Verification of neural approach to radar-AIS tracks association for maneuvering targets based on kinematic spatial information

The paper undertakes a problem of tracking radar and AIS fusion in maritime navigation. The focus is laid on target association process, which is the basis for track-to-track fusion. The analysis provides neural approach to association process in which neural network classifies if the association should be stated or not. The paper includes theoretical basics of the proposed approach, as well as numerical experiment in which the performance of various neural networks for association is analyzed. Research scenario includes course maneuver of the target and the conclusions confirms promising results obtained with the use of neural networks.