeknik Negeri Jember Kautsar , Angga; Politeknik Negeri Jember Dwinanda , Alfan; Politeknik Negeri Jember Ahmad Berlian
This paper presents a robust heading stabilization system for a mecanum wheel mobile robot by integrating a Kalman filter (KF) with sliding mode control (SMC). A two-state KF estimates the robot’s heading angle and gyroscope bias from ICM20948 inertial measurement unit (IMU) measurements, reducing sensor noise by 65% and compensating for bias drift of 0.3° per second. The estimated heading is regulated using SMC with a boundary layer to minimize chattering. Implemented on a Raspberry Pi 3B, the system was validated under varying surface friction conditions and external disturbances. The controller achieved heading stabilization with root mean square error (RMSE) between 0.380° and 0.589° across all surfaces, steady-state error within ±0.5°, and convergence within 2.0–2.6 seconds. Under severe disturbances causing heading deviations up to 238°, rapid recovery within 0.5 seconds was achieved with only 3.04° final steady state error. The results demonstrate the feasibility of implementing robust heading stabilization on low-cost embedded platforms for autonomous navigation applications.
Volume: 24
Issue: 4
Page: 1396-1408
Publish at: 2026-08-01

Semantic-aligned multimodal human activity recognition using visual and audio data

10.11591/ijece.v16i4.pp2087-2095
Yeeun Park , Junhoo Byun , Siwoo Byun
Human activity recognition (HAR) requires robust performance under heterogeneous sensing conditions for practical deployment. However, single-modality approaches are limited in capturing the rich contextual information inherent in complex human behaviors. This paper presents a semantic-aligned multimodal HAR framework that integrates visual and audio information without assuming instance-level synchronization. To address dataset heterogeneity, samples from the HMDB51 video dataset and the ESC-50 audio dataset are aligned by mapping fine-grained classes into a shared high-level activity label space. For each modality, ResNet-18-based models are trained independently using frame-based visual inputs and 64-bin Mel-spectrogram-based audio representations. During inference, the output logits of the two models are combined through score-level weighted linear fusion. Experimental results show that the proposed multimodal approach consistently outperforms unimodal baselines in terms of accuracy and Macro-F1 score, with particularly notable improvements in activity groups where environmental context plays a significant role. These findings indicate that semantic-aligned score-level fusion can improve recognition robustness even under mismatched dataset conditions.
Volume: 16
Issue: 4
Page: 2087-2095
Publish at: 2026-08-01

A convolutional neural network -based driver monitoring system for drowsiness and distraction detection

10.11591/ijece.v16i4.pp2220-2229
Sara Benkouider , Nasreddine Lagraa
Road accidents caused by driver drowsiness and distraction are a major global concern, as fatigue and inattention significantly slow reaction time and increase accident risk. To address this issue, this paper proposes a vision-based driver monitoring system using facial analysis from an on-board camera. The system detects the face and extracts key regions of interest, including the eyes, mouth, and head, which are analyzed independently using convolutional neural networks. Temporal information is captured by aggregating the convolutional neural network (CNN) outputs over a fixed time window. Drowsiness is estimated by fusing eye and mouth features with a multilayer perceptron, while distraction is detected based on head movements. An important advantage of the proposed approach is its robustness to partial input loss, allowing the system to remain functional even when some facial regions are missing or occluded, such as when wearing sunglasses or face masks. Experimental results show high detection accuracy, reaching 97.3% for drowsiness and 98% for distraction under ideal conditions, with only limited performance degradation in challenging scenarios. These results confirm the suitability of the proposed system for real-time driver monitoring applications.
Volume: 16
Issue: 4
Page: 2220-2229
Publish at: 2026-08-01

Enhancement of YOLOv8 for object detection in adverse weather conditions using generative adversarial network

10.11591/ijece.v16i4.pp2230-2246
Talifhani Calvin Tshipota , Chunling Tu , Mukatshung Claude Nawej , Sempe Thom Leholo
Detecting objects in bad weather like rain, fog, snow, or low light is still difficult because visibility drops, noise increases, and contrast gets worse, all of which hurt detection accuracy. Most current methods either improve detector designs or use image preprocessing on their own. They usually focus on just one type of weather and do not use a common way to evaluate results. This paper introduces a YOLOv8 framework improved with a generative adversarial network (GAN) for image enhancement before detection. Instead of just making images look better, the GAN is trained to help the object detector work better in tough conditions, so it can extract features more effectively when images are degraded. The model was tested on datasets with different weather conditions using standard metrics like Precision, Recall, F1-score, and mean average precision (mAP). Results show that this method consistently improves performance, with up to a 6.5% increase in mAP@0.5 over YOLOv8-STE and 9.2% over IA-YOLO, especially in foggy and low-light situations. These results show that adding GAN-based preprocessing to YOLOv8 makes detection more reliable and still keeps the process fast. This framework offers a practical and scalable solution for real-world uses like self-driving cars, smart transportation, and surveillance.
Volume: 16
Issue: 4
Page: 2230-2246
Publish at: 2026-08-01

Intelligent routing-based attack detection in Internet of Things networks using artificial intelligence

10.11591/ijece.v16i4.pp2169-2181
Huda Saloom Sultan , Asseel Jabbar Almahdi , Murteza Hanoon Tuama , Athar Hussein Mohammed
The fast-growing Internet of Things (IoT) networks have posed considerable security risks because of decentralized network designs, dynamic topologies, and inadequate computation capabilities. Current intrusion detection strategies are primarily traffic-based, but without paying attention to routing-layer dynamics, which are paramount in multi-hop IoT systems. To overcome this drawback, this paper suggests a smart routing-conscious attack detection model which combines routing-layer monitoring with methods of artificial intelligence to improve the security of IoT networks. The suggested framework constantly compares routing metrics, such as packet loss, change in hop count, end to end delay and energy consumption to detect malicious routing behavior in real time. Two types of artificial neural networks, feedforward neural network (FFNN) and convolutional neural network (CNN) are used to categorize routing activities as normal or malicious. The experimentation on simulation was carried out by using NS-2 in a dynamic multi-hop IoT environment where routing-based DoS attacks were implemented. The experimental results reveal that CNN model had a higher detection accuracy of 85.76% with lower execution time of 17 s compared to the FFNN model which had an accuracy of 82.76% and an execution time of 18 s. Moreover, the suggested framework enhanced reliability of routing by minimizing the packet loss and communication delay and having low routing overhead. These results support the hypothesis that routing-aware intelligence can be used to enhance AI-based intrusion detection to create an adaptive, routing-aware, and resource-efficient security solution to decentralized networks of IoT devices.
Volume: 16
Issue: 4
Page: 2169-2181
Publish at: 2026-08-01

Physics-based modeling of cobalt-doped nickel-zinc on-chip ferrite inductors

10.11591/ijece.v16i4.pp1805-1816
Bambang Mulyo Raharjo , Dicky Rezky Munazat , Sudirman Rohadi
The miniaturization of integrated voltage regulators (IVRs) for multi-core processors is fundamentally bottlenecked by the high-frequency magnetic losses of conventional inductor cores. This study presents a rigorous computational framework to optimize Cobalt-doped Nickel-Zinc ferrite (Ni_(1-x) Zn_0.4 Co_x Fe_2 O_4) for 10 MHz on-chip power delivery. Utilizing Landau-Lifshitz-Gilbert (LLG) relaxation dynamics and Maxwell-Wagner interfacial polarization, the complex electromagnetic dispersion was modeled and quantitatively validated against recent empirical literature. Furthermore, high-temperature power loss density was bounded using the trust region reflective (TRF) numerical curve fitting algorithm to validate an anisotropy-compensated "thermal valley" at 80 °C. A multi-objective sensitivity analysis identified a high-efficiency Cobalt "sweet spot" at a concentration of x=0.04. This specific formulation optimally stiffens domain walls, safely shifting the resonance frequency to 40 MHz and maximizing the quality factor (Q) at the 10 MHz operational target. When applied to a simulated 3.5 V to 1.0 V DC-DC buck converter, the optimized x=0.04 core demonstrated its adequacy for on-chip applications by maintaining >90% power efficiency under a rigorous 2.0 A load. These predictive results mathematically prove that precision Cobalt doping is a highly viable strategy for suppressing parasitic losses in next-generation 3D-IC power delivery networks.
Volume: 16
Issue: 4
Page: 1805-1816
Publish at: 2026-08-01

From data to intelligence: foundations of learning systems, representation, and computational perception

10.11591/ijece.v16i4.pp1669-1676
Tole Sutikno
The rapid evolution of intelligent systems has shifted the focus of electrical and computer engineering from isolated data processing toward integrated models of machine cognition. This editorial introduces a foundational perspective on machine intelligence systems, emphasizing the transformation from raw data to meaningful intelligence through learning systems, representation mechanisms, and computational perception. In contemporary AI-driven environments, intelligence is no longer defined solely by algorithmic performance, but by the ability to construct structured representations of the world and interpret complex multimodal signals. Learning systems, particularly those grounded in machine learning and deep learning paradigms, serve as the core mechanism enabling this transformation. Representation learning provides the bridge between unstructured data and abstract knowledge, while computational perception enables machines to interpret visual, auditory, and sensor-based information in real time. Together, these components form the foundational architecture of intelligent systems that underpin emerging applications in engineering, automation, and cyber-physical environments. This editorial sets the stage for understanding intelligence as an emergent computational construct, highlighting its role as the first phase in the broader cognitive intelligence systems continuum that progresses toward adaptive, autonomous, and socio-cognitive systems in future research directions.
Volume: 16
Issue: 4
Page: 1669-1676
Publish at: 2026-08-01

Beyond adoption: measuring the success of mandatory information systems through an integrated ECM and ISSM

10.11591/ijece.v16i4.pp2031-2041
Muhammad Rosyid Ridlo , Muhammad Fachri Shandika Iman , Reny Yuliati
The successful implementation of mandatory organizational information systems depends not only on system adoption but also on user satisfaction. However, most post-adoption evaluation studies have focused on voluntary systems, leaving mandatory public sector deployments substantially underexplored. This study evaluates the determinants of employee satisfaction with the Coretax Administration System, a nationwide integrated tax platform implemented by the Directorate General of Taxes in Indonesia. To provide a comprehensive explanation of post-adoption evaluation, this research integrates the Expectation Confirmation Model (ECM) and the Information System Success Model (ISSM), examining how system quality, information quality, and service quality influence confirmation and perceived usefulness, which in turn determine user satisfaction. Using a quantitative approach, data were collected from 292 employees actively using the system and analyzed through Partial Least Square Structural Equation Modeling (PLS-SEM). The results demonstrate that the integrated model exhibits strong predictive power, explaining 76.4% of the variance in user satisfaction. System quality emerged as the most influential determinant, significantly affecting confirmation and perceived usefulness, which subsequently drives satisfaction. Meanwhile, information quality and service quality showed selective effects, indicating that technical reliability plays a more critical role than supportive features in a mandatory environment. The findings offer actionable guidance for policymakers and IS architects engaged in large-scale compulsory digital transformation initiatives in the public sector.
Volume: 16
Issue: 4
Page: 2031-2041
Publish at: 2026-08-01

Interpretable machine learning for digital soil pH mapping using an optimized AdaBoost algorithm

10.12928/telkomnika.v24i4.27724
Zakiyyan; Universitas Jenderal Soedirman Zain Alkaf , A’isya Nur; Universitas Jenderal Soedirman Aulia Yusuf , Elsa Sari; Universitas Jenderal Soedirman Hayunah Nurdiniyah , Tri; Universitas Jenderal Soedirman Wisudawati
Soil pH is a fundamental parameter determining nutrient availability, microbial activity, and crop productivity. Unlike previous studies that often prioritize prediction accuracy over explainability, this study proposes an interpretable machine-learning framework integrating hyperparameter optimized adaptive boosting (AdaBoost) with Shapley Additive exPlanations (SHAP) to unravel the spatial drivers of soil pH. A systematic workflow was implemented to evaluate a diverse set of algorithms, followed by Bayesian optimization to fine-tune the best-performing models. The results demonstrated that the optimized AdaBoost model yielded the largest performance improvement (~7.5%), achieving excellent accuracy on independent test data with a coefficient of determination (R²) of 0.817 and a mean absolute error (MAE) of 0.293. Furthermore, SHAP analysis identified iron (Fe) and calcium carbonate (CaCO₃) as the most influential predictors, revealing that Fe exhibits a strong inverse relationship with pH, while CaCO₃ shows a positive association. This framework successfully balances high predictive accuracy with pedological interpretability, offering a robust tool for digital soil mapping and precision agriculture.
Volume: 24
Issue: 4
Page: 1197-1203
Publish at: 2026-08-01

Extended-state observer control with online payload identification for HRI in series elastic actuators

10.12928/telkomnika.v24i4.27787
Edwin; Fundación Universitaria Los Libertadores Villarreal-López , Horacio; Universidad de San Buenaventura Bogotá Coral-Enriquez , Luini L.; Universidad Distrital Francisco José de Caldas Hurtado-Cortés
Safe human–robot interaction requires actuators that combine compliance with accurate force inference. Series elastic actuators (SEAs) are particularly suitable for this purpose; however, separating payload-induced torques from voluntary human interaction forces without dedicated sensors remains a critical challenge. This paper proposes a unified observer-based control framework for SEAs that integrates three key components: a linear parameter-varying extended state ob server (LPV-ESO), an online algebraic payload estimator, and a disturbance driven motion intention identification (MII) mechanism. The payload estimator continuously updates the observer gains to adapt to load variations, while the LPV-ESO decouples payload dynamics from human interaction forces. The dis turbance estimate is then processed through the MII block to infer operator intent in real time without additional sensing hardware. Numerical simulations under nominal conditions and ±10% parametric uncertainty confirm bounded trajec tory tracking error, rapid convergence of payload estimation, and effective sepa ration of payload-induced torques from voluntary interaction forces. The results demonstrate that the proposed framework achieves sensorless force decompo sition and reliable motion intention inference, offering a practical solution for SEA-based collaborative robots in rehabilitation, assistance, and industrial ap plications.
Volume: 24
Issue: 4
Page: 1409-1418
Publish at: 2026-08-01

Real-time anomaly detection system using best performed machine learning model

10.12928/telkomnika.v24i4.27561
Victor; North-West University Mathebula , Bukohwo; North-West University Michael Esiefarienrhe
Effective anomaly detection is critical for protecting organizational networks against increasingly sophisticated cyber threats. However, most machine learning-based intrusion detection models are developed and validated using public benchmark datasets, which may not reflect the operational characteristics, traffic behavior, and threat patterns of real institutional networks. In the case of Umalusi, there is currently no anomaly detection model customized and validated using Umalusi-specific network traffic, creating a practical gap in deployable cybersecurity capability. This study proposes a hybrid machine learning framework tailored to support accurate, efficient, and operationally relevant anomaly detection. Using knowledge discovery in databases (KDD) process, network traffic data were collected and pre-processed through normalization, label encoding, missing value treatment, and dimensionality reduction using principal component analysis (PCA). The 16 hybrid models integrating unsupervised anomaly detection with supervised classification were implemented and comparatively evaluated. Experimental findings indicate that the density-based spatial clustering of applications databasescan (DBSCAN) + random forest (RF) model achieved 99.92% accuracy while maintaining a low false positive (FP) cost, making it suitable for a security operations centre (SOC). In addition, a Flask-based web application was developed to enable real-time deployment by sniffing live network traffic, executing inference, and persisting results in an SQLite database.
Volume: 24
Issue: 4
Page: 1204-1215
Publish at: 2026-08-01

Sentiment analysis in telecommunications: a systematic review of applications and challenges

10.12928/telkomnika.v24i4.27762
Achraf; LaGeS Laboratory, Hassania School of Public Works Bouhamidi , Naziha; LaGeS Laboratory, Hassania School of Public Works Laaz , Zineb; LaGeS Laboratory, Hassania School of Public Works Rachik
The increasing amount of digital content on online platforms creates both opportunities and challenges for telecommunications operators aiming to improve customer experience. Sentiment analysis (SA), a subfield of natural language processing (NLP), allows for the extraction of subjective information from data generated by users. This paper presents a systematic literature review (SLR), conducted in accordance with Kitchenham’s guidelines, examining the applications and challenges of SA within the telecommunications sector. A structured methodology was used to select and analyze over 100 studies published between 2017 and 2025. The review categorizes SA methods and evaluates their application in practical telecom contexts, including customer satisfaction assessment, churn prediction, service monitoring, and reputation management. Findings indicate that machine learning and deep learning models can achieve performance levels up to 97% in specific experimental contexts. However, such results may vary depending on the datasets, evaluation protocols, and application contexts. Key challenges identified include handling informal language, domain dependency, multilingualism, sarcasm detection. Finally, the review highlights future directions such as real-time sentiment tracking, multimodal analysis, and the integration of federated learning for privacy-preserving customer analytics. This review provides a foundational reference for researchers and practitioners aiming to deploy effective sentiment-driven systems in the telecom industry.
Volume: 24
Issue: 4
Page: 1224-1240
Publish at: 2026-08-01

From climate time series to planting windows in chili (Capsicum frutescens): a SARIMA–SVM–XGBoost framework with balanced-accuracy thresholding

10.11591/ijece.v16i4.pp2210-2219
Efrans Christian , Nova Noor Kamala Sari , Ressa Priskila , Septian Geges
This study proposes a spatio-temporal decision-support framework that integrates Seasonal Autoregressive Integrated Moving Average (SARIMA), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) to derive adaptive planting windows for chili (Capsicum frutescens) at the sub-district level. The framework addresses key challenges in climate-sensitive agriculture, including spatial data leakage and class imbalance, by employing Leave-One-Group-Out (LOGO) cross-validation and Balanced Accuracy–based threshold optimization. The proposed system transforms heterogeneous environmental data into actionable recommendations by combining climate forecasting, land suitability assessment, and yield prediction within a unified pipeline. Experimental results indicate that the framework effectively captures seasonal climate dynamics and produces consistent planting recommendations aligned with agronomic conditions, enabling multiple planting cycles per year. The primary contribution of this work lies in a transparent and generalizable integration of statistical and machine learning models into a practical decision-support framework. The proposed approach bridges predictive modeling and real-world agricultural decision-making and can be extended to other crops and regions for climate-adaptive agricultural planning.
Volume: 16
Issue: 4
Page: 2210-2219
Publish at: 2026-08-01

Indirect adaptive neural network control for constant power conversion in wave energy system

10.11591/ijece.v16i4.pp2120-2133
Jesus de la Cruz-Alejo , Hugo Beatriz Cuellar , J. Antonio Lobato Cadena , Edwin Christian Becerra-Alvarez
The conversion of ocean wave energy into electrical energy occurs near beaches and is important for the design and implementation of wave energy conversion (WEC) systems. However, its generation depends on environmental conditions, which complicates the design and control of the devices. This work presents an approach to indirect adaptive control based on artificial neural networks to detect wave conditions for the proper functioning of WEC structures. The method involves generating a constant output voltage using a voltage boost converter and a direct current-alternating current (DC-AC) converter. Maintaining a constant output power despite variations in wave conditions to generate a voltage of 24 V with a current of 2 A is the primary proposal for the control design. The mechanical design integrates a rack and pinion system and a pulley transmission that connects a floating device to an electric generator. The implementation of control is carried out on an Arduino platform. The control system was implemented on an Arduino platform, occupying 48% of the available memory, with a convergence time of 4.29 ms, a mean squared error (MSE) of 0.13715, and a root mean squared error (RMSE) of 0.37034. These low values indicate that the proposed control system has greater accuracy. The experimental results validate the proposed control system, which reduces energy conversion errors and achieves greater efficiency.
Volume: 16
Issue: 4
Page: 2120-2133
Publish at: 2026-08-01

Study of the effect of changes in cell output temperature on the net efficiency of solid oxide fuel cell power generation

10.12928/telkomnika.v24i4.27828
Handrea; PT PLN (Persero) Bernando Tambunan , Ignatius; Politeknik Negeri Bandung Riyadi Mardiyanto , Lina; Politeknik Negeri Bandung Troskialina , Sri; Politeknik Negeri Bandung Paryanto Mursid , Lidya; Politeknik Negeri Bandung Elizabeth , Dhyna; Politeknik Negeri Bandung Analyes Trirahayu , Retno; Politeknik Negeri Bandung Dwijayanti
Solid oxide fuel cell (SOFC) power plants offer a promising pathway toward highly efficient energy conversion compared to conventional systems. This study explores the impact of output temperature variations on system performance through simulation analysis. The results demonstrate that increasing the temperature difference between the fixed input (700 °C) and the cell output significantly enhances net efficiency, while gross efficiency remains relatively stable. At an output temperature of 875 °C, the system achieves a net efficiency approaching 50% and a gross efficiency of approximately 61%. These findings emphasize the critical role of output temperature in determining overall system efficiency and highlight the importance of thermal management in SOFC operation.
Volume: 24
Issue: 4
Page: 1419-1426
Publish at: 2026-08-01
1 2 3
...
2060 2061
Show 7 of 2061

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