Perancangan Sistem Kontrol Propulsi USV Berbasis Navigasi Visual Pada Kondisi Malam Hari

Ma'ana, Syauqy Mahabbatillah (2026) Perancangan Sistem Kontrol Propulsi USV Berbasis Navigasi Visual Pada Kondisi Malam Hari. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5019221072-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Operasi Search and Rescue (SAR) maritim pada malam hari menghadapi tantangan besar akibat keterbatasan visibilitas dan kondisi lingkungan yang dinamis. Autonomous Surface Vehicle (USV) berpotensi mendukung operasi tersebut, namun keberhasilan manuver otonom tidak hanya ditentukan oleh kemampuan deteksi visual, melainkan juga oleh respon dan kestabilan sistem kontrol propulsinya. Penelitian ini merancang dan menguji sistem kontrol propulsi USV berbasis PID yang memanfaatkan input navigasi visual berupa sudut bearing target hasil deteksi kamera night vision terhadap reflektor pada life jacket korban. Arsitektur navigasi dirancang dengan pendekatan distributed control melalui protokol UART dan I2C, terdiri atas dua mode operasi yang bertransisi secara otonom. Mode Search yang memanfaatkan sinkronisasi GPS (ATGM336H) dan IMU untuk navigasi global berbasis Line of Sight (LOS) Guidance, dan Mode Approach yang mengalihkan referensi arah menjadi navigasi lokal begitu kamera night vision mendeteksi pantulan reflektor target. Pemrosesan visi dilakukan pada Raspberry Pi, sementara eksekusi kendali PID dan differential thrust dua thruster BLDC dijalankan pada ESP32. Pengujian lapangan skala prototype menunjukkan sistem mampu meredam error heading dengan RMSE 1,11° hingga 1,35° dan steady state error 0,47°, menghentikan USV pada jarak akhir 0,41 hingga 0,73 meter dari target, serta menghasilkan respon dinamis yang memenuhi seluruh batas desain (rise time 2,45 s, settling time 19,86 s, overshoot maksimum 9,47°). Hasil ini membuktikan bahwa keluaran navigasi visual dapat digunakan sebagai basis tunggal kendali arah otomatis USV pada kondisi malam hari, sekaligus menunjukkan bahwa pemisahan pemrosesan visi dan eksekusi kendali antar mikrokontroler efektif menjaga kestabilan tanpa menimbulkan delay komunikasi yang berarti.
=======================================================================================================================================
Maritime Search and Rescue (SAR) operations conducted at night face significant challenges due to limited visibility and dynamic environmental conditions. Autonomous Surface Vehicles (USVs) have the potential to support such operations, however, successful autonomous maneuvering depends not only on visual detection capability but also on the response and stability of the propulsion control system. This study designs and tests a PID based USV propulsion control system that utilizes visual navigation input in the form of target bearing angle derived from night vision camera detection of reflectors on a survivor's life jacket. The navigation architecture is designed using a distributed control approach via UART and I2C protocols, consisting of two operation modes that transition autonomously. Search Mode, which synchronizes GPS (ATGM336H) and IMU data for global navigation based on Line of Sight (LOS) Guidance, and Approach Mode, which switches the heading reference to local navigation once the night vision camera detects the target's reflector. Vision processing is performed on a Raspberry Pi, while PID control execution and dual BLDC thruster differential thrust are handled by an ESP32. Prototype scale field testing shows that the system suppresses heading error with an RMSE of 1.11° to 1.35° and a steady state error of 0.47°, stops the USV within a final distance of 0.41 to 0.73 meters from the target, and produces a dynamic response meeting all design limits (rise time 2.45 s, settling time 19.86 s, maximum overshoot 9.47°). These results demonstrate that visual navigation output can serve as a standalone basis for automatic heading control of a USV under nighttime conditions, while also showing that separating vision processing and control execution across microcontrollers effectively maintains stability without significant communication delay.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Night Vision , PID Controller, SAR, USV, Vision Based Navigation, Night Vision , PID Controller, SAR, USV, Vision Based Navigation
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors V Naval Science > VK > VK555 Navigation.
Divisions:	Faculty of Marine Technology (MARTECH) > Marine Engineering > 36202-(S1) Undergraduate Thesis
Depositing User:	Syauqy Mahabbatillah Ma`ana
Date Deposited:	07 Aug 2026 12:54
Last Modified:	07 Aug 2026 12:54
URI:	http://repository.its.ac.id/id/eprint/142771

Actions (login required)

View Item