Passive sonar and AIS track fusion method based on optimal linear matching and track prediction

In the detection and data fusion of passive sonar targets, there is often a challenge of associating target trajectories with non-acoustic information such as AIS. To address this issue, this paper proposes a probabilistic correlation method for passive sonar target trajectories. This method aims at the problem of passive sonar target detection and recognition when the AIS information of the Unmanned Underwater Vehicle cannot be received in real time during underwater operation. It utilizes a recurrent neural network model to predict the ship trajectory after the underwater unmanned vehicle enters the water. Then, the ellipsoid model is used to calculate the distance between the two tracks on the earth surface, so as to obtain the probability correlation coefficient between the tracks, and then the linear matching method is used to improve the accuracy of trajectory correlation. Experimental results demonstrate that under different values of virtual variables, this method successfully achieves accurate matching between AIS information and passive sonar detected target trajectories, providing support for enhancing the accuracy and reliability of passive sonar target detection.