| Since the1980s to the1990s, the application of the geocentric coordinate system is increasingly popular due to development of space technology. The geocentric coordinate system, in which the center of the earth is defined as the origin, can greatly improve the accuracy of measurement, and can quickly obtain accurate three-dimensional geocentric coordinates. At present, the positioning and imaging results obtained by space technology are based on the geocentric coordinate system as the reference system. The development and wide use of space technology urgently requires countries to provide high accuracy, geocentric, dynamic, practical, unified geodetic coordinate system as the basic guarantee of the social and economic activities.China Geodetic Coordinate System2000(CGCS2000) has been adopted as the new national geodetic reference system since July2008, which will be used to replace the old systems. CGCS2000is referred to ITRF97(International Terrestrial Reference Frame1997) at the epoch of2000.0, the reference frame of CGCS2000currently consists of national GPS control network2000and the national astro-geodetic network after combined adjustment with the GPS network. In the future, CORS will play an important role in the maintenance of the reference frame of CGCS2000. However, the CGCS2000cannot provide accurate velocities at present stage, which makes it can’t meet the needs of precise precision and navigation.To do research on movement characteristics and model of framework using the latest mathematical and surveying techniques to attempt to realize the TRS as precisely as possible is very important. It will improve the precision of domestic coordinate system by establishing non-linear model of CGCS2000coordinate frame, which include two aspects, the first aspect is studying the deformation modeling under the loading of surface water/atmosphere mass, the second aspect is studying the non-linear movement characteristics of stations which is affected by earthquake,which is always difficult to be modeled with using short,noisy time series. Some research has been done to study the method of linear velocity estimation of CORS Network in China based on the CGCS2000frame. There are also many studies about non-linear movement characteristics of CORS network in china based on traditional methods. However, few studies on non-linear model of CGCS2000Framework have been published.Singular Spectrum Analysis (SSA) is introduced in this paper, which is based on principal component analysis in the vector space of delay coordinates for a time series. This method has been used in oceanography for years. In this study, singular spectrum analysis for missing data (SSA-M) is introduced and Singular Spectrum Analysis-Inter Quartile (SSA-IQR), a gross error detection method, is proposed. These two methods are used for pretreatment of GPS time series. After that, SSA is combined with nonparametric test of Kendall and the maximum entropy method (MEM) to refine the interpretation of trend and oscillatory behavior. At last a prediction method, singular spectrum analysis for prediction (SSA-P), is proposed, which is used in this paper to establish forecast model for CGCS2000.It mainly contains the following sections:1) In this paper, the original data of500GPS stations in the Crustal Movement Observation Network of China, from1999to2009, was processed. We adopted two different data processing methods differ in choosing reference coordinators to process the data. Part of ITRF stations which are superior in stability were selected in the first method and the ITRF stations in the neighborhood were selected in the second. Comparison of the result through processing the data relatively, systematic deviations have been found between the two results. It also shows that the magnitude and characteristics of the systematic deviations act differently in different directions, and the functional mechanism is relative to the spatial and temporal distribution of the sites. Through analysis on the phenomena and findings referred, some relative conclusions are derived and the advices on establishment regional coordinate system are also given.2) It is of great practical significance to study the effect of the earthquake on GPS observation, which can provide important information for the work of updating and maintaining domestic framework. The CGCS2000is the high-precision geocentric coordinate system in our country, which functions by defining the coordinates of framework stations at the epoch2000.0and the velocity of the framework stations. And the GPS observation data has shown that not only the coordinates but also the velocity of the GPS stations may change after the earthquake, which means that to update the information of the framework stations after the earthquake using appropriate method is very necessary and exigent. In this paper, the influence scope of earthquake and the coseismic displacement were calculated by processing the GPS observation data. In order to analyse the movement characteristic of the stations before and after the earthquake, the time series and the velocity of fiducial stations in the Crustal Movement Observation Network of China (CMONOC) were also obtained by processing the observation data from1999to2009.In the end, some advices on the method to update and maintain domestic geocentric coordinate frame is given. The long-term coordinate frame can be updated by updating the coordinate and the velocity obtained by processing the observation data when that has been accumulated for several months. Before that data is collected, a useful method to maintain the system is establishing temporary short-time coordinate frame depending on CORS and temporary reference stations.3)As the traditional method of gross error detection and data interpolation for GPS time series has some flaws and limitations, the SSA-IQR (Singular Spectrum Analysis-Inter Quartile Range) is used to detect gross error and SSA-M (Singular Spectrum Analysis for Missing Data) is used to fill the gaps of several data sets in this paper. SSA is a data-adaptive, nonparametric method, which can extract as much reliable information as possible from short and noisy time Series without using prior knowledge about underlying physics of the system, based on the information it also provide the gross error detection and data interpolation models. SSA-M is a self-consistent method to infer missing data from GPS data series and extract the relevant empirical orthogonal functions. Since the filling and error detection method does not need a priori information about the data structure, the method is self-consistent and parameter free. Based on the theories of SSA and IQR criteria, the SSA-IQR method for gross error detection in observation data series is also studied. In this paper, the experiment is performed to identify the reliability of the methods, and the incomplete data of IGS stations in china is reconstructed. The result shows that SSA can be used in data interpolation successfully.4) Establishing Non-linear model of CGCS2000framework can improve the precision of domestic coordinate system, which can provide support for precise orbit determination of Beidou satellites and precise point positioning. Based on studying the theory of SSA, we propose some new methods to resolve the questions that may present when we use SSA in practical experiment. For example, the pseudo-data method is proposed to resolve the phase-shift at the end of data when we reconstruct the signal. This new methods are combined with Kendall and MEM methods to establish non-linear model of CGCS2000framework. The new methodology has been proved to be of high precision, reliability and self-adaptability.5) It is very important to study the prediction method to improve the CGCS2000real-time precision. In this paper, we proposed SSA-P (Singular Spectrum Analysis for prediction) to accomplish this work. Through experiments, this method has been proved to be of high precision, reliability and self-adaptability... |
PDF Full Text Request