Trajectory Prediction Of Hypersonic Vehicles In Near Space Based On Neural Network

In modern military warfare,the aircraft of major military powers are basically equipped with supersonic and stealth capabilities,while the hypersonic aircraft in nearby space has extremely high flight altitude and extremely fast flight speed.With these two advantages,it quickly becomes a new highland that major military powers in the world pay attention to and compete for.Now,each military power has its own hypersonic flight vehicle program,and some countries have successfully flown many times.As a defensive country,China has an urgent need to track and forecast hypersonic aircraft in nearby space.In recent years,with the rise of artificial intelligence technology and computer technology,more and more fields begin to integrate various artificial intelligence technologies such as machine learning and deep learning,among which neural network technology based on deep learning has developed particularly rapidly in recent years.This paper studies and designs a combination model based on Kalman filter and neural network with the background of tracking and prediction of US hypersonic vehicle HTV-2 to realize the tracking and prediction function of HTV-2.In this paper,the tracking problem of HTV-2 is taken as the starting point,the motion analysis of HTV-2 is carried out,the state equation of the system is given,and the Extended Kalman Filter is designed and implemented.Secondly,in the prediction stage of the target,the requirement of using Neural Network to model the state variable is proposed.Based on the filtering problem,several design schemes to solve the state variable prediction problem are discussed.According to the requirement of trajectory prediction,the standard neural network model is modified to make it more efficient and accurate in predicting state variables.Then,in order to solve the problem of tracking and forecasting,this paper designs and gives two schemes of combining Kalman filter and neural network,and improves the combined model based on the characteristics of the model,so that it can complete the tracking and forecasting task more accurately and efficiently.Finally,simulation experiments are carried out to verify the feasibility of the corresponding models.The simulation results are given under different task conditions to evaluate the performance of the model.