Prediction and Analysis of Chinese Earth Rotation Parameters Based on Robust Least-Squares and Autoregressive Model

摘要

Earth Rotation Parameters (ERP) are the necessary parameters to achieve mutual conversion of the celestial reference frame and earth reference frame. It is the requirements of national economic construction and national major strategic defense, and. the needs of key frontier disciplines and production applications. In this paper, we take into account the consistency and stability of the Chinese ERP data, and the least-squares combination of Autoregressive (LS+AR) model and the Robust least-squares combination of Autoregressive (RLS+AR) model are proposed to use to predict different span of ERP based on Chinese ERP data, as the cycles and trends have the characteristic of time-varying in the data. The results show that, the RLS+AR model can reduce Influence of crude differential on forecast accuracy effectively, and then achieve the IERS prediction accuracy, and meet the need of national defense in time and space datum height consistency.