Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/163996
Title: Robot Cerdas ROV (Remotely Operated Vehicle) Berbasis Kendali Jarak Jauh Untuk Pengamatan Kehidupan Ikan Berbantukan Kamera 180?
Other Titles: Remote Control-Based ROV Intelligent Robot For Fish Life Observation Assisted 180? Camera
Authors: Siskandar, Ridwan
Nisa, Afifah Rodhiyatun
Issue Date: 2025
Publisher: IPB University
Abstract: Pengembangan robot cerdas ROV (Remotely Operated Vehicle) mampu mendeteksi kehidupan ikan berdasarkan kondisi lingkungan air dengan menggunakan kamera 180? yang dilengkapi dengan nirkabel kendali jarak jauh. Robot ini dirancang untuk parairan air tawar, yang dikendalikan dengan remote control. Prediksi akurasi data sensor kualitas air menggunakan metode logika fuzzy ANFIS. Dimensi robot adalah lebar 40 cm, panjang 46 cm, dan diameter acrylic tabung sebesar 10 cm dengan tinggi 40 cm, dengan berat 4 kg. Robot dapat bergerak sejauh 25 cm, dengan kedalaman air hingga 30-40 cm. Pengamatan ikan akan ditampilkan pada website monitoring. Pengujian sensor polutan air dilakukan pada pagi, siang, dan malam di 5 kolam yang berbeda kemudian dilakukan evaluasi kinerja model dengan menggunakan Random Forest. Kemudian pada pengamatan kehidupan ikan menggunakan model YOLOv8l dengan akurasi mAP50(B) sebesar 0,859; mAP50-95(B) sebesar 0,68; precision(B) sebesar 0.904, dan recall sebesar 0,779 yang menunjukkan deteksi ikan hampir sempurna.
The development of an intelligent ROV (Remotely Operated Vehicle) robot is able to detect fish life in real-time based on water environmental conditions using a 180? camera equipped with a wireless remote control. This robot is designed for freshwater, which is controlled by remote control. Prediction of water quality sensor data accuracy using the ANFIS fuzzy logic method. The dimensions of the robot are 40 cm wide, 46 cm long, and an acrylic tube diameter of 10 cm with a height of 40 cm, weighing 4 kg. The robot can move as far as 25 cm, with a water depth of up to 30-40 cm. Fish observations will be displayed on the monitoring website. Water pollutant sensor testing was carried out in the morning, afternoon, and evening in 5 different ponds, then an evaluation of model performance was carried out using Random Forest. Then in the observation of fish life using the YOLOv8l model with an accuracy of mAP50 (B) of 0.859, mAP50-95 (B) of 0.68, precision (B) of 0.904, and recall of 0.779 which indicates almost perfect fish detection.
URI: http://repository.ipb.ac.id/handle/123456789/163996
Appears in Collections:UT - Computer Engineering Tehcnology

Files in This Item:
File Description SizeFormat 
cover_J0304211119_e95747e0a6914eaeb72d3ca3ced98efc.pdfCover1.09 MBAdobe PDFView/Open
fulltext_J0304211119_13f6da27f98e435cbce4348d80baafda.pdf
  Restricted Access
Fulltext5.3 MBAdobe PDFView/Open
lampiran_J0304211119_124fd48c3a7b4e42a1b6d41b9cc670e8.pdf
  Restricted Access
Lampiran1.89 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.