cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
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
Published by Universitas Negeri Makassar
ISSN : 2745925X     EISSN : 2722273X     DOI : -
Core Subject : Science,
Computer Science & IT
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
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 240 Documents
 20   21   22   23   24 
Vision-AI Roadside Unit Simulator with SUMO Integration for Visually Impaired Mobility Optimization Bambang Cahyono; Fara Triadi; Karyo Budi Utomo
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.v7i2.2625

Abstract

Purpose – This study develops and evaluates a low-cost Vision-AI-based roadside unit simulator integrated with SUMO to support adaptive traffic signal control and safer crossing for visually impaired pedestrians. Design/methods/approach – The proposed system combines a Raspberry Pi 4 edge platform, a camera sensor, YOLOv5s-based vehicle detection, and SUMO traffic simulation through the TraCI interface. Three control strategies were compared: fixed-time control, SUMO adaptive control, and the proposed Vision-AI-assisted control. Experiments were conducted under low, medium, and high traffic density scenarios, with 30 simulation runs for each condition. Performance was measured using average waiting time, queue length, travel time, pedestrian crossing success rate, detection accuracy, latency, and statistical significance testing. Findings - The proposed system outperformed both baseline methods across all scenarios. It reduced vehicle waiting time by up to 41%, queue length by approximately 35%, and travel time by around 22% compared with fixed-time control. The assistive crossing mechanism increased pedestrian crossing success from 62% to 93%. The edge platform achieved 18–22 FPS, latency below 85 ms, and mAP@0.5 of 0.87. Research implications/limitations – The findings demonstrate the feasibility of low-cost edge-based intelligent transportation systems, although validation remains limited to simulation and a single-intersection case. Originality/value – This study integrates Vision-AI traffic perception, adaptive signal control, and accessibility-aware pedestrian support within one simulation framework.
A Multi-Attribute Utility Theory-Based Ranking System for PKH Social Assistance Recipients Rezki Kurniati; Aldi Suhandi
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.v7i2.2628

Abstract

Purpose – This study aims to design and evaluate a mobile-based decision support system for ranking prospective recipients of the Program Keluarga Harapan (PKH) in Parit Satu Api-Api Village. The system is intended to support objective, transparent, and traceable recipient selection, particularly in village-level contexts where manual selection may lead to subjectivity, delays, and limited accountability. Design/methods/approach – The system was developed using the waterfall model. Multi-Attribute Utility Theory (MAUT) was applied to calculate utility and preference values based on six criteria: household income, family dependents, social status, occupation, housing condition, and school-age children. A simulation using seven alternatives was conducted, and system functionality was evaluated through black-box testing. Findings - The system provides criteria, subcriteria, alternative, assessment, ranking, user management, and PDF reporting features. The simulation results show that alternative A1 obtained the highest preference value of 0.653, followed by A4 and A6 with equal values of 0.650. The corrected MAUT calculation produced a more consistent ranking because higher subcriteria scores represented greater socioeconomic vulnerability and higher assistance priority. The system also automates normalization, weighting, ranking, and PDF-based reporting. Research implications/limitations – The system can support faster, more systematic, and better-documented PKH recipient selection at the village level. However, this study is limited to a small simulation dataset and black-box functionality testing. Further evaluation using larger real-world data, controlled timing tests, and user acceptance assessment is required. Originality/value – This study contributes a mobile-oriented MAUT-based decision support system tailored to village officer workflows, with emphasis on calculation traceability, recipient ranking, and PDF reporting.
A CNN–LSTM–DQN Policy with Prioritized Experience Replay for Cost-Aware Intrusion Detection on CSE-CIC-IDS2018 Rushendra; Kalamullah Ramli; Prima Dewi Purnamasari
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.v7i2.2629

Abstract

Purpose – This study aims to examine intrusion detection as a policy learning problem and determine how replay strategy alone controls the operational trade-off between attack recall and alert volume under extreme class imbalance. Design/methods/approach – A controlled ablation study was conducted using a fixed CNN–LSTM feature extractor, fixed feature set, and fixed reward structure combined with a Deep Q-Network (DQN) agent. Three configurations were compared: naïve DQN without replay, uniform experience replay with a target network, and Prioritized Experience Replay (PER). Experiments used the CSE-CIC-IDS2018 dataset, consisting of 10,788,508 training flows and 2,697,128 testing flows, with attack events occurring at fewer than 70 per million flows. Performance was assessed through recall and alerts per million flows (ARMF). Findings - Supervised CNN–LSTM baselines achieved recall above 95% but generated 31,000–45,000 ARMF. Naïve DQN reduced ARMF to 383 but sharply decreased recall to 42.47%. Uniform replay improved recall to 84.95% but increased ARMF to 12,728. PER achieved the most balanced operating point, reaching 91.40% recall at 1,031 ARMF, approximately 30 times fewer alerts than the supervised CNN–LSTM reference, with a 5.91-percentage-point recall cost. Research implications/limitations – The findings indicate that replay distribution is a critical operational design variable for controlling alert volume in highly imbalanced intrusion detection settings. However, the study is limited to a fixed backbone, feature set, reward shape, and dataset. Originality/value – This study demonstrates that replay strategy can substantially reshape IDS operating points independently of model architecture or feature representation.
Machine Learning for Predicting Poverty and Educational Outcomes: A Comparative Simulation Study for Evidence Based Social Policy Syamsul Bhahri; Renny; Rachmat
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.v7i2.2630

Abstract

Purpose – This study aims to develop and test a comparative machine learning evaluation framework for predicting poverty status and educational risk as a methodological basis for evidence-based social policy. Design/methods/approach – A comparative simulation study was conducted using a controlled simulated dataset of 10,000 observations, sixteen input features, and two binary targets: poverty status and educational risk. Five supervised classification models were evaluated: Logistic Regression, Decision Tree, Random Forest, XGBoost, and LightGBM. The models were assessed using accuracy, F1-score, AUC, Brier Score, per-class performance, cross-validation stability, explainability, and a proposed Policy Readiness Index. The dataset included predefined prevalence assumptions, missing values, outliers, and simulated nonlinear and interaction effects. Findings - Within the controlled simulation setting, XGBoost achieved the strongest technical performance across both prediction tasks, with the highest accuracy, F1-score, AUC, and calibration quality. However, Random Forest obtained the highest Policy Readiness Index because it provided the best balance between predictive performance, cross-validation stability, and interpretable feature attribution. The findings show that the technically best model is not automatically the most policy-ready model. Research implications/limitations – The study offers a structured decision-support approach for comparing machine learning models in poverty and education policy contexts. However, all results are derived from simulated data and should be interpreted as a methodological proof of concept rather than empirical evidence for a specific real-world population. Originality/value – This study contributes a policy-oriented machine learning evaluation framework that integrates predictive quality, calibration, stability, explainability, and policy usability into a transparent Policy Readiness Index.
Integration of Machine Learning-Based Preeclampsia Risk Prediction Model in Digital Health Economics System for Cost Efficiency Evaluation of Obstetric Services Qatrunnada Salsabila Delfi; Muhamad Isa Sani; Alfiah Hasanah
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.v7i2.2633

Abstract

Purpose – This study aims to develop and evaluate a machine learning-based preeclampsia risk prediction model integrated with payer-stratified health-economic analysis to assess the clinical feasibility, cost efficiency, and equity implications of digital obstetric screening in a multipayer healthcare system. Methods – A retrospective prediction-model and budget-impact study was conducted using de-identified electronic medical record and hospital billing data from 465 pregnant patients in an Indonesian tertiary referral hospital. Logistic Regression and XGBoost models were developed using routine clinical variables. Model performance was assessed through five-fold Group K-Fold cross-validation, calibration using Isotonic Regression, SHAP-based interpretability, Decision Curve Analysis, and payer-stratified economic evaluation across Private, Commercial Insurance, and BPJS groups. Findings – The XGBoost model achieved strong predictive performance with ROC-AUC of 0.9315 and PR-AUC of 0.8013. Calibration error approached 0.0000, indicating reliable probability estimates. SHAP analysis showed clinically plausible predictors, particularly systolic blood pressure, mean arterial pressure, and gestational age. A 1% threshold achieved 100% sensitivity, eliminating missed preeclampsia cases. Economically, model-guided screening reduced costs for Private and Insurance patients but increased costs for BPJS patients due to reimbursement misalignment. Research Implications – The findings indicate that clinical AI implementation requires not only predictive accuracy but also payer-aware financing strategies to prevent inequitable access. Originality – This study offers a novel integration of calibrated machine learning prediction, clinical interpretability, threshold optimization, and payer-stratified economic evaluation in preeclampsia screening.
Real-Time Ambulance Detection System at Traffic Intersections Using Raspberry Pi and YOLOv5 Muhammad Isro’ Risqi; Raka Pratindy; Dzaki Putra Prakosa
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.v7i2.2635

Abstract

Purpose – Delays experienced by ambulances at signalized intersections remain a critical issue in emergency transportation, particularly in dense urban traffic conditions. This study aims to develop and evaluate a low-cost real-time ambulance detection system using Raspberry Pi 4B and YOLOv5 to support intelligent transportation monitoring and emergency vehicle prioritization. Design/methods/approach – This study employed an experimental research design by integrating CCTV cameras, Raspberry Pi 4B, YOLOv5s object detection, ONNX Runtime INT8 optimization, and Telegram Bot API notification. The model was trained using 3,250 annotated ambulance images divided into training, validation, and testing subsets. System performance was evaluated under five operational scenarios: daytime, nighttime, heavy traffic, long-distance detection, and low-lighting conditions. Findings – The proposed YOLOv5s model achieved precision of 95.4%, recall of 93.8%, mAP@0.5 of 96.1%, and sustained throughput of 22 FPS on Raspberry Pi 4B. The Telegram notification subsystem achieved a transmission success rate of 98.7% with an average delay of 1.8 seconds. However, detection performance decreased under low-lighting conditions, with a true positive rate of 78.5% and false positive rate of 11.2%. Research implications/limitations – The system demonstrates the feasibility of deploying embedded computer vision for cost-effective ambulance detection, although nighttime reliability and traffic signal integration require further improvement. Originality/value – This study contributes an ONNX INT8-optimized YOLOv5s implementation on Raspberry Pi 4B with multi-condition evaluation and real-time Telegram notification for ambulance detection at traffic intersections.
Analysis of Optimal Epoch Selection for YOLO26 Model in Detecting Graves and Free Slots Using UAV Photogrammetry Hafiz Irsyad; Muhammad Rizky Pribadi; Dedy Hermanto; Dina Lestari Putri
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.v7i2.11573

Abstract

Purpose – This study investigates the optimal number of training epochs for the YOLO26 model in detecting graves and free burial slots from UAV photogrammetry imagery, with particular attention to model convergence, generalization, and detection performance. Design/methods/approach – A publicly available cemetery dataset containing two object classes, namely graves and free burial slots, was preprocessed using auto-orientation, 2×2 tiling, resizing to 640 × 640 pixels, grayscale conversion, and data augmentation. The YOLO26 model was trained using transfer learning under six epoch configurations: 50, 100, 150, 200, 250, and 300 epochs. Performance was evaluated using precision, recall, F1-score, mAP@50, mAP@50–95, confusion matrices, and training and validation loss curves. Findings – Model performance improved substantially as training progressed and began to stabilize after approximately 200 epochs. The highest observed performance occurred at epoch 289, with a precision of 98.78%, recall of 98.26%, F1-score of 99%, mAP@50 of 99.40%, and mAP@50–95 of 90.94%. Although the 300-epoch configuration produced similarly strong results, the additional gains were marginal, indicating diminishing returns after convergence. Research implications/limitations – The findings provide practical guidance for selecting an appropriate training duration in UAV-based cemetery mapping and small-object detection. However, the study relies on a relatively small, single-source dataset, which may limit generalizability across different cemetery layouts, environmental conditions, and UAV imaging configurations. Originality/value – This study provides a domain-specific multi-epoch benchmark for YOLO26 and demonstrates the importance of metric- and convergence-based checkpoint selection rather than relying solely on the maximum predefined number of epochs.
An Enhanced Informed-RRT* Integrating Bidirectional Exploration, Fast Sampling, and Local Path Optimization Heru Suwoyo; Taufik Hidayat; Yingzhong Tian; Julpri Andika; Andi Adriansyah
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.v7i2.12223

Abstract

Purpose – This study aims to improve the efficiency and path quality of Informed-RRT* for mobile robot path planning in complex environments by integrating bidirectional exploration, fast sampling, and local path optimization. Design/methods/approach – The proposed method follows a two-stage framework. In the initial exploration phase, two search trees are expanded bidirectionally using fast sampling to accelerate feasible path discovery. An ancestor-based parent selection mechanism based on the triangular inequality principle is then applied to improve local path quality and reduce inefficient rewiring. After an initial path is found, the algorithm switches to an informed optimization phase using ellipsoidal sampling while maintaining simultaneous two-tree expansion. The proposed method was evaluated in three two-dimensional static environments with different levels of complexity and compared with Fast-RRT* and Informed-RRT*. Performance was assessed using path cost, computation time, number of iterations, and Wilcoxon signed-rank testing. Findings – The results show that the proposed method consistently achieves lower path cost, shorter computation time, and fewer iterations than the baseline methods. Across simple to maze environments, the method reduces computation time by approximately 24–75% and improves path optimality by approximately 3.9–4.8%. Statistical testing confirms significant improvements in more constrained and maze-like environments. Research implications/limitations – The study demonstrates that combining accelerated exploration and local optimization can improve sampling-based path planning performance. However, the evaluation is limited to two-dimensional static environments and selected baseline methods. Originality/value – The proposed framework provides an enhanced Informed-RRT* approach that jointly addresses initial exploration efficiency, local path refinement, and informed optimization within a unified planning strategy.
A Hybrid VAE-CNN Framework for Unsupervised URL-Based Phishing Detection Hendra Kurniawan; Yoga Syahputra; Novrizal Fattah Fahmitra
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.v7i2.12237

Abstract

Purpose – URL-based phishing remains a persistent cybersecurity threat because attackers continuously generate deceptive links that evade blacklist-based and supervised detection systems. This study aims to develop and evaluate an unsupervised phishing detection framework that learns legitimate URL patterns without requiring labeled phishing samples during training. Design/methods/approach – A hybrid Variational Autoencoder–Convolutional Neural Network (VAE-CNN) model was proposed and compared with a Convolutional Autoencoder (CAE) baseline. Both models were trained only on legitimate URLs from the PhiUSIIL dataset and evaluated using repeated holdout validation across five random seeds. The experiments included four latent dimensions, five percentile-based anomaly thresholds, and McNemar’s exact test. In addition, external validation was conducted using PhishTank phishing URLs combined with PhiUSIIL legitimate holdout samples. Findings – In the main PhiUSIIL benchmark, the best VAE-CNN configuration achieved an accuracy of 0.9458, precision of 0.9091, recall of 0.9907, and F1-score of 0.9482, outperforming the best CAE baseline with an F1-score of 0.8920. In external validation, VAE-CNN achieved an F1-score of 0.9494, compared with 0.8470 for CAE. McNemar’s test confirmed statistically significant paired-prediction differences across all evaluated comparisons. Research implications/limitations – The findings indicate that probabilistic latent modeling improves unsupervised phishing detection, although the study remains limited to character-level URL representations and benchmark-based evaluation. Originality/value – The proposed VAE-CNN provides a statistically validated unsupervised framework for URL-based phishing detection and demonstrates promising cross-source performance under the evaluated PhiUSIIL–PhishTank setting, particularly when labeled malicious URLs are limited, delayed, or rapidly outdated.
Phishing URL Detection Using URL Characteristics: A Multi-Algorithm Machine Learning and Explainable AI (XAI) Approach Resky Ayu Dewi Talasari; Eka Alifia Kusnanti
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.v7i2.12398

Abstract

Purpose – This study aims to develop and evaluate a phishing URL detection system using URL-based characteristics and multiple machine learning algorithms, while integrating Explainable Artificial Intelligence (XAI) to improve model transparency. Design/methods/approach – The study used 24,390 URLs consisting of 4,878 phishing URLs and 19,512 legitimate URLs collected from a public Kaggle dataset and manually collected sources. Thirteen URL-based features were extracted, including URL length, protocol usage, number of subdomains, domain length, IP address usage, and suspicious keywords. Six machine learning algorithms were compared: Random Forest, XGBoost, Decision Tree, Support Vector Machine, Logistic Regression, and Naïve Bayes. The dataset was evaluated using an 80:20 stratified train-test split, and model performance was measured using accuracy, precision, recall, F1-score, and confusion matrix. XAI was implemented using model-based feature importance and SHAP local explanations. Findings – XGBoost achieved the best overall performance with 97.89% accuracy, 93.09% precision, 96.62% recall, and 94.82% F1-score for the phishing class, followed closely by Random Forest. Feature importance analysis showed that protocol usage, suspicious keywords, and domain length were the most influential features. SHAP explanations further demonstrated that phishing predictions were influenced by the combined contribution of multiple URL characteristics. Research implications/limitations – The findings indicate that ensemble-based machine learning combined with explainable output can support transparent phishing URL screening. However, the study was limited to URL-based features, a fixed train-test split, and manually validated URLs without systematic verification from external threat-intelligence platforms. Originality/value – This study contributes by combining multi-algorithm phishing URL detection, URL-characteristic-based feature analysis, SHAP-based explanation, whitelist checking, and typosquatting indicators within an interpretable cybersecurity detection approach.

Page 22 of 24 | Total Record : 240

 20   21   22   23   24 

Filter by Year

2020 2026


Reset
Filter By Issues
All Issue Vol 7 No 3 (2026): September 2026 Vol 7 No 2 (2026): June 2026 Vol 7 No 1 (2026): March 2026 Vol 6, No 4 (2025): Desember 2025 Vol 6, No 3 (2025): September 2025 Vol 6, No 2 (2025): June 2025 Vol 6, No 1 (2025): March 2025 Vol 5, No 3 (2024): November 2024 Vol 5, No 2 (2024): July 2024 Vol 5, No 1 (2024): March 2024 Vol 4, No 2 (2023): November 2023 Vol 4, No 1 (2023): May 2023 Vol 3, No 2 (2022): November 2022 Vol 3, No 1 (2022): May 2022 Vol 2, No 2 (2021): November 2021 Vol 2, No 1 (2021): May 2021 Vol 1, No 2 (2020): November 2020 Vol 1, No 1 (2020): May 2020 More Issue