Font Size: a A A

Research On Application Of Singular Spectrum Analysis In Geodetic Survey Time Series

Posted on:2015-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:C L LvFull Text:PDF
GTID:2180330434453892Subject:Surveying the science and technology
Abstract/Summary:PDF Full Text Request
Abstract:The global continuous GPS observations has accumulated a large number of data, the GPS time series analysis has become a hot research topic. Due to observation precision demand is higher and higher, more and more influence factors, the research shows that the time series contains very rich, and even not clear kinds of information, how to separate them has become a research of difficulty. Traditional parametric model already can’t solve the complexity problem, the SSA is a kind adaptive analysis method which is started from the time series itself, it has been applied to the filtering, denoising, interpolation, detection of periodic signal, and trend. At present, the SSA method is widely used in the analysis of time, but there is few research in geodetic field. In generally, the main work and results of this dissertation is summarized as follows:1. By detailedly introducing the basic principle of SSA method, and based on the basic idea of Singular Spectrum Analysis, using the Hurst index of noise and signal has a significant difference of the properties, a new filtering method was proposed; and then introduces the methods to determine and evaluate whether the reconstructed components of SSA were trend or periodic signals.2. During the GPS Dynamic deformation monitoring, strong temporal correlation noises exists in the multipath effects errors, and then the method was applied to processing GPS multipath for the multipath correction model; at the same time, with the wavelet and EMD method and the SSA filtering method by other scholars put forward were compared. Results show that SSA filtering method is an effective denoising method, the denoising effect is the same as wavelet filter and EMD filter method and better than the SSA filtering method by other scholars put forward. The correction model was established based on the filter method, which can effectively weaken the influence of the multipath effect, thus improving the accuracy of GPS dynamic deformation monitoring. The standard of embedding dimension L and reconstruction order P are also given at the same time. 3. By combing with the time series of California ZIMM station, the SSA method is utilized to extract the height direction of periodic signal, and the periodic signals were removed from the direction of U, considering the case of colored noise, the rate of ZIMM station were estimated by the maximum likelihood estimation again. Results show that SSA method can effectively extract the periodic signal, and the ZIMM station’s periodic signal amplitude is time-varying, so model error will be introduced by using traditional model of least squares fitting, under the condition of colored noise, the model error will influence the rate.4. In view of the ionosphere is nonlinear and non-stationary characteristics in space-time, we us the ionosphere grid data provided by IGS, SSA method is introduced into the ionosphere TEC value forecast, the experimental results show that compared with the single ARMA forecast model, combination of SSA and ARMA model can effectively improve the prediction precision.25figures,10tables and87references.
Keywords/Search Tags:GPS, time series, Singular Spectrum Analysis, EmpiricalMode Decomposition, Wavelet
PDF Full Text Request
Related items