This project focuses on analyzing vessel trajectories using AIS (Automatic Identification System) data, aiming to predict trajectories that will help in collision avoidance, destination prediction etc with data cleaning, clustering, and forecasting techniques. The project is conducted under the guidance of Dr. Luo Wei at GeoSpatialX Lab, National University of Singapore.
- Initial Dataset Used:
AIS_2022_01_01 - Dataset Procured From: Marine Cadastre AIS Data Portal
- Dataset Size: 7,239,758 × 17 (7,239,758 data points and 17 variables)
- Vessel Type: Vessel Type ‘80’ was selected, which are tanker vessels.
- Data Cleaning:
- Remove invalid data points, such as incorrect MMSI entries i.e., no. of digits in MMSI greater or lesser than 9.
- Filter latitude values to the range of 20–25 degrees while keeping longitude unrestricted to allow longer trajectory length and narrow down the area of focus for easier analysis.
- Clustering:
- Applied the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm to identify clusters and outliers.
- Resulted in 18 clusters, with 1 identified outlier removed.
- Trajectory Prediction:
- Used the VARMAX (Vector Autoregressive Moving-Average with Exogenous Regressors) model for trajectory prediction.
- Predictions were validated by comparing them with known trajectory points.
- Data cleaning and filtering ensure high-quality inputs for clustering and forecasting.
- DBSCAN effectively segmented the dataset, identifying clusters and removing noise.
- VARMAX provided robust trajectory predictions when trained on sufficient high-quality data.
- Perform advanced trajectory prediction for individual vessels using multiple models.
- Expand the dataset by incorporating additional AIS datasets (e.g., AIS_01_02, AIS_03).
- Develop dynamic and interactive visualizations for the predictions using React.js.
- Further optimize accuracy and computational efficiency of the prediction models.