GNSS形变序列的信号提取方法比较分析研究

Comparison of the Signal Extraction Methods for GNSS Time Series

如何从观测数据序列中准确提取信号是GNSS数据处理的主要研究内容之一。经验模态分解和局部均值分解是提取信号的较为常用的2种时频分析方法,与其相比,奇异谱分析法能够准确识别提取时间序列中的趋势周期信号且不需要先验信息。因此文中利用奇异谱分析法对GNSS数据进行深入分析提取变形信号,并与局部均值分解和经验模态分解2种方法进行对比,分析3种方法提取信号的效果。鉴于奇异谱分析识别周期和趋势信号的优势,实验结果表明奇异谱分析与其它2种方法相比能更为准确的提取信号,是一种有效的GNSS数据处理分析信号提取方法。

How to extract the signals from the observation data series exactly is one of the main research contents for GNSS data processing. Among the methods, empirical mode decomposition (EMD) and local mean decomposition (LMD) are two commonly used time-frequency analysis methods for extracting signals. Compared with the two methods, singular spectrum analysis can identify and extract trend periodic signals of time series easily without the prior information. Thus in this paper, singular spectrum analysis is used to extract the signals from GNSS time series, and compared with LMD and EMD methods to verify the effect of three methods. In view of the advantages of SSA in identifying periodic and trend signals, experimental results showed that SSA can extract more signals accurately than the other two methods, and it is an effective signal extraction method for GNSS data processing and analysis.