PEMETAAN VOLATILE ORGANIC COMPOUND (VOC) MENGGUNAKAN UNMANNED SURFACE VEHICLE (USV) BERBASIS MACHINE LEARNING DENGAN SENSOR GAS ARRAY

Gibran, Muhammad Kaisar (2024) PEMETAAN VOLATILE ORGANIC COMPOUND (VOC) MENGGUNAKAN UNMANNED SURFACE VEHICLE (USV) BERBASIS MACHINE LEARNING DENGAN SENSOR GAS ARRAY. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5019201158-Undergraduate-JointDegree_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (4MB) | Request a copy

Abstract

Pollutants in water-mediated environments is crucial matter in the marine environment. To help monitoring operation efficiently, Unmanned Surface Vehicles (USVs) emerged as a promising solution for real-time sampling and monitoring activities. Recent technological advancement has improved the practical applications of USVs, enabling them to be equipped with sensors for pollutant detection and GPS for tracing and mapping sources. This integration allows for real-time data analysis using Machine Learning (ML) algorithms, enhancing the identification of gas patterns. In this study, a USV equipped with multiple gas sensors was used to perform experiments around a lake. The USV detected vapors from alcohol and various fuels that could generate enough vapor gas for training in Machine Learning. The Mission Planner software enabled the USV to autonomously maneuver and map the route while detecting and mapping gas concentrations. The results showed that the ML model successfully identified and differentiated gases, with Pertalite and Biosolar achieving the highest accuracy. Alcohol was identified with moderate accuracy, while the remaining gases had less than 40% accuracy, likely due to wind turbulence affecting detecting the gas vapor.
========================================================================================================================
Pencemaran di lingkungan perairan merupakan masalah krusial dalam lingkungan laut. Untuk membantu operasi pemantauan secara efisien, Unmanned Surface Vehicle (USV) muncul sebagai solusi menjanjikan untuk pengambilan sampel dan pemantauan secara real-time. Kemajuan teknologi terbaru telah meningkatkan aplikasi praktis USV, memungkinkan mereka dilengkapi dengan sensor untuk deteksi polutan dan GPS untuk pelacakan serta pemetaan sumber. Integrasi ini memungkinkan analisis data real-time menggunakan algoritma Machine Learning (ML), meningkatkan identifikasi pola gas. Dalam studi ini, sebuah USV yang dilengkapi dengan beberapa sensor gas digunakan untuk melakukan eksperimen di sekitar danau. USV mendeteksi uap dari alkohol dan berbagai bahan bakar yang dapat menghasilkan cukup gas uap untuk pelatihan dalam Machine Learning. Perangkat lunak Mission Planner memungkinkan USV untuk bermanuver secara otonom dan memetakan rute sambil mendeteksi dan memetakan konsentrasi gas. Hasilnya menunjukkan bahwa model ML berhasil mengidentifikasi dan membedakan gas, dengan Pertalite dan Biosolar mencapai akurasi tertinggi. Alkohol diidentifikasi dengan akurasi sedang, sementara gas lainnya memiliki akurasi kurang dari 40%, kemungkinan disebabkan oleh turbulensi angin yang mempengaruhi deteksi gas uap.
Kata

Item Type:	Thesis (Other)
Uncontrolled Keywords:	USV, VOC, Gas Detection, Vapor Detection, Gas Sensors, Machine Learning, GPS Mapping. USV, VOC, Deteksi Gas, Deteksi Vapor, Sensor Gas, Machine Learning, Pemetaan GPS
Subjects:	V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM365 Remote submersibles. Autonomous vehicles.
Divisions:	Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis
Depositing User:	Muhammad Kaisar Gibran
Date Deposited:	09 Aug 2024 07:03
Last Modified:	09 Aug 2024 07:03
URI:	http://repository.its.ac.id/id/eprint/113234

Actions (login required)

View Item