Research On Satellite Orbits Predicting Algorithm Based On Deep Learning

The orbital prediction of space targets plays an important role in the national space safety,and has a wide range of applications in orbit tracking and navigation.Traditional orbital prediction methods based on dynamic models need to model the perturbation forces of satellites in orbit.However,due to the complexity of space forces and the uncertainty of satellite parameters,it is impossible to get the accurate dynamic models of the satellite,so that the orbit prediction can not achieve high accuracy.In this paper,LSTM algorithm is applied to satellite orbit prediction,which improves the accuracy of orbit prediction.Therefore,the research of satellite orbit prediction algorithm based on depth learning proposed in this paper has important theoretical significance and application value.Firstly,this paper analyses the current situation of orbit prediction at home and abroad,introduces the time system and coordinate system needed to describe the satellite in-orbit operation,and analyses the perturbation model of satellite in-orbit operation.The principles of numerical and analytical algorithms commonly used in satellite orbit prediction are introduced.The orbit prediction by numerical integration algorithm is simulated and the results of numerical algorithms under various perturbation conditions are obtained.Secondly,the initial orbit determination method of space non-cooperative target is studied.The space-based orbit observation is modeled,and the observability of the orbit determination method based on space-based angle measurement is analyzed,and the simulation results are given.Then,on the basis of Laplace orbit determination method,the iteration formula of J2 order perturbation orbit determination method is deduced and simulated.Finally,the disadvantages of the existing orbit prediction are analyzed,the time series modeling and prediction are studied,and a satellite orbit prediction method based on LSTM algorithm is proposed.The feasibility and effectiveness of the algorithm are verified by simulation.Finally,an improved LSTM algorithm based on firefly algorithm is given and the methods of its application in orbit prediction is proposed.The feasibility and superiority of the method are verified by simulation.