TRAJECTORY ESTIMATION FOR MARITIME VESSELS USING LSTM NETWORKS ON LARGE-SCALE AIS DATA

Tóm tắt

Accurate trajectory estimation of maritime vessels plays a pivotal role in ensuring navigational safety, enhancing operational efficiency, and facilitating real-time decision-making in maritime surveillance systems. With the increasing availability of Automatic Identification System (AIS) data, deep learning models have shown considerable promise in modeling the complex spatio-temporal dynamics of vessel movement. This study proposes a Long Short-Term Memory (LSTM) network-based approach to predict the future trajectories of vessels using large-scale AIS datasets. We preprocess and segment the AIS records to construct structured sequences for training, and apply an LSTM architecture tailored to capture both short-term and long-term dependencies in the trajectory data. The results demonstrate that our approach outperforms traditional methods in terms of accuracy and robustness, especially in diverse maritime environments. This research provides a solid foundation for deploying real-time vessel movement forecasting systems and supports maritime anomaly detection and traffic management applications.