A UUV target detection method based on Informer

Aiming at the problem of slow convergence and high environmental complexity of traditional bearing-only algorithm when underwater platform detects targets, Informer time series prediction model is applied to deeply mine and extract features of the intrinsic relationship between target observation data and vehicle information and predict the motion elements of the target in future time series. The experimental data sets of thirty kinds of situations are selected to conduct comparative experiments and target track prediction. Experiments show that the Informer model has achieved satisfactory results in predicting target motion elements in complex environments and can provide element reference and judgment basis for traditional bearing-only target tracking algorithms.