A method for predicting orbit errors of broadcast ephemeris by improved BP neural network

Abstract

The invention discloses a method for predicting the orbit errors of a broadcast ephemeris by using an improved BP neural network, belonging to the technical field of satellite navigation data processing. The method for realizing the invention comprises the follow steps of: data required for predicting orbit error of broadcast ephemeris is collected by use of a BP neural network; the data needed byBP neural network is preprocessed; then the BP neural network model is constructed by using the preprocessed data; then the initial weights and thresholds of BP neural network model are optimized byPSO algorithm; the improved BP neural network model is trained again; finally, using the test data to evaluate the accuracy of the improved BP neural network model, the orbit error of broadcast ephemeris is predicted and compensated. The invention can reduce sensitivity to initial parameters, avoid falling into local minimum value, effectively improve satellite orbit determination accuracy and reduce system-level error of satellite navigation system.

Description

technical field [0001] The invention relates to a method for predicting broadcast ephemeris orbit errors by using an improved BP neural network, and belongs to the technical field of satellite navigation data processing. Background technique [0002] The BeiDou Navigation Satellite System (BDS) is an important national space infrastructure. The master control station uses the orbit determination value to extrapolate and form the broadcast ephemeris, and the receiver can calculate the satellite orbit according to the broadcast ephemeris parameters. Real-time location, and users can locate and navigate themselves through the location of several satellites, so orbit accuracy is one of the key factors restricting the performance of satellite navigation services. [0003] One way to improve the accuracy of orbit determination is to establish a more accurate mathematical model of the system. Satellites will be affected by various perturbations in actual operation. Some known pertu...

AI Technical Summary

Problems solved by technology

[0006] Aiming at the complicated problem of improving the accuracy of orbit determination by refining the dynamic model in the prior art, the BP neural network is used to train the nonlinear mapping relationship between broadcast ephemeris orbit error and related parameters, but the BP neural network has The following defects: easy to fall into local minimum; and poor robustness, sensitive to the initial setting of parameters

Images