Smart Buoy Model and Early Warning System (EWS) Based on Ship-Buoy Domain Threshold

Sidik, Hamdan (2026) Smart Buoy Model and Early Warning System (EWS) Based on Ship-Buoy Domain Threshold. Other thesis, Institut Teknologi Sepuluh Nopember (ITS).

Text 5019221030-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Maritime safety relies heavily on the reliability of aids to navigation (AtoN) specifically buoys in accurately providing positional data and information on the marine environment. Issues affecting buoy based AtoN include positional shifts caused by environmental forces, mooring system failures, asset loss, and potential ship collisions, all of which can compromise navigational functions. Consequently, there is a need to develop a technology-based monitoring system capable of real-time surveillance of buoy conditions and vessel activity, while also providing early warnings regarding high-risk situations. This study proposes a "smart buoy" system integrating the Automatic Identification System (AIS) and an Early Warning System (EWS). The research methodology begins with an AIS-based grid density analysis to quantify vessel traffic intensity around the buoys; this data is transformed into vessel presence probabilities using a Poisson distribution approach and compared against a bivariate Gaussian distribution-based ship domain model to define spatial occupancy zones. Threshold probability (Pth) and threshold distance (Dth) values are determined via Receiver Operating Characteristic (ROC) analysis optimizing the difference between the True Positive Rate (TPR) and False Positive Rate (FPR) to establish an optimal early warning radius for each buoy. The validity of this risk classification model is subsequently tested using the Area Under the Curve (AUC) metric from the ROC curve to ensure the system's ability to accurately discriminate between high-risk and low-risk zones. Model analysis yielded compatible results, with simulations conducted in accordance with IALA G1039 standards. Warning limits were established through a Ship-Buoy Domain analysis that accounted for vessel movement characteristics, safe distances from the buoy, and traffic patterns derived from AIS data. Calculations were performed for three buoys KP51, KP54, and KP46 resulting in threshold values ranging from 114 m to 628 m. The system is designed to detect when a vessel enters the threshold area and generate early warnings that serve as supporting information for operational decision-making. The implementation of AIS and EWS based smart buoys is expected to enhance the reliability of aids to navigation, accelerate the identification of potential hazards, and support a maritime monitoring system that is more effective, adaptive, and data-driven, thereby improving navigational safety.
========================================================================================================================
Keselamatan pelayaran sangat dipengaruhi oleh keandalan sarana bantu navigasi dalam memberikan informasi posisi dan kondisi lingkungan perairan secara akurat. Salah satu permasalahan yang dapat terjadi pada sarana bantu navigasi berupa buoy adalah perubahan posisi akibat gaya lingkungan, kerusakan sistem tambat, kehilangan aset, maupun potensi tubrukan dengan kapal yang dapat mengganggu fungsi navigasi. Oleh karena itu, diperlukan pengembangan sistem monitoring berbasis teknologi yang mampu melakukan pemantauan kondisi buoy dan aktivitas kapal secara real-time serta memberikan peringatan dini terhadap kondisi berisiko. Sistem smart buoy dengan integrasi Automatic Identification System (AIS) dan Early Warning System (EWS). Metode penelitian ini diawali dengan analisis kepadatan grid (grid density) berbasis data AIS untuk mengkuantifikasi intensitas lalu lintas kapal di sekitar buoy, yang kemudian ditransformasikan ke dalam bentuk probabilitas kehadiran kapal menggunakan pendekatan distribusi Poisson, serta dibandingkan dengan model ship domain berbasis distribusi Gaussian bivariat untuk menentukan zona okupansi spasial kapal. Nilai probabilitas ambang batas (Pth) dan jarak ambang batas (Dth) selanjutnya ditentukan melalui analisis Receiver Operating Characteristic (ROC) dengan mengoptimalkan selisih antara True Positive Rate (TPR) dan False Positive Rate (FPR), sehingga diperoleh radius peringatan dini yang optimal di sekitar masing-masing buoy. Validitas model klasifikasi risiko ini kemudian diuji menggunakan nilai Area Under Curve (AUC) dari kurva ROC untuk memastikan kemampuan diskriminasi sistem dalam membedakan zona berisiko dan tidak berisiko secara akurat.. Berdasarkan hasil analisis model didapatkan hasil model yang kompatibel, dimana simulasi dilakukan berdasarkan IALA G1039. Penentuan batas peringatan dilakukan melalui analisis Ship-Buoy Domain dengan mempertimbangkan karakteristik pergerakan kapal, jarak aman terhadap buoy, serta pola lalu lintas kapal berdasarkan data AIS. Hasil perhitungan dilakukan untuk 3 buoy yaitu KP51, KP54, dan KP46. Nilai threshold yang didapatkan adalah pada rentang 114 m sampai 628 m. Sistem ini dirancang untuk mendeteksi kondisi ketika kapal memasuki area threshold dan menghasilkan peringatan dini yang dapat digunakan sebagai informasi pendukung dalam pengambilan keputusan operasional. Penerapan smart buoy berbasis AIS dan EWS diharapkan mampu meningkatkan keandalan sarana bantu navigasi, mempercepat proses identifikasi potensi bahaya, serta mendukung sistem monitoring maritim yang lebih efektif, adaptif, dan berbasis data guna meningkatkan keselamatan pelayaran.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	AI; AUC; Buoy; Domain; Smart buoy; Threshold AIS; AUC; Ambang Batas; Domain Pelampung; Pelampung Pintar
Subjects:	V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering > VM293 Shipping--Indonesia--Safety measures
Divisions:	Faculty of Architecture, Design, and Planning
Depositing User:	Hamdan Sidik
Date Deposited:	03 Aug 2026 07:46
Last Modified:	03 Aug 2026 07:46
URI:	http://repository.its.ac.id/id/eprint/142186

Actions (login required)

View Item