Researches On The Prediction Methods Of Polar Motion Of The Earth Rotation Parameters

High precision PM have important theoretical and practical significance.And it is necessary to obtain the final PM parameters through complex calculation process, so the PM can not be obtained in real time.Because the excitation mechanism of the polar shift is relatively complex,the prediction results are not very satisfactory. Therefore, the prediction of polar motion still needs further research.With the increase of polar sequence observation precision and the optimization of polar motion prediction method, the prediction accuracy of polar motion is gradually improved. However, the method of prediction of polar motion needs further improvement. This research, then, has the following contents:(1)In order to improve the accuracy of short-term prediction,This paper further studies the ARMA model in the prediction of PM .The LS+ARMA is used to predict PM within 5 days, and the results are compared with other researchers, The feasibility of the model is verified.(2)Based on the stability of the data, this paper proposes a prediction method based on the double difference LS+ARMA model, to predict PM within 100 days, and the results are compared with LS+ARMA, The result shows that the accuracy of the double difference LS+ARMA model is better than the LS+ARMA model, this model is both feasible and effective to predict the polar motion parameters.(3)Taking into account the existence of high frequency signal will have a certain impact on the accuracy of the pole shift prediction.Therefore, this paper proposes by using wavelet analysis to decompose the time serise into a low frequency signal and several high frequency signals, The influence of high frequency signal on the accuracy of PM prediction is studied by different combinations of high frequency and low frequency.The result shows that the high frequency signal can reduce the short-term prediction accuracy ,but the effect of the medium and long term prediction accuracy is negligible. This method is used to predict PM within 365 days, and the results are compared with IERS, The feasibility of the model is verified.