Prediction of Ionosphere F2 Layer Critical Frequency Based on EMD-ARIMA-LSTM

Abstract

Improper selection of flight test communication frequency will lead to the failure of long-distance high-frequency communication test, which will reduce flight test efficiency and increase flight test cost. The predicted value of the critical frequency foF2 in the F2 layer of the ionosphere is one of the important parameters for frequency selection calculation before the flight test. Taking the ionospheric F2 layer critical frequency prediction as the research object, based on the analysis of historical foF2 data, the EMD method is used to stabilize and reduce the noise of the original sequence. After obtaining period terms and trend terms with different time-frequency characteristics, combining the advantages of ARIMA and LSTM models in linear fitting and nonlinear fitting respectively, establish an ARIMA-LSTM model to predict each component. First, establish an ARIMA model to predict according to the autocorrelation characteristics of each component, and then use the prediction results of the ARIMA model and the historical data with strong correlation as the input of the LSTM model to construct the component LSTM model again. Finally, the ARIMA-LSTM model prediction results of each component are reconstructed to obtain the final prediction value of the foF2 sequence. Predict the 2016 foF2 data from the Beijing Ionospheric Vertical Observation Station and compare it with multiple models, the results show that the EMD-ARIMA-LSTM combined model can predict the chang