改进BIC的GNSS站坐标时序有色噪声特性分析  

Analysis of colored noise characteristics in GNSS station coordinate time series based on improved BIC

作　　者：鲁铁定[1] 罗新赣 孙喜文 LU Tieding;LUO Xingan;SUN Xiwen(School of Surveying and Geoinformation Engineering,East China University of Technology/Key Laboratory of Mining Environment Monitoring and Treatment around Poyang Lake,Ministry of Natural Resources,Nanchang,Jiangxi 330013,China)

机构地区：[1]东华理工大学测绘与空间信息工程学院/自然资源部环鄱阳湖区域矿山环境监测与治理重点实验室,江西南昌330013

出　　处：《导航定位学报》2024年第2期53-61,共9页Journal of Navigation and Positioning

基　　金：国家自然科学基金(42061077,42064001,42104023,42374040)。

摘　　要：为了有效解决传统的极大似然估计方法存在的未知数多导致结果不准确的问题,提出一种改进BIC的GNSS站坐标时序有色噪声特性分析方法:选取全球范围内100个基准站2008—2022年坐标时序及仿真的100组坐标时序,采用赤池信息准则(AIC)、贝叶斯信息准则(BIC)、改进的贝叶斯信息准则(BIC_tp)对不同有色噪声模型进行噪声特性分析,获取基准站坐标序列的最优噪声模型及速度参数;探讨BIC_tp对噪声模型及速度估计的影响。结果表明,基准站坐标序列噪声模型呈现出多样性特征,在北(N)和东(E)方向主要表现为闪烁噪声+白噪声(FNWN)模型,天顶(U)方向主要表现为幂律噪声+白噪声(PLWN)模型;相比于AIC和BIC噪声模型估计准则,BIC_tp方法对有色噪声模型估计具有较高的准确性,尤其是闪烁噪声+随机游走噪声(RW)+白噪声(FNRWWN)模型对全球卫星导航系统(GNSS)站速度不确定度的影响在N与E方向远大于U方向;正确获取模型参数估计的实际不确定度及改正噪声分量对于合理应用GNSS坐标时间序列数据具有重要意义。In order to effectively solve the problem of inaccurate results caused by many unknowns in the traditional maximum likelihood estimation method,the paper proposed an analysis method of colored noise characteristics in GNSS station coordinate time series based on improved Bayesian information criteria(BIC):the reference station coordinate time series from 2008 to 2022 and the simulation of 100 sets of coordinate time series were selected worldwide;and the noise characteristics of different color noise models were analyzed by using the Akaike information criteria(AIC),Bayesian information criteria(BIC)and the improved BIC(BIC_tp);then the optimal noise model and velocity parameters of the reference station coordinate time series were obtained;finally,the effect of BIC_tp on noise models and velocity estimation was discussed.Results showed that the reference station coordinate time series noise model could exhibit diversity characteristics,which would be mainly manifested as flicker noise(FN)+white noise(WN)model in the N and E directions,and power law noise(PL)+WN model in the U direction;moreover,compared with the estimation criteria of AIC and BIC noise models,the BIC_tp method would have high accuracy for estimating colored noise models,especially the influence of the FN+random walk noise(RW)+WN model on the velocity uncertainty of global navigation satellite system(GNSS)stations would be much greater in the N and E directions than that in the U direction;in general,it is of great significance to correctly obtain the actual uncertainty of model parameter estimation and correct the noise component for the reasonable application of GNSS coordinate time series data.

关 键 词：赤池信息准则(AIC) 贝叶斯信息准则(BIC) 改进的贝叶斯信息准则(BIC_tp) 噪声模型估计 全球卫星导航系统(GNSS)坐标时间序列 

分 类 号：P228[天文地球—大地测量学与测量工程]