Long-term ship trajectory prediction method

Abstract

The invention discloses a long-term ship trajectory prediction method, which comprises the following steps: S1, dividing a predicted water area space according to regular grids, and calculating a ship trajectory distribution density map; s2, an initial state domain is generated according to the initial state points of the ship to be predicted, and a starting state point set is obtained in the initial state domain through similarity search; s3, calculating a point position set after the starting state point set delta t; s4, clustering the point position set by using a density clustering algorithm, extracting a clustering result main cluster, and calculating representative points to obtain a representative point set Rc; s5, calculating a density value corresponding to each point of Rc in the ship trajectory distribution density map; and S6, calculating current-round prediction state point information according to the Rc. According to the method, the defects that an existing ship trajectory prediction method is not high enough in precision and only can achieve single-trajectory prediction are overcome, trajectory prediction of multiple navigation modes is achieved in a mode of flexibly adjusting the threshold value, and important information such as the navigation speed and the navigation direction at different moments is predicted at the same time.

Description

technical field [0001] The invention relates to a ship track prediction method, in particular to a long-term ship track prediction method. Background technique [0002] Ship trajectory prediction is an important part of water traffic situation awareness and a key technology for realizing intelligent shipping. From the perspective of forecasting time, ship trajectory forecasting can be divided into two fields: short-term forecasting and long-term forecasting. Short-term prediction is often used for ship collision avoidance. Unlike road traffic, due to the huge inertia of ships, short-term forecasting is relatively easy to achieve. The traditional constant velocity forecasting model (Constant Velocity Model, CVM) is relatively mature and has been applied to commercial products. Among other common prediction methods, Chinese Patent Document No. CN110070565B discloses a ship trajectory prediction method based on image superposition, which uses Kalman filter technology to estim...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a long-term ship trajectory prediction method, which improves the defects of the existing ship trajectory prediction method that the accuracy is not high enough and can only realize single trajectory prediction, and realize the trajectory of multiple navigation modes by flexibly adjusting the threshold Prediction, taking into account the forecasting of important information such as speed and course at different times, is suitable for application fields such as inland river maritime supervision, intelligent shipping information push, and port arrival time estimation

Images