基于EEMD-MHM模型的观测值域多路径误差改正研究

Research on multipath correction of observation range based on EEMD-MHM model

为削弱多路径误差以提高GNSS定位精度,引入集合经验模态分解EEMD与多路径半天球图MHM进行组合构建多路径误差改正模型。该组合模型首先基于“零均值”假设进行GNSS双差到单差的转换,然后利用EEMD尺度自动分解与重构精确提取单差残差序列中的多路径误差,最终构建MHM模型用于观测值域多路径误差的实时改正。静态多天实验结果表明,EEMD-MHM组合的多路径误差改正模能提高相邻观测间多路径误差序列的相关性,提高多路径误差模型的精度,将多路径误差改正后的静态多天基线分量标准差STD稳定在1 mm以内,并且东北天方向上的均方根误差RMSE可分别提升75%、42.9%、62.5%,验证了该EEMD-MHM组合多路径误差改正模型的有效性和稳定性。

In order to solve this problem and improve the accuracy of GNSS,this paper proposes a correction model of multipath which is formed from EEMD and MHM.First,the single-difference residuals are obtained by converting the double-difference residuals based on the zero-mean assumption,and then the EEMD is used to automatically decompose and reconstruct the multipath for each satellite.Finally,the MHM model is constructed by the azimuth,elevation and multipath.The results of multi-day static experiments show that compared to smooth-MHM and non-correction,the EEMD-MHM model can enhance the correlation of multipath between two days and improve the accuracy of multipath.Meanwhile,the STD of the multi-day baseline vector is stabilized within 1mm,and the RMSE of the final baseline vector in the east,north and up direction can be increased by 75%,42.9%,and 62.5%respectively,which shows the effectiveness and stability of the EEMD-MHM model for multipath correction.