A New Intelligent Recognition Method for Maneuver Modes of Re-entry Gliding Vehicle

Abstract

Aiming at the problem of maneuver mode recognition for re-entry gliding vehicles (RGV), this article proposes a new intelligent method for maneuver mode recognition of re-entry gliding vehicles. Based on the extracted feature parameters that fit the maneuver characteristics of the vehicle trajectory and the constructed RGV maneuver modes trajectory library, an LSTM deep learning neural network was built to train the extracted new feature parameters. Compared with other typical feature parameters in network training, the results show that our proposed feature parameters converge faster and more stably in LSTM maneuver mode recognition network training, and achieve high recognition accuracy.