地基GPS反演大气水汽关键问题的理论与方法研究
发布时间：2018-09-19 17:15
【摘要】：研究利用地基GPS反演大气水汽不仅对天气变化的预报有重要意义,而且可以促进GPS定位精度的提高,从而发挥GPS在应用方面的潜力。基于此,本文对地基GPS反演大气水汽所涉及的主要问题进行了系统的研究,整个论文的主要研究内容和创新点概括如下： 1.概述了影响GPS反演大气水汽各主要问题的研究现状,并指出了本文的研究目标,对GPS的观测模型及估计方法进行了详细的介绍,从而为本文后续数据处理奠定基础。 2.分析了广播星历外推卫星轨道的精度,并给出了Lagrange插值、Neville插值、Chebyshev多项式拟合和Legendre多项式拟合进行精密星历加密的方法,通过对实际精密星历数据的计算,分别从内插和拟合的阶数、计算精度、计算效率进行了比较分析,并给出了最佳阶数插值算法和拟合算法计算卫星轨道的点位误差对比图、所有卫星轨道的精度对比效果图。随后,分别以实例对卫星信号发射时刻的两种计算方法与影响卫星轨道的地球自转误差进行了探讨分析。 3.基于TurboEdit方法对GPS观测数据进行周跳探测的特点,设计了固定弧段长度的滑动窗口拟合模型,对其中的Geometry-Free组合法进行改进,在探测出周跳后,提出利用最小二乘Chebyshev多项式拟合来修复周跳。实验结果表明：改进后的TurboEdit算法可以探测出等周的1周小周跳、等周的大周跳和连续的小周跳、大周跳,同时,最小二乘Chebyshev多项式拟合可以精确修复以上周跳对。 4.根据双差相位观测序列在相邻历元间求差后,周跳能真实地以粗差形式反映出来的特点,在利用Chebyshev多项式拟合双差观测量时,顾及抗差估计的思想、Chebyshev多项式拟合阶数、拟合弧段长度,提出基于移动窗口的抗差Chebyshev多项式拟合算法来探测与修复周跳,应用采样间隔为60s和15s的L1频率的载波相位数据测试表明,该算法可有效的探测多历元间隔周跳及连续周跳,能对周跳进行准确的修复。 5.针对不存在先验信息时常规GPS单历元数据处理中存在的问题,提出了一种新的GPS整周模糊度单历元算法。该算法先采用一个历元的码观测值进行最小二乘定位,求取初始模糊度,并根据解的中误差来构造模糊度原始搜索空间。再采用两种不同线性组合的扩波方法进行模糊度变换,使原模糊度的搜索空间变小。在模糊度的新搜索空间确定后,通过线性组合的逆变换求取模糊度N1及N2,并以模糊度函数法进行真值的搜索,实现单历元解算。采用基线长度不同的两组数据作为实例进行测试,试验结果表明了本文方法的可行性和正确性。 6.针对单频GPS动态定位中常用模糊度求解方法存在的问题,提出了一种新的整周模糊度快速解算方法。该方法先通过对双差观测方程中坐标参数的系数阵进行QR分解变换以消除坐标参数,从而仅对模糊度参数建立Kalman滤波方程进行估计,然后利用排序和双Cholesky分解对滤波得到的模糊度进行降相关处理,并结合收缩模糊度搜索空间的思想来搜索固定整周模糊度。以实测的动态数据为例对该方法进行测试,其分析结果表明,该方法不但可以改善模糊度浮点解精度,而且具有良好的模糊度降相关效果,可正确有效地实现整周模糊度的快速解算。 7.通过考虑模糊度的整数特性及相位波长与基线非参考站初始坐标误差之间的约束条件,提出了一种基于梯级递推的无模糊度基线解算方法。该方法不受周跳的影响,并且在基线求解过程中不用考虑模糊度参数。分别以单个历元和移动窗口的多历元两种求解方案验证了该方法的可行性和正确性,并且基线解算结果具有较高的精度。 8.利用参考站坐标已知的先验信息,基于组合后的超快星历,提出了一种参考站对流层湿延迟近实时估计的三步Kalman滤波算法,该方法先利用Kalman滤波分离宽巷模糊度与伪距多路径误差,再基于电离层无关组合模型,启动Kalman滤波器进行L1模糊度与相对对流层湿延迟的分离,然后利用将正确固定的L1双差模糊度进行回代的方法,重新构建Kalman滤波器来估计准确的相对对流层湿延迟参数,通过实例验证了该方法的可行性和正确性。随后,利用该方法获取的近实时对流层湿延迟计算了相应的近实时可降水量,并与GAMIT解算的可降水量结果进行了比较分析。
[Abstract]:The research on the retrieval of atmospheric water vapor by ground-based GPS is not only of great significance to the forecast of weather changes, but also can promote the improvement of GPS positioning accuracy, so as to bring into play the potential of GPS in application. Based on this, the main problems involved in the retrieval of atmospheric water vapor by ground-based GPS are systematically studied in this paper, and the main contents and contents of the whole paper are The innovations are summarized as follows:
1. The main problems affecting the retrieval of atmospheric water vapor by GPS are summarized, and the research objectives of this paper are pointed out. The GPS observation model and estimation method are introduced in detail, so as to lay a foundation for subsequent data processing in this paper.
2. The precision of extrapolating satellite orbit from broadcast ephemeris is analyzed. The methods of precise ephemeris encryption by Lagrange interpolation, Neville interpolation, Chebyshev polynomial fitting and Legendre polynomial fitting are given. By calculating the actual precise ephemeris data, the order of interpolation and fitting, the calculating precision and the calculating efficiency are compared and analyzed. Then, the comparisons between the best order interpolation algorithm and the fitting algorithm for calculating the position error of the satellite orbit and the precision comparison results of all the satellite orbit are given.
3. Based on the characteristics of Turbo Edit method in detecting cycle slip of GPS observation data, a sliding window fitting model with fixed arc length is designed, and the Geometry-Free combination method is improved. After detecting cycle slip, the least square Chebyshev polynomial fitting is proposed to repair cycle slip. The algorithm can detect the small cycle slip, the large cycle slip and the continuous small cycle slip. At the same time, the least square Chebyshev polynomial fitting can accurately repair the above cycle slip pairs.
4. According to the characteristic that the cycle slip can be truly reflected in the form of gross error after the difference between adjacent epochs of the double-difference phase observation sequence is calculated, the robust Chebyshev polynomial fitting order and arc length based on the moving window are proposed when the Chebyshev polynomial is used to fit the double-difference observation. Combined algorithm is used to detect and repair cycle slips. Carrier phase data of L1 frequency with sampling intervals of 60 s and 15 s are tested. The results show that this algorithm can detect multi-epoch interval cycle slips and continuous cycle slips effectively and can repair cycle slips accurately.
5. Aiming at the problem of conventional GPS epoch metadata processing without prior information, a new GPS integer ambiguity single epoch algorithm is proposed. Firstly, an epoch code observation is used to locate the GPS integer ambiguity single epoch, and the initial ambiguity is obtained. Then, the original search space of ambiguity is constructed according to the median error of the solution. The ambiguity N1 and N2 are obtained by the inverse transformation of the linear combination, and the true value is searched by the ambiguity function method to realize the single epoch resolution. As an example, the test results show that the method is feasible and correct.
6. Aiming at the problems of ambiguity resolution in single-frequency GPS dynamic positioning, a new method for fast ambiguity resolution is proposed. The method first eliminates the coordinate parameters by QR decomposition transformation of the coefficient matrix of coordinate parameters in the double-difference observation equation, and then establishes Kalman filter equation to estimate only the ambiguity parameters. Then, the fuzzy degree is reduced by ranking and double Cholesky decomposition, and the fixed ambiguity is searched by using the idea of contraction ambiguity search space. Moreover, it has good ambiguity reduction correlation effect, and can effectively and effectively achieve integer ambiguity resolution.
7. Considering the integer characteristic of ambiguity and the constraint condition between phase wavelength and the initial coordinate error of baseline non-reference station, a new method of ambiguity-free baseline solution based on cascade recursion is proposed. The feasibility and validity of the method are verified by the two solutions of multi-epoch with moving window, and the results of baseline calculation have high precision.
8. A three-step Kalman filtering algorithm for near-real-time estimation of tropospheric wet delay at reference stations is proposed based on the combined ultrafast ephemeris and the priori information of reference station coordinates. Firstly, the wide lane ambiguity and pseudo-range multipath error are separated by Kalman filtering, and then the Kalman filter advance is started based on the ionospheric independent combination model. After separating the L1 ambiguity from the relative tropospheric wet delay, the Kalman filter is reconstructed to estimate the accurate relative tropospheric wet delay parameters by using the method of regenerating the correctly fixed L1 ambiguity. The feasibility and correctness of the method are verified by an example. The near-real-time precipitable water is calculated by layer wet delay and compared with the results calculated by GAMIT.
【学位授予单位】：西南交通大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P228.4;P412.27
本文编号：2250790
[Abstract]:The research on the retrieval of atmospheric water vapor by ground-based GPS is not only of great significance to the forecast of weather changes, but also can promote the improvement of GPS positioning accuracy, so as to bring into play the potential of GPS in application. Based on this, the main problems involved in the retrieval of atmospheric water vapor by ground-based GPS are systematically studied in this paper, and the main contents and contents of the whole paper are The innovations are summarized as follows:
1. The main problems affecting the retrieval of atmospheric water vapor by GPS are summarized, and the research objectives of this paper are pointed out. The GPS observation model and estimation method are introduced in detail, so as to lay a foundation for subsequent data processing in this paper.
2. The precision of extrapolating satellite orbit from broadcast ephemeris is analyzed. The methods of precise ephemeris encryption by Lagrange interpolation, Neville interpolation, Chebyshev polynomial fitting and Legendre polynomial fitting are given. By calculating the actual precise ephemeris data, the order of interpolation and fitting, the calculating precision and the calculating efficiency are compared and analyzed. Then, the comparisons between the best order interpolation algorithm and the fitting algorithm for calculating the position error of the satellite orbit and the precision comparison results of all the satellite orbit are given.
3. Based on the characteristics of Turbo Edit method in detecting cycle slip of GPS observation data, a sliding window fitting model with fixed arc length is designed, and the Geometry-Free combination method is improved. After detecting cycle slip, the least square Chebyshev polynomial fitting is proposed to repair cycle slip. The algorithm can detect the small cycle slip, the large cycle slip and the continuous small cycle slip. At the same time, the least square Chebyshev polynomial fitting can accurately repair the above cycle slip pairs.
4. According to the characteristic that the cycle slip can be truly reflected in the form of gross error after the difference between adjacent epochs of the double-difference phase observation sequence is calculated, the robust Chebyshev polynomial fitting order and arc length based on the moving window are proposed when the Chebyshev polynomial is used to fit the double-difference observation. Combined algorithm is used to detect and repair cycle slips. Carrier phase data of L1 frequency with sampling intervals of 60 s and 15 s are tested. The results show that this algorithm can detect multi-epoch interval cycle slips and continuous cycle slips effectively and can repair cycle slips accurately.
5. Aiming at the problem of conventional GPS epoch metadata processing without prior information, a new GPS integer ambiguity single epoch algorithm is proposed. Firstly, an epoch code observation is used to locate the GPS integer ambiguity single epoch, and the initial ambiguity is obtained. Then, the original search space of ambiguity is constructed according to the median error of the solution. The ambiguity N1 and N2 are obtained by the inverse transformation of the linear combination, and the true value is searched by the ambiguity function method to realize the single epoch resolution. As an example, the test results show that the method is feasible and correct.
6. Aiming at the problems of ambiguity resolution in single-frequency GPS dynamic positioning, a new method for fast ambiguity resolution is proposed. The method first eliminates the coordinate parameters by QR decomposition transformation of the coefficient matrix of coordinate parameters in the double-difference observation equation, and then establishes Kalman filter equation to estimate only the ambiguity parameters. Then, the fuzzy degree is reduced by ranking and double Cholesky decomposition, and the fixed ambiguity is searched by using the idea of contraction ambiguity search space. Moreover, it has good ambiguity reduction correlation effect, and can effectively and effectively achieve integer ambiguity resolution.
7. Considering the integer characteristic of ambiguity and the constraint condition between phase wavelength and the initial coordinate error of baseline non-reference station, a new method of ambiguity-free baseline solution based on cascade recursion is proposed. The feasibility and validity of the method are verified by the two solutions of multi-epoch with moving window, and the results of baseline calculation have high precision.
8. A three-step Kalman filtering algorithm for near-real-time estimation of tropospheric wet delay at reference stations is proposed based on the combined ultrafast ephemeris and the priori information of reference station coordinates. Firstly, the wide lane ambiguity and pseudo-range multipath error are separated by Kalman filtering, and then the Kalman filter advance is started based on the ionospheric independent combination model. After separating the L1 ambiguity from the relative tropospheric wet delay, the Kalman filter is reconstructed to estimate the accurate relative tropospheric wet delay parameters by using the method of regenerating the correctly fixed L1 ambiguity. The feasibility and correctness of the method are verified by an example. The near-real-time precipitable water is calculated by layer wet delay and compared with the results calculated by GAMIT.
【学位授予单位】：西南交通大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P228.4;P412.27
【参考文献】
相关博士学位论文 前1条
1 张绍成;基于GPS/GLONASS集成的CORS网络大气建模与RTK算法实现[D];武汉大学;2010年
，本文编号：2250790
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2250790.html
