Prediction of vessels locations and maritime traffic using similarity measurement of trajectory

Maritime traffic prediction is a crucial task for increasing the efficiency of port operations and safety, especially in congested regions. A huge amount of automatic identification system (AIS) data is constantly transmitting from vessels to receivers that contain information about vessels' movements and characteristics. These historical AIS data can be utilized in movement analyses of vessels. This paper proposes a novel point-based model for location and traffic prediction using vessels' trajectories adapted from AIS measures. The location prediction procedure is setup based on similarity analysis of historical AIS data. The model is applied to a real dataset of hundreds of vessels' trajectories in the Strait of Georgia, USA. The correlation results of 0.9976, 0.9887, and 0.9794 for the next 10, 20, and 30 minutes, respectively, imply sufficient correspondence between predicted and actual coordinates. The traffic prediction procedure considers the probability of the appearance of new vessels inside an area of interest (AoI) at different time intervals. The Sorenson similarity index (SSI) is used to measure the accuracy of the traffic prediction model. The SSIs for time intervals of 10, 20, and 30 minutes are 70%, 66%, and 59%, respectively, which show the robustness of the model to predict hot spots inside the AoI.

Automated summary

The study introduces a novel maritime traffic prediction model based on AIS data, which is successfully applied to a real dataset in the Strait of Georgia, USA, demonstrating high accuracy and prediction capabilities. The model, utilizing similarity analysis of historical AIS data, effectively predicts vessel locations and traffic conditions.