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

Edge-aware coffee aroma classification using multi-representation feature extraction and LightGBM on Jetson Nano

10.11591/ijece.v16i4.pp2096-2105
Denda Dewatama , Erni Yudaningtyas , Muhammad Fauzan Edy Purnomo , Setyawan Purnomo Sakti
Objective coffee aroma evaluation remains challenging outside controlled laboratory settings, and most electronic nose studies neglect embedded deployment constraints. This work proposes an edge-aware coffee aroma classification framework that integrates multi-representation feature extraction with LightGBM and evaluates both predictive performance and computational efficiency. A six-sensor metal-oxide semiconductor (MOS) e-nose was developed, producing a balanced dataset of 1,080 trials from 12 aroma classes. Five feature representations were investigated, including baseline signals, autoencoder embeddings, and convolutional features derived from pseudo-image transformation. Experiments on an NVIDIA Jetson Nano using stratified five-fold cross-validation showed that residual-based representations significantly improved performance. The lightweight residual network achieved an accuracy of 0.9972 with low training time and memory usage. Pareto analysis confirms that optimal performance is achieved by balancing accuracy and resource constraints, thereby enabling reliable deployment in edge and IoT environments.
Volume: 16
Issue: 4
Page: 2096-2105
Publish at: 2026-08-01

Scenario-driven fault injection for realistic bugs in web application testing

10.11591/ijece.v16i4.pp2014-2030
Asri Maspupah , Joe Lian Min , Yadhi Aditya
Conventional mutation-based fault injection techniques generally produce single-line syntactic faults, which often fail to represent realistic errors at the functional requirement level because requirement context, execution paths, and functional dependencies are not considered. To address this limitation, this study proposes scenario-driven fault injection (SDFI), a scenario-based fault insertion approach that derives faults from functional requirements and test cases. SDFI integrates operational fault localization, web fault taxonomy, fault injection patterns, and functional scenario mapping to produce targeted fault injections at relevant code locations, resulting in a realistic bug dataset with multi-line faults. An experimental evaluation on a real web application produced 29 mutants, achieving a fault detection rate of 89.29% based on the RIP model. Further analysis shows that the generated mutants replicate common real-world bug characteristics, including logic errors, validation anomalies, inter-function data propagation, and multi-line faults affecting client–server application behavior. These results demonstrate that SDFI is effective in producing realistic bug datasets for evaluating software testing quality, improving test case effectiveness, and supporting further research on requirement-based fault realism.
Volume: 16
Issue: 4
Page: 2014-2030
Publish at: 2026-08-01

The adoption of artificial intelligence in government technology in developing countries: a systematic literature review

10.11591/ijece.v16i4.pp2192-2209
Maria Florentina Rumba , Flourensia Sapty Rahayu
This study systematically reviews empirical evidence on Artificial Intelligence adoption in Government Technology across developing countries, addressing three research questions regarding determinant factors for success, strategies to overcome absorptive capacity constraints, and governance frameworks for responsible implementation. To answer these questions, a systematic literature review was carried out using a structured methodology: The Scopus database was searched with a comprehensive keyword strategy, which initially yielded ninety-three publications. Through rigorous screening and eligibility assessment, fourteen articles meeting all inclusion criteria were analyzed thematically. The findings demonstrate that successful AI adoption requires holistic alignment among institutional factors including policy frameworks and public trust, organizational elements such as human resource capacity and bureaucratic culture, and environmental conditions encompassing infrastructure and societal demands. Effective strategies identified include strengthening digital infrastructure, developing human capital, implementing adaptive governance, fostering multi-stakeholder collaboration, and ensuring contextual adaptation. The study further reveals that legitimate and sustainable implementation necessitates integrating collaborative governance as a legitimacy foundation, Governance, Risk, and Compliance (GRC) frameworks as ethical control systems, and Explainable AI as a transparency mechanism. This integrated approach enables AI to generate both operational efficiency and strategic public value, including social inclusion and progress toward sustainable development objectives.
Volume: 16
Issue: 4
Page: 2192-2209
Publish at: 2026-08-01

GenAI as an IoT programming assistant: a case study on automated debugging for air quality monitoring systems

10.12928/telkomnika.v24i4.27707
Steven; Pradita University Imanel Bawole , Handri; Pradita University Santoso
The rapid expansion of internet of things (IoT) technology has necessitated the development of user-friendly programming solutions for non–experts. While generative artificial intelligence (GenAI) offers the potential to democratize code development, its ability to assist in the intricate task of automated debugging, particularly regarding hardware integration remains a critical area of research. A design research approach was employed, employing a structured four – phase workflow: error analysis, diagnostic execution through prompting, iterative solution analysis, and functional verification. The methodology was applied to an experimental case study involving an air quality (AQ) monitoring system. The study tested the artificial intelligence (AI)’s capacity to debug C++ code intended for the Arduino integrated development environment (IDE). Gemini AI successfully identified and resolved three critical logic errors arising from mismanaged MQ135 calibration variables, incorrect loop sequencing, and data desynchronization between the organic light emitting diode (OLED) display and the internal status logic. GenAI proved effective as a programming assistant for resolving bugs in IoT applications. However, effective debugging still depends on well-structured prompts and a basic understanding of the underlying IoT hardware.
Volume: 24
Issue: 4
Page: 1278-1286
Publish at: 2026-08-01

Multiobjective framework for congestion management through coordinated scheduling of generation and demand

10.11591/ijece.v16i4.pp1688-1703
Jayesh Priolkar , Govind Kunkolienkar
The safe and secure operation of power system networks remains a significant challenge due to the ever-increasing demand for electrical energy. In deregulated environments, there is a strong emphasis on the optimal and efficient utilization of existing resources. This work aims to address line congestion by optimally re-dispatching generation resources and proactively managing demand through advanced demand response (DR) programs. An elasticity based, multi-period load model is employed to enhance the realism and effectiveness of DR strategies. The novelty of the proposed work is the holistic approach that simultaneously addresses economic, environmental, and technical objectives, incorporating realistic DR behavior and the advanced modified elephant herding optimization (MEHO) technique. This work proposes a MEHO algorithm for multi-objective congestion management with coordinated generation and DR programs, with comparative analysis against MPSO on both IEEE 30-bus and IEEE 118-bus systems. The MEHO algorithm generates seven unique Pareto-optimal solutions that represent various trade-offs between the conflicting objectives, demonstrating the implementation's remarkable performance on the IEEE 30-bus and IEEE 118-bus test system. MEHO achieves 3.5 to 6.2% better cost solutions, 2.3 to 5.2% lower emissions, and 60 to 62.5% better congestion indices across both test systems.
Volume: 16
Issue: 4
Page: 1688-1703
Publish at: 2026-08-01

A systematic review and conceptual roadmap for sky computing: AI-enabled orchestration, interoperability, and governance beyond multi-cloud

10.11591/ijece.v16i4.pp2247-2253
Abdullah Al-Bakri
The concept of sky computing is becoming known as an industry-independent “cloud of clouds” approach intended to address the issues of fragmentation inherent in today’s multi-cloud and hybrid cloud implementations. The objective of this systematic literature review is to examine the challenge of existing multi-cloud architectures, which despite delivering high reliability and purchasing flexibility, suffer from API heterogeneity, fragmented intercloud orchestration, insufficient workload mobility, and unresolved sovereignty and compliance challenges. Based on the PRISMA methodology, 325 papers released between January 2020 and June 2025 have been systematically selected in five scientific databases: ACM Digital Library, IEEE Xplore, SpringerLink, ScienceDirect, and Scopus. Upon applying a process of elimination for duplicates, title-and-abstract screening, full-text evaluation, and quality assessment, a total of 35 peer-reviewed publications from the same timeframe have been thematically analyzed. Four thematic areas were examined: architectural architecture, intercloud orchestration, automation through artificial intelligence/machine learning (AI/ML), and security, privacy, and compliance. Not a single article predating the year 2020 was part of the final systematic literature review (SLR) database or bibliography. The results reveal that compatibility layers and intercloud brokers increase portability of workloads; scheduling based on artificial intelligence is useful in automating operations and achieving optimal cost performance; while zero trust architecture, self-sovereign identity, confidential computing, and policy-driven compliance are key in achieving trustworthy cross jurisdictional operations. The two major conclusions that arise from this study are: Firstly, future studies need to investigate explainability and energy-efficient AI orchestration in a real-world setting with multiple cloud providers, whereas secondly, cloud computing professionals need to incorporate principles of privacy, sovereignty, and compliance directly into the orchestration policies, not as an afterthought.
Volume: 16
Issue: 4
Page: 2247-2253
Publish at: 2026-08-01

Design and evaluation of a simple load balancing prototype using the round robin algorithm in local networks

10.12928/telkomnika.v24i4.27560
Muh. Fahmi; Universitas Sulawesi Barat Rustan , Wawan; Universitas Sulawesi Barat Firgiawan , Wiwi; Universitas Sulawesi Barat Nopiana
Load balancing plays a crucial role in ensuring efficient workload distribution and maintaining stable performance in web service systems. This study presents the design and experimental evaluation of a round robin–based load balancing system implemented in a multi-client local area network (LAN) environment using NGINX as a centralized controller. The system consists of three physical machines, comprising one load balancer and two backend servers hosting identical web applications. Multiple clients generate simultaneous hypertext transfer protocol (HTTP) requests, which are distributed alternately to the backend servers using the default round robin mechanism provided by NGINX. Experimental evaluation was conducted under three workload scenarios of 50, 100, and 200 concurrent requests. The results show that the round robin algorithm consistently distributes requests evenly between the backend servers. The average response time increased from approximately 110 ms at 50 requests to 165 ms at 100 requests and 290 ms at 200 requests, indicating stable performance under light to moderate load conditions. These findings demonstrate that the proposed system is lightweight, modular, and easy to deploy in resource limited environments. The implementation is particularly suitable for campus-scale networks and small institutional settings, serving as a practical platform for local server deployment, academic applications, and experimental learning in networking and distributed systems.
Volume: 24
Issue: 4
Page: 1121-1130
Publish at: 2026-08-01

Classification of P300 event-related potentials using SNN, CNN and LSTM deep learning models

10.12928/telkomnika.v24i4.27659
Ahlaam; Bright Star University M. Saed , Ibtihal; College of Electrical and Electronics Technology Fawzi Elshami , Ali; University of Benghazi I. Elgayar
Accurate classification of P300 event-related potentials remains challenging due to the complex, non-stationary, and low signal-to-noise characteristics of electroencephalography (EEG) signals in brain-computer interface (BCI) systems. P300-based devices, such as the P300 speller, enable communication for patients with severe motor impairments, including those with locked-in syndrome; however, reliable brain signal classification is still a critical limitation. This study presents a comparative evaluation of deep learning models, including convolutional neural networks (CNN), long short-term memory (LSTM) networks, and spiking neural networks (SNN), for P300 signal classification. SNNs represent a biologically inspired paradigm that models the discrete, time-dependent behavior of neural spiking activity and offers advantages in terms of energy efficiency and hardware implementability. Experimental results demonstrate that CNN achieved the highest average classification accuracy (81.04%), followed closely by SNN (80.94%) and LSTM (80.60%). Although CNN slightly outperformed the other models, SNNs showed comparable accuracy while requiring fewer training samples and offering potential benefits for low power and real-time BCI systems. These findings highlight the trade-offs between classification performance and computational efficiency and underline the promise of SNNs as an efficient alternative for P300-based BCI applications.
Volume: 24
Issue: 4
Page: 1294-1306
Publish at: 2026-08-01

Decision-tree-based machine learning for detecting coffee agroforestry using SPOT-7

10.12928/telkomnika.v24i4.27747
I Made; IPB University Khrisna Yoga Devandra , I Nengah; IPB University Surati Jaya , Tatang; IPB University Tiryana
This study develops a decision-tree-based machine-learning (ML) approach to identify coffee agroforestry plants using SPOT-7 satellite imagery. The algorithm was developed by examining the combination of image indices derived from SPOT-7 and biophysical variables. Detection using spectral variables is often hampered by spectral similarity between vegetation cover classes. This study found that a ML method that combines spectral and biophysical variables can significantly improve overall accuracy, from 60.4% (using conventional spectral variables alone) to 94% (using integrated spectral-biophysical variables). For detecting and identifying agroforestry coffee classes typically found under tree canopies, the addition of the “land cover” variable published by the Ministry of Environment and Forestry contributes significantly to the classification of agroforestry coffee. Important variables identified in this model are normalized difference vegetation index (NDVI), visible difference vegetation index (VDVI), normalized red-green vegetation index (NRGI), elevation, and land cover.
Volume: 24
Issue: 4
Page: 1307-1319
Publish at: 2026-08-01

NLP-driven hate speech detection on TikTok: a case study from UIN Sunan Ampel Surabaya

10.12928/telkomnika.v24i4.27419
Achmad; UIN Sunan Ampel Surabaya Teguh Wibowo , Aris; UIN Sunan Ampel Surabaya Fanani , Mujib; UIN Sunan Ampel Surabaya Ridwan , Bramasta; UIN Sunan Ampel Surabaya Kurnia Aji
This study examines hate speech detection in TikTok comments using natural language processing (NLP) techniques within the student community of UIN Sunan Ampel Surabaya. A dataset of 10,000 comments associated with the hashtag #PBAKUINSA2023 was analyzed using a lexicon-based sentiment analysis approach implemented through the TextBlob library, combined with Indonesian text preprocessing techniques, including tokenization, normalization, stopword removal, and stemming using the Sastrawi library. The results indicate that the proposed approach achieved an accuracy of 0.85, with precision of 0.88, recall of 0.83, and an F1-score of 0.854. Most comments were classified as neutral, while 31.8% were positive, and only a small proportion were negative. These findings suggest that discussions related to campus activities tend to be neutral or supportive. However, the findings also reveal that sentiment polarity does not always directly correspond to hate speech, as certain harmful expressions may appear neutral in lexicon-based analysis. This limitation highlights the need for more context-aware approaches. Overall, the proposed method provides an efficient solution for monitoring online discourse in academic environments.
Volume: 24
Issue: 4
Page: 1157-1167
Publish at: 2026-08-01

Integration of public electric vehicle charging stations into a single gateway platform for net zero emissions goals

10.12928/telkomnika.v24i4.27786
Fajar; STMIK AMIK Bandung Sidiq Arrizal , Shafira; STMIK AMIK Bandung Febriani
The rapid adoption of electric vehicles (EVs) in Indonesia necessitates robust charging infrastructure. However, a major challenge is the highly fragmented network of public EV charging stations (SPKLU). Various operators utilize disparate protocols, resulting in isolated data silos and user range anxiety. This study proposes a nationally scalable application programming interface (API)-based gateway for cross-operator SPKLU integration. The system uses an event-driven internet of things (IoT) push mechanism. A standardized JavaScript object notation (JSON) API ensures interoperability and normalizes heterogeneous SPKLU data into a unified national dashboard. Pilot testing across 20 SPKLU locations in 10 major Indonesian cities validated cross-regional network reliability. The results demonstrate successful real-time status synchronization with a low margin of error (MoE) of 4.65%. To mitigate minor discrepancies caused by temporary network latency, a user interface (UI)-level timestamp transparency feature is proposed. Furthermore, by utilizing dynamic, region specific grid emission factors, the system accurately recorded an aggregated electricity consumption of approximately 3.5 million kWh in June 2025. This translates to an estimated 2.7 million kgCO₂eq in carbon emissions. Ultimately, this centralized platform eliminates information asymmetry for users. It also lays a crucial foundation for future smart grid integration, artificial intelligence (AI)-based demand forecasting, and Indonesia’s net zero emission (NZE) targets.
Volume: 24
Issue: 4
Page: 1113-1120
Publish at: 2026-08-01

Remaining useful life estimation for predictive battery maintenance with improved recurrent singular spectrum analysis algorithm

10.11591/ijece.v16i4.pp1817-1831
Chutipongse Boonyakitmaitree , Suchada Sitjongsataporn
As the global electric vehicle (EV) battery market is projected to reach a valuation of over USD 100 billion by the end of 2026, the demand for sophisticated battery management systems (BMS) has become more critical than ever. Accurate remaining useful life (RUL) prediction is essential for ensuring vehicle safety, optimizing maintenance, and evaluating retired batteries for second-life applications. However, existing prognostic methods often struggle to balance computational efficiency with predictive accuracy, especially during the early stages of battery usage. This research proposes combined weighted similarity-based and recurrent singular spectrum analysis (CWS-RSSA), a hybrid forecasting framework that integrates RSSA with a similarity-based approach through a weighted logistic switching mechanism. The algorithm is designed to be computationally lightweight, making it suitable for resource-constrained BMS hardware. The proposed method was validated using NASA and a large-scale dataset from MIT-Stanford consisting of 124 lithium-ion cells. Experimental results demonstrate that CWS-RSSA is capable of early-stage prediction with a relative error of 19.8%, whereas existing methods are unable to provide predictions. In later stages, once sufficient data becomes available, the algorithm achieves near-perfect accuracy with a negligible relative error on the NASA dataset and an average relative error of only 0.14% across the 124 MIT-Stanford batteries. Furthermore, the algorithm demonstrates robust performance in handling capacity regeneration phenomena. These findings suggest that CWS-RSSA represents a scalable and practical advancement for battery health management, supporting the transition toward a sustainable circular energy economy and providing a reliable foundation for second-life battery certification.
Volume: 16
Issue: 4
Page: 1817-1831
Publish at: 2026-08-01

Integrating principal component analysis in spatial-spectral fusion models for hyperspectral image segmentation

10.11591/ijece.v16i4.pp2074-2086
Alexander Calvin , Laksmita Rahadianti
Hyperspectral imaging (HSI) from unmanned aerial vehicles (UAVs) provides rich spatial-spectral data, but its high dimensionality presents significant computational challenges for semantic segmentation. While state-of-the-art models like the transformer-based HSI-TransUnet are often employed, they introduce massive computational overhead. This study adapts a lightweight, dual-tunnel deep convolutional neural network (DCNN) framework for land-use segmentation on hyperspectral images by integrating PCA-based spatial reduction in the spatial branch, and benchmarks it on the UAV-HSI-Crop dataset against HSI- TransUnet. For further analysis, an ablation study compares principal component analysis (PCA) and local similarity projection (LSP) as spatial feature ex- tractors. The results demonstrate a significant performance and efficiency advantage. Our proposed PCA-based model (271.1K parameters) obtained a Kappa (κ) of 0.8582, overall accuracy (OA) of 0.8800, and average accuracy (AA) of 0.4918, outperforming the LSP-based model by 0.65% in κ, 0.51% in OA, and 2.16% in AA and the HSI-TransUnet baseline by 2.35% in κ, 1.95% in OA, and 8.10% in AA. On our experimental setup, this result was achieved with a 152.7-fold reduction in model size, a 14.2-fold decrease in training time, and a 4.6-fold speedup in inference relative to the reported HSI-TransUnet baseline. These findings show that the PCA-based dual-tunnel DCNN provides a favor- able trade-off between class-balanced accuracy and computational efficiency for this HSI segmentation task.
Volume: 16
Issue: 4
Page: 2074-2086
Publish at: 2026-08-01

Spatial and channel attention mechanism for speech disfluency detection using deep learning technique

10.11591/ijece.v16i4.pp2106-2119
Kusuma H. R. , G. Seshikala
Stuttering is a speech communication disorder, it is characterized by repetitions, prolongation, and unusual pauses that cause interference with the natural flow of speech. In recent times, automatic speech recognition and speech processing systems have gained enormous attention because they are used in most of the human machine interaction applications. However, the performance of these systems is affected by stutter speech, stutter detection is the major challenge due to speech disfluencies. To address this major challenge, this paper introduced a novel deep learning (DL) based paradigm, which integrates a hybrid feature extraction algorithm, with the Spatial and Channel attention mechanism to refine the features and for reliable detection of speech disfluency. This study is conducted on multiple stutter data set which includes UCLASS (Release 1, Release 2), FluencyBank and SEP-28k. The major drawback of all these data sets is data imbalance. To reduce this imbalance, the author used data augmentation techniques, which includes, noise, music, reverberation and pitch shifting methods. However, increasing the stutter detection accuracy remains a challenging issue. To address this issue, the author proposed a hybrid feature extraction model, which extracts temporal, contextual, spectral, and pitch information from the speech signal. The obtained features are then processed through the attention mechanism where channel and spatial attention models help to refine the features. Finally, a multiclass convolutional neural network (CNN) classifier is used to detect the stutter event in the speech signals. The results show that our model with spatial and channel attention mechanism performs better than existing deep learning approaches and accurately detects stuttering.
Volume: 16
Issue: 4
Page: 2106-2119
Publish at: 2026-08-01

Decentralized multi-agent orchestration for legacy order-to cash optimization

10.12928/telkomnika.v24i4.27807
Rahul Kumar; University of Connecticut Thatikonda , Sucharitha; Point Park University Donepudi
Legacy enterprise resource planning (ERP) systems serve as the operational backbone of global commerce but often create bottlenecks due to their rigid, monolithic design. As organizations incorporate artificial intelligence (AI), these outdated systems struggle to support high-speed, parallel workflows, creating a significant integration challenge. This paper introduces a non intrusive modernization approach that overlays a decentralized multi-agent system (MAS) onto existing infrastructure without requiring invasive code changes. By developing a digital twin of the order-to-cash (O2C) process, we train autonomous agents through multi-agent reinforcement learning (MARL) to manage credit validation, inventory allocation, and fulfillment. We adapt the centralized training, decentralized execution (CTDE) framework to meet O2C constraints, enabling agents to learn globally optimal strategies while operating independently. Simulation results show that this architecture surpasses rule-based robotic process automation (RPA) baselines, increasing total throughput by 6.9% over a monolithic setup, though at a 6.3% error rate due to aggressive allocation policies. These results indicate that decentralized agent-based orchestration provides a scalable approach for modernizing legacy ERPs, offering increased agility without the risks associated with platform replacement.
Volume: 24
Issue: 4
Page: 1216-1223
Publish at: 2026-08-01
1 2 3
...
2060 2061
Show 5 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