Analysis and prediction of BDS-3 satellite differential code bias based on LSTM

When the satellite differential code bias (DCB) constraints and benchmarks change, there will be a relatively large difference in its value,which affects the accuracy of navigation and positioning. This paper analyzes the time series changes of the BDS-3 satellite DCB in 2021, synthesizes the solar radiation flux and the geomagnetic index,and uses the LSTM neural network to predict and analyze the accuracy of the satellite DCB. The experimental results show that the prediction effect of the LSTM neural network model is better than that of the polynomial fitting method. The mean absolute deviation (MAE) and root mean squared error (RMSE) are less than 0.2 ns and 0.5 ns respectively. The errors of the forecast results for many days in the future are all less than 0.2 ns. LSTM neural network can effectively predict satellite DCB and provide reference for missing DCB products.