顾及RMLE的GNSS时序噪声特性及环境负载修正

GNSS time series noise characteristics and environmental loading correction considering RMLE algorithm

针对环境负载易对全球卫星导航系统(GNSS)站坐标时序修正前后噪声模型特性及其站速度造成影响的问题,提出一种顾及约束最大似然估计(RMLE)的GNSS时序噪声特性及环境负载修正方法:选用206个连续13a的GNSS站时间序列,采用改进的贝叶斯信息模型估计准则,利用约束最大似然估计(RMLE)方法探讨和分析大气压负载、非海洋潮汐负荷、积雪负载及土壤水负载等环境负载对GNSS时序修正前后噪声模型特性及其站速度的影响。结果表明:经环境负载修正后,GNSS时序中的噪声主要表现为闪烁噪声+白噪声(FNWN)与幂律噪声+白噪声(PLWN),北(N)、东(E)、天(U)3个分量上噪声模型修正前后分别有约19.9%、36.5%、40.8%的噪声模型发生变化,N、E、U分量上分别有96.6%、84.5%、88.3%的测站速度不确定度减小,对垂直速度估计的最大影响可达0.5mm/a;证明环境负载修正能够提高速度估值的精度。

Aiming at the problem that the environmental loads are easy to affect the characteristics of noise models and station speeds before and after the global navigation satellite system(GNSS)station coordinate time series correction,the paper proposed a method of GNSS time series noise characteristics and environmental loading correction considering RMLE algorithm:an improved Bayesian information model estimation criterion was put forward with the selected time series of 206 GNSS stations for 13 consecutive years,and the effects of the environmental loads including atmospheric pressure loading,non-tidal ocean loading,snow cover mass loading and soil moisture mass loading on noise model characteristics and station speed before and after GNSS time series correction were discussed and analyzed by using restricted maximum likelihood estimation(RMLE)algorithm.Results showed that after environmental load correction,the noise in GNSS timing series would mainly be manifested as flicker noise+white noise(FNWN)and power law noise+white noise(PLWN),moreover,before and after the correction of the noise model on the North(N),East(E)and Upper(U)components,about 19.9%,36.5%and 40.8%of the noise models would have changed,respectively,meanwhile,on the N,E and U components,96.6%,84.5%and 88.3%of the station speed uncertainty would have decreased,with a maximum impact up to 0.5 mm/a on vertical speed estimation,which verifies that the environmental load correction could improve the accuracy of speed estimation.