Elman neural network-based orbit prediction algorithm

Abstract

The invention discloses an orbit prediction algorithm based on an Elman neural network, and the algorithm is characterized in that the algorithm comprises the steps: inputting the orbit coordinates of a to-be-predicted ballistic missile into an Elman neural network prediction model, and outputting the predicted orbit coordinates; constructing an Elman neural network prediction model: constructing an Elman neural network model; and substituting the acquired reconnaissance data into the Elman neural network model for iterative training until the conditions are met, and outputting a final Elman neural network prediction model. Training a constructed Elman neural network prediction model based on reconnaissance data; based on the number of iterations or error precision, the Elman neural network prediction model is constructed, and the predicted orbit coordinates of the ballistic missile are more accurate.

Description

technical field [0001] The invention relates to an orbit prediction algorithm based on an Elman neural network, belonging to the technical field of orbit prediction. Background technique [0002] The trajectory of a ballistic missile is a complex physical model, involving many fields such as theoretical mechanics, earth gravity, aerodynamics, structural mechanics, missile ballistics and modern mathematics. In the process of surveillance, detection and tracking of ballistic missile targets, high-precision ballistic calculations are required to predict the ballistic trajectory of ballistic missile targets. [0003] The movement process of the ballistic missile target is very complicated, especially in the process of entering and exiting. Due to the influence of the dense atmosphere, the movement of the ballistic missile target is highly nonlinear, and the small initial value error and model error will also bring great influence to the ballistic forecast. error. Furthermore, ...

AI Technical Summary

Problems solved by technology

Therefore, in order to achieve high-precision ballistic prediction in the past, it is necessary to improve the accuracy of the initial prediction value, reduce the prediction propagation error and improve the real-time processing from the three aspects of high-precision dynamic model, high-precision nonlinear smoothing / filtering and high-precision extrapolation prediction. The computational efficiency is used for high-precision real-time ballistic prediction, but the establishment of such models is very complicated, and there are many parameters to be considered, and the initial value optimization is difficult

Images