The use of AIS historical data provides a great opportunity for traffic streams analysis and also for discovering related processes that govern the traffic in confined areas. To model traffic streams, a statistical analysis is necessary to determine the most optimized spatial distributions that can represent the traffic with its evolution throughout the analyzed area. In this paper, a framework is designed where AIS historical data is utilized to explore the existing maneuvering patterns in the area and model the parameters describing the traffic flow. This work is the first step leading to the creation of a macroscopic model of traffic flow through statistical inference analysis of vessel traffic streams within a selected area.

The results of this work shed light on the behavior of vessel traffic streams at a macroscopic level (group vessel level), through statistical modeling of the traffic flow into distribution functions. This work sets the basis for exploring the behavior of vessel traffic streams at specific locations and times; and eventually predicts and assesses the evolution of future traffic through simulation.