Intelligent trajectory recognition and prediction method for hypersonic glider vehicle
A method of target maneuver pattern recognition and trajectory prediction based on deep learning neural network is proposed for hypersonic gliding vehicle with strong maneuverability and long flight duration.The motion model of hypersonic glider is established,and the trajectory library of target maneuver model is set up,which is composed of longitudinal quasi-balance glide,jump glide,lateral serpentine maneuver and turn maneuver.A maneuvering pattern recognition network based on long short-term memory(LSTM)network is designed,which can effectively output the cor-responding tag sequence through the motion parameters of the target after training.For each maneu-ver mode,the regression fitting extrapolation trajectory prediction method based on LSTM is con-structed.The simulation results show that the method of combining intelligent maneuvering pattern recognition classification with regression fitting trajectory prediction is proposed in this paper,it can predict the long-time maneuvering trajectory of hypersonic gliding target with high accuracy in a short time and has a good adaptability.
NETL
NSTL
万方数据
