An LSTM-based Method for Recognition and Prediction of Aircraft Formation

Aircraft formation recognition and prediction are of great significance in modern air combat. Aiming at the problems of many manual interventions and complex implementation through traditional aircraft formation recognition methods, an intelligent recognition and prediction method of aircraft formation is proposed. First, a formation coding method is designed, which is combined with Support Vector Machine (SVM) to construct a formation recognition model. Then, a formation prediction model is constructed based on the Long-Short-Term Memory network (LSTM) and the recognition model. Finally, a dataset is generated to train the two models, and the trained model can be used for formation recognition and prediction. After experimental verification, the method proposed in the paper has better recognition and prediction effects on formations, the recognition accuracy can reach 95.5%, and the accuracy of formation prediction can reach 95%.