Article Per Year (5 Year)
p-Index From 2021 - 2026
1.066
P-Index
This Author published in this journals
All Journal Jurnal Teknik Elektro SITEKIN: Jurnal Sains, Teknologi dan Industri JITROKOM International Journal of Artificial Intelligence Research E-JOINT (Electronica and Electrical Journal Of Innovation Technology) Prosiding Seminar Nasional Hasil Penelitian Sains, Teknik dan Aplikasi Industri Fakultas Teknik Universitas Lampung Jurnal Rekayasa elektrika
Melvi .
Universitas Lampung
Chemical Engineering, Chemistry & Bioengineering Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Electrical & Electronics Engineering Energy Engineering Industrial & Manufacturing Engineering Other

Published : 12 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 12 Documents
Search

 1   2 
Sistem Peringatan Dini Untuk Pengendalian Pembatasan Jarak Fisik Dengan Metode RSSI Menggunakan Modul Wemos D1 Mini Noni Mastiana; Ardian Ulvan; Melvi Ulvan
Jurnal Rekayasa Elektrika Vol 17, No 4 (2021)
Publisher : Universitas Syiah Kuala

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.17529/jre.v17i4.21650

Abstract

Covid-19 was first reported in Indonesia in March 2020. A month later the confirmed cases had reached 1528 cases and the death toll was 136 cases. Distance restrictions, also known as physical and social distance, are one of the government's efforts to reduce the rate of increase in the number of positive Covid-19 patients. However, this policy is still not fully effective, due to the low level of community compliance. When personal awareness has not been awakened, the use of technology in the context of early warning to maintain distance is an effective solution. This article discusses the use of received signal strength indicator (RSSI) from the Wemos D1 Mini module as a model of an early warning system device to anticipate unsafe distances when someone is in a crowd. The use of RSSI in this study shows that the device made works well, where the alarm is active when the distance between devices is 1 meter. The system performance was analyzed by calculating and measuring the average RSSI error value of -34.46 dBm and the average distance error of 0.26 meters. Overall, the early warning system using this method can be used properly to estimate physical distance restrictions during a pandemic.  
DESAIN ALGORITMA SELF-SUPERVISED VISION TRANSFORMER LEARNING UNTUK PERANGKAT LOW ENERGI IOE PERINGATAN DINI TSUNAMI: Aryanto , Melvi , A Ulvan, E Komalasari Jurusan Teknik Elektro, Universitas Lampung, Jl. Prof Soemantri B. No 1. Bandar Lampung, Lampung Aryanto Aryanto; Melvi Melvi; Ardian Ulvan; Endah Komalasari
Prosiding Seminar Nasional Ilmu Teknik Dan Aplikasi Industri Fakultas Teknik Universitas Lampung Vol. 8 (2025): Prosiding Seminar Nasional Ilmu Teknik dan Aplikasi Industri (SINTA) 2025
Publisher : Fakultas Teknik Universitas Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Peningkatan intensitas bencana tsunami di wilayah pesisir menuntut hadirnya sistem peringatan dini yang lebih akurat, adaptif, dan hemat energi. Penelitian ini mengusulkan desain algoritma Self-Supervised Vision Transformer Learning (SSL-ViT) yang dioptimalkan untuk perangkat Internet of Everything (IoE) berdaya rendah. Berbeda dengan pendekatan supervised konvensional yang membutuhkan label data dalam jumlah besar, SSL-ViT memanfaatkan pembelajaran representasi visual secara mandiri melalui pretext task seperti masking patch prediction dan contrastive learning, sehingga mampu belajar dari data citra pesisir tanpa anotasi. Arsitektur Vision Transformer dimodifikasi dengan mekanisme lightweight attention dan model pruning guna menekan konsumsi memori dan komputasi. Sistem ini diintegrasikan dengan jaringan perangkat IoE untuk melakukan deteksi anomali visual pada permukaan laut secara real-time. Hasil pengujian awal menunjukkan bahwa algoritma SSL-ViT mencapai akurasi deteksi anomali sebesar >85% dengan latensi inferensi <300 ms dan penghematan energi hingga 40% dibandingkan model CNN konvensional. Temuan ini membuktikan bahwa kombinasi self-supervised learning dan optimasi arsitektur transformer pada perangkat low-energy IoE mampu menjadi fondasi sistem peringatan dini tsunami yang efisien, skalabel, dan siap diimplementasikan di kawasan pesisir berinfrastruktur terbatas.
Co-Authors Ardian Ulfan Ardian Ulvan Ardian Ulvan Ardian Ulvan Ardian Ulvan Aryanto Aryanto Bagus Wahyu Pratomo Dwi Satrio Wicaksono, Dwi Satrio Handrio Handrio Hendra Widodo Irfan Mirda Komalasari, Endah Mardiana Mardiana Mardiyah Azzahra Muvianto, Cahyo Mustiko Okta Noni Mastiana Nora Aditiyan Nur Fadillah Pattarakorn Sajjaporn Nuttachot Robert Bestak Yetti Yuniati