Prediksi Trajektori Kapal Dan Identifikasi Critical Collision Zone (CCZ) Berdasarkan Automatic Identification System (AIS), Kondisi Cuaca, Dan Dimensi Kapal (Studi Kasus : Selat Sunda)

Angelina, Vira (2026) Prediksi Trajektori Kapal Dan Identifikasi Critical Collision Zone (CCZ) Berdasarkan Automatic Identification System (AIS), Kondisi Cuaca, Dan Dimensi Kapal (Studi Kasus : Selat Sunda). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221077-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Tubrukan kapal merupakan salah satu kecelakaan laut dengan tingkat risiko tinggi yang dapat menimbulkan korban jiwa, kerusakan kapal, kerugian ekonomi, serta pencemaran lingkungan. Risiko tersebut semakin meningkat pada jalur pelayaran padat seperti Selat Sunda sehingga diperlukan pendekatan prediktif untuk meningkatkan keselamatan pelayaran. Penelitian ini penting dilakukan karena model prediksi di Indonesia umumnya masih terbatas pada pemanfaatan data AIS tanpa mempertimbangkan faktor eksternal seperti cuaca dan dimensi kapal. Penelitian ini bertujuan memprediksi trajektori kapal berdasarkan data Automatic Identification System (AIS), cuaca laut, dan karakteristik kapal menggunakan algoritma Bidirectional Gated Recurrent Unit (Bi-GRU), mengidentifikasi Critical Collision Zone (CCZ) menggunakan Density-Based Spatial Clustering of Applications with Noise (DBSCAN), serta memperkirakan peluang tubrukan menggunakan simulasi Monte Carlo. Hasil penelitian menunjukkan bahwa mayoritas kapal yang melintas di Selat Sunda merupakan Bulk Carrier dengan kategori panjang kapal didominasi ukuran medium. Panjang kapal berkisar antara 8 meter hingga 360,97 meter, sedangkan negara asal keberangkatan terbanyak adalah Indonesia. Model Bi-GRU berhasil memprediksi trajektori kapal dengan baik dimana penambahan variabel dimensi kapal meningkatkan akurasi prediksi dibandingkan model tanpa dimensi kapal. Model dengan dimensi kapal menghasilkan nilai MAE sebesar 0,004318 dan MSE sebesar 0,000190, sedangkan model tanpa dimensi kapal menghasilkan MAE sebesar 0,007597 dan MSE sebesar 0,000394. Penambahan variabel dimensi kapal mampu menurunkan nilai MAE sebesar 43,22% yang menunjukkan kontribusi dimensi kapal terhadap peningkatan akurasi prediksi trajektori. Penerapan DBSCAN menghasilkan tiga cluster yang merepresentasikan CCZ, yaitu cluster 1 pada koridor pelayaran utama Selat Sunda, cluster 2 pada bagian utara Selat Sunda, dan cluster 3 pada bagian timur Selat Sunda. Berdasarkan simulasi Monte Carlo, cluster 3 memiliki peluang tubrukan tertinggi pada model dengan dimensi kapal sebesar 0,4019, diikuti cluster 1 sebesar 0,2222 dan cluster 2 sebesar 0,1578 sehingga bagian timur Selat Sunda diidentifikasi sebagai wilayah dengan tingkat risiko interaksi kapal tertinggi atau CCZ.
=================================================================================================================================
Ship collisions are one of the marine accidents with a high level of risk that can cause casualties, ship damage, economic losses, and environmental pollution. This risk is increasing on congested shipping lanes such as the Sunda Strait, so a predictive approach is needed to improve shipping safety. This research is important because prediction models in Indonesia are generally still limited to the use of AIS data without considering external factors such as weather and ship dimensions. This study aims to predict ship trajectory based on Automatic Identification System (AIS) data, sea weather, and ship characteristics using Bidirectional Gated Recurrent Unit (Bi-GRU) algorithm, identify Critical Collision Zone (CCZ) using Density-Based Spatial Clustering of Applications with Noise (DBSCAN), and estimate the chance of collision using Monte Carlo simulation. The results of the study show that the majority of ships passing through the Sunda Strait are Bulk Carriers with the category of ship length dominated by medium size. The length of the ship ranges from 8 meters to 360.97 meters, while the country of origin of the most departures is Indonesia. The Bi-GRU model managed to predict the trajectory of the ship well where the addition of ship dimension variables increased the prediction accuracy compared to the model without ship dimensions. Models with ship dimensions produce an MAE value of 0.004318 and MSE of 0.000190, while models without ship dimensions produce an MAE of 0.007597 and MSE of 0.000394. The addition of the ship dimension variable was able to reduce the MAE value by 43.22% which shows the contribution of ship dimensions to improving the accuracy of trajectory predictions. The application of DBSCAN resulted in three clusters representing CCZ, namely cluster 1 in the main shipping corridor of the Sunda Strait, cluster 2 in the northern part of the Sunda Strait, and cluster 3 in the eastern part of the Sunda Strait. Based on Monte Carlo simulations, cluster 3 has the highest chance of collision in the model with a ship dimension of 0.4019, followed by cluster 1 of 0.2222 and cluster 2 of 0.1578 so that the eastern part of the Sunda Strait is identified as the area with the highest level of risk of ship interaction or CCZ.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Automatic Identification System (AIS), Critical Collision Zone (CCZ), Cuaca, Dimensi, Trajectory Kapal, Tubrukan Kapal, Ship Dimensions, Ship Trajectory, Weather
Subjects:	Q Science Q Science > Q Science (General) Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Vira Angelina
Date Deposited:	03 Aug 2026 04:11
Last Modified:	03 Aug 2026 04:11
URI:	http://repository.its.ac.id/id/eprint/142044

Actions (login required)

View Item