cover
Contact Name
M. Miftach Fakhri
Contact Email
fakhri@unm.ac.id
Phone
+6282290603030
Journal Mail Official
wahid@unm.ac.id
Editorial Address
Program Studi Teknik Komputer, UNM Parangtambung, Daeng Tata Raya, Makassar, South Sulawesi, Indonesia
Location
Kota makassar,
Sulawesi selatan
INDONESIA
Journal of Embedded Systems, Security and Intelligent Systems
ISSN : 2745925X     EISSN : 2722273X     DOI : -
Core Subject : Science,
The Journal of Embedded System Security and Intelligent System (JESSI), ISSN/e-ISSN 2745-925X/2722-273X covers all topics of technology in the field of embedded system, computer and network security, and intelligence system as well as innovative and productive ideas related to emerging technology and computer engineering, including but not limited to : Network Security System Security Information Security Social Network & Digital Security Cyber Crime Machine Learning Decision Support System Intelligent System Fuzzy System Evolutionary Computating Internet of Thing Micro & Nano Technology Sensor Network Renewable Energy Wearable Devices Embedded Robotics Microcontroller
Articles 216 Documents
Decision Support System Integrating Entropy Weighting and MARCOS Ranking for Multi-Criteria Data-Driven Prioritization Lince Tomoria Sianturi; Berto Nadeak; Asyahri Hadi Nasyuha; Moses Adeolu Agoi
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/r0d49p45

Abstract

Purpose – This study aims to develop a data-driven decision support system that integrates entropy-based objective weighting with the MARCOS ranking method to improve multi-criteria prioritization in credit risk assessment by enhancing objectivity, consistency, and robustness of decision-making outcomes. Design/methods/approach – A hybrid MCDM framework is proposed, combining entropy weighting to determine criterion importance based on data variability and the MARCOS method to rank alternatives relative to ideal and anti-ideal solutions. The approach is evaluated using the Statlog German Credit dataset consisting of 1,000 applicants and six evaluation criteria. Performance is assessed through comparative analysis with conventional methods (TOPSIS and VIKOR), sensitivity testing under weight perturbation, and stability analysis using Spearman rank correlation. Findings - The results demonstrate that the proposed Entropy–MARCOS framework produces reliable and consistent prioritization outcomes. The model achieves a high ranking stability with a Spearman correlation of 0.91 and outperforms conventional MCDM methods in terms of ranking consistency. The findings also indicate that criteria such as age and employment duration have the highest discriminative importance, and the method remains robust under moderate variations in criterion weights. Research implications/limitations – However, the evaluation is limited to a single dataset and static criteria weights, which may affect generalizability across different domains or dynamic environments. Future research should explore adaptive weighting mechanisms and validate the model on more diverse datasets. Originality/value – This research contributes a unified hybrid framework that combines entropy-based objective weighting with the MARCOS ranking method, providing a more transparent, data-driven, and stable approach for multi-criteria decision-making, particularly in credit risk prioritization contexts.
A Hybrid Clustering Approach Integrating K-Means and DBSCAN for Customer Segmentation Zulfi Azhar; Soeb Aripin; Azanuddin
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/2qjpr814

Abstract

Purpose – This study aims to improve customer segmentation quality by developing a hybrid clustering approach that integrates centroid-based and density-based techniques to better capture complex data structures and noise in customer datasets. Design/methods/approach – The proposed method combines K-Means and DBSCAN in a sequential hybrid framework. K-Means is first applied to identify the global structure of customer groups using centroid similarity. Subsequently, DBSCAN is performed within each cluster to refine cluster boundaries and detect noise points. The dataset is preprocessed using Min–Max normalization, and clustering performance is evaluated using the Silhouette Score and Davies–Bouldin Index. Findings - Experimental results show that the hybrid approach outperforms standalone methods. The proposed model achieves a Silhouette Score of 0.71 and a Davies–Bouldin Index of 0.42, indicating improved cluster compactness and separation. Additionally, the method successfully identifies 6% of data points as noise, enhancing segmentation reliability and interpretability. Research implications/limitations – This study demonstrates the effectiveness of combining clustering paradigms for improved segmentation. However, the evaluation is limited to a relatively small dataset with three features, and DBSCAN parameter selection remains data-dependent. Future research may explore larger datasets, higher-dimensional features, and automated parameter optimization techniques. Originality/value – This research contributes a practical hybrid clustering framework that integrates K-Means and DBSCAN in a structured manner, enabling more robust, interpretable, and noise-aware customer segmentation suitable for data-driven marketing analytics and decision support systems.
IoT–DSS-Based Fleet Management System for Enhancing the Operational Efficiency of Fishing Vessels Romadhoni Roma; Budhi Santoso; Johny Custer; M. Nur Faizi; Mohamed Nasir Alivi; Abdul Razak Shaari
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i1.2615

Abstract

Purpose - This study aims to develop an IoT–Decision Support System (DSS)-based Smart Fleet Management System (SFMS) to improve the operational efficiency of small- to medium-scale fishing vessels through real-time monitoring and data-driven decision-making. Design/methods/approach – The proposed system integrates IoT sensors (GPS, fuel flow, temperature) with an ESP32 microcontroller for real-time data acquisition and transmission via MQTT to a cloud platform. A hybrid DSS combining linear regression and fuzzy logic is developed to analyze vessel performance and generate operational recommendations. Field validation was conducted on three fishing vessels, with 1,200 telemetry samples for regression modeling, 300 decision samples for DSS evaluation, and 4,320 data packets for communication analysis. Findings - The implementation of SFMS resulted in significant improvements in operational performance. Fuel consumption decreased by 20.0% (from 50.7 L/h to 40.6 L/h), idle operational time was reduced by 28.9% (from 3.8 to 2.7 hours/day), and the Operational Efficiency Index (OEI) improved by 22.7% (from 0.110 to 0.135 kn/L). The DSS achieved an accuracy of 92.7% in decision recommendations, while system reliability reached 99.2% uptime with low latency and acceptable packet loss (1.48%). Research implications/limitations – Although effective, the study is limited by a small number of vessels and the lack of synchronized environmental data, suggesting the need for broader validation. Originality/value – This study presents a cost-effective and scalable IoT–DSS framework tailored for small-scale fisheries, supporting sustainable operations and maritime digital transformation.
Multiclass IoT Intrusion Detection Based on Particle Swarm Optimization-Tuned Light Gradient Boosting Machine Fajar Ratnawati; Agus Tedyyana
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i1.2612

Abstract

Purpose – This study aims to develop a robust multiclass intrusion detection system (IDS) for Internet of Things (IoT) environments by optimizing Light Gradient Boosting Machine (LightGBM) using Particle Swarm Optimization (PSO), with a focus on improving performance under severe class imbalance. Design/methods/approach – A PSO-based hyperparameter tuning framework is applied to LightGBM, where Macro F1-score is used as the fitness function to ensure balanced class performance. The model is evaluated on the RT-IoT2022 dataset using a leakage-safe stratified 70:15:15 split. Performance is assessed using Accuracy, Macro Precision, Macro Recall, Macro F1-score, Weighted F1-score, and Matthews Correlation Coefficient (MCC). Experiments are repeated across 10 runs, and statistical significance is validated using the Wilcoxon signed-rank test. Findings - The proposed PSO-LightGBM model significantly outperforms the baseline LightGBM. It achieves 99.75% accuracy, 97.10% macro F1-score, 99.75% weighted F1-score, and 99.37% MCC, compared to 85.36%, 25.44%, 84.23%, and 63.67%, respectively, for the baseline. The model demonstrates substantial improvement in minority-class detection, reducing misclassification and preventing class collapse observed in the baseline. Research implications/limitations – The findings highlight the effectiveness of Macro-F1-guided optimization for imbalanced multiclass IoT intrusion detection. However, the evaluation is limited to a single dataset and centralized experimental setting, which may affect generalizability. Originality/value – This study contributes a leakage-safe, Macro-F1-driven PSO-LightGBM framework with comprehensive evaluation, including class-wise analysis, repeated runs, and statistical testing, providing strong evidence for balanced multiclass IoT intrusion detection.
Wazuh-Based Security Monitoring for Public Service Web Systems: Detection Effectiveness, Alert Latency, and Resource Overhead Mansur; Nurmi Hidayasari; Kasmawi; Zuliar Efendi
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2636

Abstract

Purpose - This study evaluates the effectiveness of Wazuh-based security monitoring in detecting cyber threats in public service web systems, particularly XSS, SQL Injection, SSH brute-force attacks, and file integrity violations. Design/methods/approach – An experimental quantitative approach was conducted in a controlled virtualized environment using Proxmox VE. Wazuh was deployed as a centralized Security Information and Event Management (SIEM) system with active threat detection and File Integrity Monitoring (FIM). Simulated attacks and file modification events were executed, and system performance was assessed based on detection rate, alert latency, and resource utilization. Findings - Wazuh achieved a 100% detection rate across all tested attacks, including XSS (10/10), SQL Injection (10/10), and SSH brute-force (20/20). The average alert latency was 6.8 seconds for XSS, 132.6 seconds for SQL Injection, and 52 seconds for SSH brute-force attacks. Although CPU and memory usage increased after deployment, the overhead remained within acceptable operational limits in the experimental environment. Research implications/limitations – The system demonstrates effective real-time centralized monitoring capability; however, the evaluation was limited to a controlled virtual environment and predefined attack scenarios, which may not fully represent real-world conditions. Originality/value – This study provides empirical evidence of Wazuh’s performance as a centralized SIEM solution for public service web systems, highlighting its detection effectiveness and operational trade-offs in terms of alert latency and system resource usage.
Real-Time Intelligent IoT-Based Drum Brake Wear Monitoring System Dzaki Putra Prakosa; Setya Wijayanta
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2609

Abstract

Purpose – This study aims to develop and evaluate a real-time IoT-based intelligent monitoring system for drum brake lining wear to overcome the limitations of conventional manual inspection, which is periodic, subjective, and prone to delayed detection of critical wear conditions. Design/methods/approach – The research adopts a Research and Development (R&D) methodology consisting of design, prototyping, and laboratory testing. The system integrates an ESP32 microcontroller with a VL53L0X time-of-flight sensor to measure brake lining thickness in real time. A rule-based classification algorithm is implemented to categorize brake conditions into SAFE, WARNING, and DANGER states. Experimental evaluation was conducted through 15 trials across a thickness range of 1.0–10.0 mm. Performance metrics include accuracy, mean absolute error (MAE), root mean square error (RMSE), and response time. Findings - The system achieved an average measurement accuracy of 96.0%, MAE of 0.13 mm, and RMSE of 0.15 mm. All samples were correctly classified, resulting in 100% classification accuracy across the three condition states. The system also recorded a mean response time of 1.27 seconds, indicating fast and reliable real-time performance under controlled conditions. Research implications/limitations – The system is feasible for low-cost brake wear monitoring, but validation is limited to laboratory conditions with a small dataset, and real-world factors were not examined. Originality/value – This study presents an integrated IoT-based drum brake monitoring framework combining ToF sensing, embedded rule-based intelligence, and mobile notification in a single low-cost system. It specifically addresses drum brake applications, which remain underexplored compared to disc brake monitoring systems, offering a practical solution for resource-constrained environments.
Evaluating CNN Robustness for Face Mask Classification under Environmental Variations Bagaskara Ridho Vandio; Fatma Indriani; Andi Farmadi; Dodon Turianto Nugrahadi; Friska Abadi
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2617

Abstract

Purpose – This study aims to analyze and compare the performance of ResNet50 and MobileNetV3 for multi-class face mask classification under various environmental conditions. Design/methods/approach – ResNet50 and MobileNetV3 are trained using transfer learning for three-class face mask classification and evaluated under normal conditions and environmental variations, including illumination changes, blur, low compression, and rotation. Findings – Experimental results show that ResNet50 achieves an accuracy of 94.32% under normal conditions, slightly outperforming MobileNetV3 at 94.10%. Under environmental variations, the largest performance degradation is observed under darkening and blur conditions, while low compression and rotation have relatively minor effects. ResNet50 demonstrates higher robustness across most perturbation settings, whereas MobileNetV3 provides competitive performance with substantially better computational efficiency. Research implications/limitations – This study is limited to a controlled evaluation using synthetic environmental perturbations on a single dataset and does not consider broader dataset diversity. Therefore, the findings should be interpreted within the evaluated experimental conditions. Originality/value – This study provides a comparative analysis of model robustness under controlled environmental perturbations, highlighting the trade-off between robustness and computational efficiency for face mask classification systems.
Text Classification of 2024 Regional Head Elections Logistics Distribution in Online News Using Support Vector Machine Murni Kassa; Irene Realyta Haldy Trosi Tangkawarow; Audy Aldrin Kenap
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2619

Abstract

Purpose – This study aims to classify online news related to logistics distribution issues in the 2024 Indonesian Regional Head Elections using Support Vector Machine and imbalance handling strategies. Design/methods/approach – A total of 1,355 online news articles were collected from nine national news portals through web scraping. The research workflow involved data preprocessing, rule-based weak supervision, manual validation, TF–IDF feature extraction, oversampling using SMOTE and ADASYN, class-weighted learning, and SVM classification with Linear, RBF, Polynomial, and Sigmoid kernels. Model performance was evaluated using macro-averaged F1-score, 5-fold cross-validation, classification report, and confusion matrix analysis. Findings - The results show that Linear and RBF kernels produced more consistent performance than Polynomial and Sigmoid kernels for sparse TF–IDF representations. The RBF kernel with class-weighted learning achieved the highest hold-out macro F1-score of 0.641, although cross-validation results showed only marginal differences among top-performing configurations. The model performed well in classifying “No Issues” and “Damaged” categories but still struggled with the minority “Late” class. Research implications/limitations – The findings indicate that machine learning can support preliminary election logistics monitoring, but the model should not yet be used as a fully automated early-warning system due to minority-class limitations and weak-labeling constraints. Originality/value – This study contributes empirical evidence on SVM-based imbalanced text classification for election logistics news monitoring in the Indonesian Pilkada context.
Secure Automated Reconnaissance Using LLM Agents and a Layered Cryptographic Protection Pipeline Ikhwan Ruslianto; Wijang Widhiarso; Hafiz Muhardi
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2621

Abstract

Purpose – This study aims to design and evaluate a secure reconnaissance platform that integrates Large Language Model (LLM) agents for dynamic tool orchestration with a layered cryptographic protection pipeline to accelerate penetration-testing information gathering while protecting sensitive artefacts. Design/methods/approach – The platform unifies Nmap, WHOIS, and theHarvester under an LLM controller that generates command-line parameters through schema-constrained orchestration. Each output is validated against a strict JSON schema before execution. The protection pipeline applies AES-256-GCM with envelope keys for confidentiality, HMAC-SHA256 hash chaining for tamper-evident logs, Ed25519 signatures for report-level non-repudiation, and Argon2id-derived session keys. Evaluation was conducted on three public domains across thirty runs each, measuring latency, cryptographic overhead, verification integrity, signature validation, and an internal CVSS-informed triage score. Findings - The prototype showed that automated reconnaissance and cryptographic auditability can be combined with limited performance cost. A full pass over untan.ac.id completed in 14.97 seconds and produced an internal triage-heuristic score of 78/100. Cryptographic operations added 312 ms on average, equal to about 2.08% of total latency. All hash-chain links were verified, and Ed25519 signatures were validated in 71 µs. Research implications/limitations – The findings support red-team and blue-team workflows requiring faster, auditable reconnaissance reporting. However, the evidence is limited to three public domains under one network condition; therefore, the results should be interpreted as feasibility evidence, not generalisable performance claims. The risk score is an internal prioritisation heuristic, not a validated severity instrument. Originality/value – The study contributes a secure LLM-orchestrated reconnaissance framework that integrates structured command orchestration with cryptographic safeguards for confidentiality, integrity, and non-repudiation.
Comparative Analysis of MLP, 1D CNN, and Hybrid Models for Elbow Joint Angle Classification Using Surface Electromyography Signals Riky Tri Yunardi; Nasa Zata Dina; Deny Arifianto
Journal of Embedded Systems, Security and Intelligent Systems Vol 7 No 2 (2026): June 2026
Publisher : Program Studi Teknik Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59562/jessi.v7i2.2622

Abstract

Purpose – This study aims to analyze and compare the performance of a Multi-Layer Perceptron (MLP), a One-Dimensional Convolutional Neural Network (1D CNN), and a hybrid model for classifying three elbow joint angle movements using surface electromyography (sEMG) signals. Design/methods/approach – This dataset consists of sEMG signals collected from 15 healthy participants performing three elbow joint angle movements at 45°, 90°, and 135° with 50 repetitions, for a total of 2,250 data points. The proposed MLP model uses time-domain features extracted using Root Mean Square (RMS) and a Kalman filter, while a 1D CNN learns features from the raw, segmented signals from Shield EMG. A hybrid model combines both features. Model performance is evaluated using accuracy, precision, recall, F1 score, and confusion matrix. Findings - The results show that the 1D CNN model achieved score accuracy 0.78, outperforming the MLP model with accuracy of 0.65 , indicating superior feature learning from the raw sEMG signal. The hybrid model achieved accuracy of 0.82 and was more stable in discriminating intermediate elbow joint angles, indicating that feature fusion improves classification reliability. Research implications/limitations – This study was limited by the relatively small number of participants and the lack of an external validation dataset, which may impact the generalizability of the results. Future research should include larger and more diverse populations and explore more advanced architectures. Originality/value – This study provides a comparison between MLP, 1D CNN, and hybrid model approaches for sEMG-based elbow joint angle classification, highlighting the strengths of each method and offering insights for the development of robust rehabilitation technologies.