GPS坐标时间序列中信号与噪声分析

发布时间：2018-06-05 20:41

本文选题：GPS坐标时间序列 + 噪声特性　；参考：《长安大学》2017年硕士论文

【摘要】：高精度GPS坐标时间序列精度已达到毫米级,这使GPS观测技术不仅可以实时监测地表剧烈运动造成的瞬时形变,也可监测到地表微小的蠕动形变。目前GPS已成为连续、动态、实时观测地表形变的最佳技术手段。本文在此背景下以GPS坐标时间序列为研究对象,主要探究了GPS坐标时间序列的趋势项、季节项振幅及相位、噪声类型等特征参数、不同GPS坐标时间序列分析方法所求特征参数的差异性、共模误差对GPS坐标时间序列特征参数的影响、及GPS坐标时间序列季节项成因分析。围绕上述几点研究内容,本文所做的研究工作和研究成果主要包括以下几点内容:(1)在介绍GPS数据处理基本原理、解算策略及方法、数据处理流程的基础上详细叙述了GAMIT/GLOBK软件中涉及到的常用坐标系及各坐标系之间的换算关系。(2)文中以中国内陆海原断裂、六盘山断裂、西秦岭北缘断裂所在区域(文中称A区域)、美国加州南部旧金山湾地区(文中称B区域)为研究区域,研究了区域内约130个GPS测站坐标时间序列的特征参数,发现A区域内37个GPS测站时间序列的特征参数在断裂两侧没有明显差异,B区域内有部分GPS测站的趋势项在断裂两侧有差异,说明目前B区域断裂两侧的地壳运动仍然活跃。(3)针对所选区域,研究了不同参考框架下GPS坐标时间序列的差异。研究结果表明不同参考框架下GPS坐标时间序列的水平(N、E方向)趋势项表现不同,而垂向(U方向)趋势项、各方向序列的季节性变化几乎不受参考框架影响。(4)对比了谐波模型、EMD获取的GPS坐标时间序列的特征参数,研究结果表明对绝大多数GPS测站而言,两种方法获取的测站趋势项基本一致,个别测站趋势项的不一致很可能是受序列季节性波动的影响,但两种方法获取的测站季节项、坐标残差时间序列有些许差异。通过频谱分析发现EMD获取的测站季节项包含了原始序列中主要的周期项,且季节项噪声水平远低于原始序列,而谐波模型获取的周期项仅包含了年、半年周期项。(5)本文探究了地表负载、区域GPS网共模误差对GPS测站时间序列的影响。对比了地表负载改正前后GPS测站时间序列特征参数的变化,发现地表负载改正前后测站各方向的趋势项变化甚微,可以忽略不计,但测站周期项振幅、噪声类型均有不同程度的变化。而区域网共模误差对测站各项特征参数均有影响,滤波后GPS测站时间序列的周期项振幅、相位均表现出局部一致性,而滤波前则比较离散。滤波后残差时间序列的标准差明显小于滤波前残差序列的标准差,另外滤波前后测站水平反向的速度场在运动方向上有较明显的差异。
[Abstract]:The precision of high precision GPS coordinate time series has reached millimeter order, which makes GPS observation technology can not only monitor the instantaneous deformation caused by the violent motion of the earth's surface in real time, but also detect the tiny creep deformation of the surface. At present, GPS has become the best technique for continuous, dynamic and real-time observation of surface deformation. In this paper, the GPS coordinate time series is taken as the research object, and the characteristic parameters of GPS coordinate time series, such as trend term, seasonal term amplitude and phase, noise type and so on, are discussed. The difference of characteristic parameters obtained by different GPS coordinate time series analysis methods, the influence of common mode error on the characteristic parameters of GPS coordinate time series, and the cause of formation of seasonal term of GPS coordinate time series are analyzed. The research work and results of this paper mainly include the following contents: 1) introducing the basic principle of GPS data processing, solving strategies and methods. On the basis of the data processing flow, this paper describes in detail the commonly used coordinate systems and the conversion relations between the coordinate systems involved in the GAMIT/GLOBK software. In this paper, the inland Haiyuan fault and Liupanshan fault in China are used. The area where the fault is located in the northern margin of the West Qinling Mountains (area A in this paper) and the San Francisco Bay area (area B in southern California) in the United States of America are studied. The characteristic parameters of the coordinate time series of about 130 GPS stations in the region are studied. It is found that the characteristic parameters of the time series of 37 GPS stations in area A have no obvious difference on both sides of the fault. It shows that the crustal movement on both sides of the fault in region B is still active. (3) for the selected area, the difference of GPS coordinate time series under different reference frames is studied. The results show that the trend term of GPS coordinate time series is different in horizontal direction and in the direction of vertical U) under different reference frames. The seasonal variation of each direction sequence is almost independent of the reference frame. (4) the characteristic parameters of the GPS coordinate time series obtained by the harmonic model EMD are compared. The results show that for the majority of GPS stations, The trend terms obtained by the two methods are basically the same, and the inconsistency of the trend terms of individual stations is probably affected by the seasonal fluctuation of the sequence, but the time series of coordinate residuals obtained by the two methods are somewhat different. Through spectrum analysis, it is found that the seasonal term obtained by EMD contains the main periodic term in the original sequence, and the noise level of the seasonal term is much lower than that of the original sequence, while the period term obtained by harmonic model contains only a year. In this paper, the influence of surface load and regional GPS network common-mode error on the time series of GPS stations is investigated. The change of time series characteristic parameters of GPS station before and after the surface load correction is compared. It is found that the trend item of each direction of the station has little change before and after the surface load correction, which can be ignored, but the amplitude of the period term of the station is negligible. The type of noise varies in varying degrees. The common-mode error of the regional network has an effect on the characteristic parameters of the station. The amplitude and phase of the periodic term of the GPS time series show local consistency after filtering, but before filtering, they are more discrete. The standard deviation of the time series of residual error after filtering is obviously smaller than that of the series of residual errors before and after filtering. In addition, the velocity field of the measured station before and after filtering is obviously different in the direction of motion.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P228.4

【参考文献】