Estimation of Vessel Link-Level Sailing Time Distribution under a Connected Network

To better improve the management of maritime traffic, transportation, and port operation scheduling, it is essential to estimate the vessel sailing time distribution. Some studies focus on generating a statistical profile of sailing time, but these methods have limitations, travel times are calculated only from vessels that completed the entire study segment journey. In this paper, we propose an estimation method of vessel link-level sailing time distribution based on Kernel Density Estimation (KDE). First, we propose a novel connected maritime network construction algorithm. The connected maritime network was constructed based on massive AIS data, and the vessel link-level sailing time in the network was extracted. To obtain the law of sailing time distribution, we use the KDE method to fit and estimate the vessel link-level sailing time. Finally, experimental results of different methods on real data demonstrate that the proposed method can effectively estimate the vessel sailing time distribution.