2026-10-10T11:26:25Z https://www.scidb.cn/oai

10.57760/sciencedb.011zm 2026-09-20T15:18:16Z

2026-09-20 Processed AIS Vessel Trajectory Dataset for Multi-Horizon Prediction in Tampa Bay and the Port of New York (2024) doi:10.57760/sciencedb.011zm en This dataset contains processed Automatic Identification System (AIS) vessel trajectories used for multi-horizon vessel trajectory prediction experiments in Tampa Bay and the Port of New York. The data were derived from 2024 AIS records and processed to construct continuous vessel trajectory sequences suitable for data-driven trajectory forecasting.The dataset was generated through a series of preprocessing procedures, including geographical filtering, invalid and abnormal record removal, trajectory sorting and segmentation, temporal processing, and construction of observation–prediction sequence pairs. The resulting trajectory samples were organized into fixed training, validation, and test subsets to ensure reproducibility of the experimental results.The dataset contains the vessel motion attributes required by the prediction models and the corresponding future trajectory targets. It supports short- and long-horizon vessel trajectory forecasting experiments, including the 20-minute and 60-minute prediction settings evaluated in the associated study.The two geographical subsets, Tampa Bay and the Port of New York, are provided separately to facilitate evaluation across different maritime traffic environments. Documentation describing the data organization, attribute definitions, preprocessing procedures, and dataset splits is included with the deposited files.The dataset is intended to support research on AIS-based vessel trajectory prediction, maritime traffic analysis, autonomous navigation, and data-driven maritime intelligent systems. Automatic Identification System (AIS); AIS; vessel; trajectory Tianhe Yang PUBLIC https://opendatacommons.org/licenses/by/1-0/ dataset Science Data Bank