cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
-
Contact Email
-
Phone
-
Journal Mail Official
-
Editorial Address
-
Location
Kab. sukabumi,
Jawa barat
INDONESIA
JTERA (Jurnal Teknologi Rekayasa)
Published by Politeknik Sukabumi
ISSN : 2548737X     EISSN : 25488678     DOI : -
Core Subject : Science, Engineering,
Civil Engineering, Building, Construction & Architecture Computer Science & IT Electrical & Electronics Engineering Industrial & Manufacturing Engineering Mechanical Engineering
TERA (Journal of Engineering Technology) is peer-review journal providing original research papers, case studies, and articles review in engineering technology field. The journal can be used as an authoritative source of scientific information for researchers, researcher academia or institution, industrial engineer, and government agencies. Paper topics of JTERA focused on engineering including informatics engineering, computer engineering, electrical engineering, mechanical engineering, industrial engineering, civil engineering, environmental engineering, and other topics related to engineering. JTERA is published by Politeknik Sukabumi and managed to be issued twice in every volume on June and December.
Arjuna Subject : -
Articles 533 Documents
 50   51   52   53   54 
Performance Comparison of YOLOv8n and Faster R-CNN for Automatic Detection of School Uniform Attributes Argi Ginanjar; Adhi Kusnadi; Risnandar
JTERA (Jurnal Teknologi Rekayasa) Vol 11 No 1: Vol. 11 No. 1: Juni 2026
Publisher : Politeknik Sukabumi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31544/jtera.v11.i1.2026.189-198

Abstract

Pemeriksaan kelengkapan atribut seragam siswa secara manual masih banyak dilakukan di lingkungan sekolah, khususnya pada sekolah kejuruan dengan jumlah siswa yang besar. Proses tersebut membutuhkan waktu, rentan terhadap subjektivitas, dan sulit terdokumentasi secara konsisten. Penelitian ini bertujuan membandingkan kinerja YOLOv8n dan Faster R-CNN MobileNetV3 320 FPN dalam mendeteksi atribut seragam sekolah secara otomatis. Dataset yang digunakan terdiri dari 638 citra seragam sekolah dengan lima kelas objek, yaitu epolet, name_tag, sabuk, sepatu_hitam, dan topi. Kedua model dilatih menggunakan dataset yang sama, kemudian dievaluasi berdasarkan Intersection over Union (IoU), mean Average Precision (mAP), dan Frames per Second (FPS) pada konfigurasi eksperimen yang digunakan. Hasil pengujian menunjukkan bahwa YOLOv8n memperoleh IoU sebesar 0,8713, mAP50 sebesar 0,9193, mAP50-95 sebesar 0,6695, dan kecepatan 83,5921 FPS. Sementara itu, Faster R-CNN memperoleh IoU sebesar 0,6714, mAP50 sebesar 0,6640, mAP50-95 sebesar 0,3853, dan kecepatan 9,5801 FPS. Hasil tersebut mengindikasikan bahwa, pada konfigurasi eksperimen dan subset evaluasi yang digunakan, YOLOv8n memiliki keseimbangan yang lebih baik antara akurasi lokalisasi, akurasi deteksi, dan efisiensi inferensi. Dengan demikian, YOLOv8n berpotensi lebih sesuai digunakan sebagai dasar pengembangan sistem pemantauan kelengkapan seragam sekolah secara semi-real-time.
Dari Risiko Terbuka ke Risiko Terkendali:Mengevaluasi Aturan Firewall MikroTik dalam Keamanan Jaringan Murhaban Murhaban; Muzakir Muzakir; Ulfa Merlinda; Teuku Farizal; Suryadi Suryadi; Mukhlizar Mukhlizar
JTERA (Jurnal Teknologi Rekayasa) Vol 11 No 1: Vol. 11 No. 1: Juni 2026
Publisher : Politeknik Sukabumi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31544/jtera.v11.i1.2026.199-208

Abstract

The increasing dependence of government institutions on internet-based services has made network security a strategic requirement for preserving data confidentiality, service availability, and public trust. This study evaluates the reliability and mitigation effectiveness of MikroTik firewall rules in securing the internet network of a local government institution against three attack scenarios: packet sniffing, port scanning, and Distributed Denial of Service (DDoS). An experimental design was employed to compare network conditions before and after the activation of firewall rules. ARP spoofing-based packet sniffing was simulated using Ettercap, port scanning was performed using Zenmap, and DDoS traffic was generated using Hping3. In contrast, Wireshark and Winbox were used to monitor network traffic and firewall responses. The initial results show that before firewall activation, the network was exposed to multidimensional risks: packet sniffing revealed DHCP packets and IP-MAC metadata, port scanning identified 8 open ports, and the DDoS attack increased CPU utilization to 100% and elevated memory usage. After the firewall rules were activated, all three attacks were successfully mitigated. Sniffing activity was limited to normal DHCP Discover traffic, port scanning no longer revealed open ports, and DDoS packets were dropped before establishing effective connections. Overall, the MikroTik firewall achieved 100% reliability within the tested scenarios, demonstrating its effectiveness in reducing traffic exposure, controlling service visibility, and maintaining system availability in a government network environment.
Tinjauan Sistematis terhadap Ensemble Semi-Supervised Network Intrusion Detection System dengan Perspektif Adversarial Robustness Sofy Fitriani; Siti Dwi Setiarini; Muhammad Irfan Nurdin; Ridwan Nurdin
JTERA (Jurnal Teknologi Rekayasa) Vol 11 No 1: Vol. 11 No. 1: Juni 2026
Publisher : Politeknik Sukabumi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31544/jtera.v11.i1.2026.209-118

Abstract

Sistem deteksi intrusi jaringan atau Network Intrusion Detection System (NIDS) berbasis pembelajaran mesin menjadi pilar utama keamanan siber modern. Namun, model berbasis data berlabel penuh sering kali terbatas dalam generalisasi dan rentan terhadap serangan adversarial. Pendekatan semi-supervised learning memanfaatkan data tak berlabel untuk memperluas cakupan pelatihan, sementara ensemble learning meningkatkan stabilitas melalui keragaman model. Penelitian ini melakukan Systematic Literature Review (SLR) untuk menganalisis tren, arah perkembangan, dan research gap pada kombinasi ensemble semi-supervised NIDS, dengan fokus utama pada robustness terhadap serangan adversarial. Proses review mengikuti pedoman PRISMA dengan sumber data dari IEEE Xplore, Scopus, SpringerLink, dan ScienceDirect untuk periode 2018–2025. Dari 96 artikel awal, 23 memenuhi kriteria inklusi. Hasil analisis menunjukkan sebagian besar penelitian menitikberatkan peningkatan akurasi (F1 > 95%), namun hanya 18% yang menguji ketahanan terhadap serangan adversarial secara eksplisit. Belum ditemukan studi yang mengintegrasikan ensemble multi-paradigma, semi-supervised learning, dan evaluasi adversarial attack secara terpadu. Kontribusi utama penelitian ini adalah penyusunan kerangka konseptual baru, Adversarial-Aware Hybrid Ensemble Framework, yang menempatkan proses adversarial design sebagai bagian integral dari pelatihan semi-supervised ensemble NIDS untuk meningkatkan keandalan sistem deteksi intrusi modern.

Page 54 of 54 | Total Record : 533

 50   51   52   53   54 

Filter by Year

2016 2026


Reset
Filter By Issues
All Issue Vol 11 No 1: Vol. 11 No. 1: Juni 2026 Vol 10 No 2: December 2025 Vol 10, No 2: Desember 2025 Vol 10, No 1: Juni 2025 Vol 9, No 2: December 2024 Vol 9, No 1: June 2024 Vol 8, No 2: December 2023 Vol 8, No 1: June 2023 Vol 7, No 2: December 2022 Vol 7, No 1: June 2022 Vol 6, No 2: December 2021 Vol 6, No 1: June 2021 Vol 5, No 2: December 2020 Vol 5, No 1: June 2020 Vol 4, No 2: December 2019 Vol 4, No 1: June 2019 Vol 3, No 2: December 2018 Vol 3, No 2: December 2018 Vol 3, No 1 (2018) Vol 3, No 1: June 2018 Vol 3, No 1: June 2018 Vol 2, No 2 (2017) Vol 2, No 2: December 2017 Vol 2, No 2: December 2017 Vol 2, No 1: June 2017 Vol 2, No 1: June 2017 Vol 1, No 1 (2016) Vol 1, No 1: December 2016 Vol 1, No 1: December 2016 More Issue