Research on Air Target Maneuver Recognition Based on LSTM Network

Abstract

Aiming at the current fact of low recognition rate and poor anti-noise performance of the existing air target maneuver recognition algorithms, a method of target maneuver recognition based on LSTM network was studied. Input of the LSTM network is getting by a series of preprocessing on the original track, including eliminating outliers and interpolation, and reconstructing the track. After training and recognition, the maneuver type recognition result of the target to be measured is obtained. By comparing with HMM model algorithm, the algorithm designed in this paper turns out to be of higher recognition rate and better anti-noise performance under the same training sample and test set.