Improved ship trajectory prediction method and device based on recurrent neural network

Abstract

The invention discloses an improved ship trajectory prediction method and device based on a recurrent neural network. At present, offshore ships are dense, the traffic environment is complex, and ship automatic identification system (AIS) data have the following characteristics: (1) a large number of anchoring tracks exist in a sea area; (2) part of non-anchoring tracks have abnormal acute-angle bends; (3) the behavior difference of the ships in different legs is large. As these characteristics can reduce the accuracy of trajectory prediction, the invention provides an improved ship trajectory prediction model based on a recurrent neural network, and the method comprises the following steps: (1) proposing an anchor trajectory elimination algorithm to eliminate an anchor trajectory; (2) proposing a probability-based trajectory repair algorithm to repair the acute angle bends; (3) designing a two-stage ship trajectory flow clustering algorithm to distinguish ship behaviors; (4) building a deep-layer two-way gate circulation unit (GRU) model. The improved ship trajectory prediction model provided by the invention has higher prediction precision, and has a certain reference value in ship trajectory prediction in an offshore area.

Description

technical field [0001] The invention belongs to the field of deep learning technology and data mining, and in particular relates to an improved fishing boat trajectory prediction method based on a cyclic neural network and a device thereof. Background technique [0002] In offshore port areas with high ship density and complex traffic environment, how to effectively predict fishing boat trajectories is a huge challenge. On the one hand, unlike vehicle or pedestrian trajectory prediction, maritime moving targets are not constrained by geometric structures, and their motion patterns are more complex than those of land targets. According to the "International Regulations for Preventing Collisions at Sea" (COLREGS), the navigation of ships is more dependent on experience than standard road traffic rules, and it is difficult to quantify and analyze. But on the other hand, the ship's automatic identification system (AIS) data contains the potential movement patterns of the ship, ...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to overcome the problem of low prediction accuracy caused by the existence of outliers in the trajectory data of fishing boats and the large differences in ship behavior characteristics, and to provide an improved fishing boat trajectory prediction method based on cyclic neural network

Images