Analysis of The ship's automatic identification system information and models for constructing spatio-temporal algorithms for predicting vessel traffic

The accuracy of solving navigation safety problems is determined by two factors: the accuracy of the initial data and the degree of adequacy of the mathematical models used. The ship's automatic identification system (AIS) has clear advantages over radar/ARPA in terms of data accuracy for predicting vessel movements. This data can be used to build Decision Support Systems for Maritime Autonomous Vessels (MANS) to track traffic routes and predict collisions in real time. AIS data makes it possible to describe movement in coordinate form, in the form of a onedimensional time series of observations of the vessel's location (x, y), identify trends in changes in movement dynamics (increase, decrease or constancy) and observe a series of periods of ship movement with different dynamics, when stationary time series change to non-stationary and vice versa. A change in movement modes is accompanied by a disorder in the time series, i.e. detection of changes in the properties of the observed parametric series occurring at an unknown point in time. Determining the moment of disorder due to the inertia of the vessel is a rather difficult task. To solve it, the method of finding the likelihood ratio of the threshold value of some decision rule after crossing the corresponding values of the time series (x,y) can be used. There is no universal method for forecasting time series; each method finds its application for different types of movement and different types of time series. Time series (x,y) can be used in various models: physical, machine learning and hybrid models. Today there is no implementation of this or that model in real time. This problem can be solved by developing hybrid models, which are a combination of spatiotemporal algorithms for predicting the movement of vessels at sea to obtain better vessel movement estimation results. Keywords: autonomous navigation, automatic identification system, real-time monitoring and forecasting of vessel movement, preliminary analysis of AIS data, vessel movement models.