GNSS整周模糊度估计方法研究

发布时间：2017-12-27 01:02

本文关键词：GNSS整周模糊度估计方法研究　出处：《中国矿业大学(北京)》2016年博士论文　论文类型：学位论文

【摘要】：随着全球导航卫星系统(Global Navigation Satellite System,GNSS)的建设和现代化发展,GNSS已广泛应用于全球各领域中。这些应用对GNSS精度和可靠性的要求越来越高,有些应用需要近实时或实时定位。整周模糊度的快速、准确解算是实现和保证GNSS快速、高精度定位的关键因素之一,为此,本文着重对GNSS定位中整周模糊度的估计方法进行了深入研究。围绕整周模糊度解算中搜索空间的确定和搜索策略进行了详细分析和探讨。依据GNSS整周模糊度估计问题与格上最近向量问题的等价性,将格中最近向量问题的搜索算法进行了改进和优化,并应用于整周模糊度解算中,解决了整周模糊度快速解算问题。详细讨论了精密单点定位模糊度解算问题,分析了基于模糊度浮点解的静态精密单点定位性能和应用效果。本文还对影响GNSS定位的电离层电子含量的计算作了一些介绍。主要的研究内容及成果如下:(1)模糊度搜索空间的大小是影响模糊度解算效率的主要因素之一。针对传统确定模糊度搜索空间的方法较为保守,致使搜索空间过大,提出了一种基于最小二乘模糊度降相关平差法(LAMBDA)的改进方法。首先,在介绍混合整数最小二乘方法的基础上,以LAMBDA方法为例,对其搜索空间确定的三种方法进行详细分析和对比,评价了这些经典方法的优缺点。然后,定义了搜索空间的一个影响因子,并结合LAMBDA方法提出了确定模糊度搜索空间的修正公式。基于仿真和实测数据进行实验,结果表明,该方法在保证获得期望模糊度组数的前提下,其确定的模糊度搜索空间包含的整数点个数更小,可保证90%以上实际模糊度组数接近于期望值。(2)在模糊度域内搜索的方法中,Bootstrapping方法解算效率高,但成功率偏低。整数最小二乘方法理论严密,模糊度解算成功率高,但解算效率相对较低。为保证整周模糊度的解算效率和成功率,综合Bootstrapping方法和LAMBDA方法,提出一种以Bootstrapping成功率为约束条件,以R-Ratio检验为确认原则,基于部分搜索的模糊度解算方法。实验分析表明,与原有算法相比,该算法在保证模糊度固定成功率满足预设值的条件下,能更快速地固定模糊度,较适用于高维模糊度快速解算。(3)模糊度估计是混合整数最小二乘问题,实际上也是格中的最近向量问题。在简要介绍解决最近向量问题的搜索算法基础上,针对现有模糊度搜索方法仍不能很好满足快速定位的需求,将M-VB搜索算法作了两方面改进,并应用于模糊度解算中:一是优化了该算法执行过程中更新上界的问题;二是提出借助Bootstrapped估计值来确定其搜索空间半径的方法。基于仿真数据和实测数据,分别在降相关和不降相关条件下,将上述改进方法与LAMBDA方法、MLAMBDA方法进行了对比分析。结果表明,改进的M-VB算法比另外两种方法能更快固定整周模糊度。(4)以基于小数(Fractional Cycle Bias,FCBs)改正的模糊度固定方法为基础,详细介绍了静态精密单点定位(Precise Point Positioning,PPP)模糊度固定的基本过程,并系统分析了基于模糊度固定解PPP定位的优缺点。在此基础上,利用Bernese 5.0软件和CSRS(Canadian Spatial Reference System)PPP软件定量分析了PPP浮点解参数估计值的精度及收敛时间等情况。最后,针对传统测量方式布设近井点成本高、精度不统一等问题,提出利用精密轨道和星历产品,采用PPP技术建立近井点。结果表明,点位精度优于5 cm,静态PPP技术可应用于近井点测量。(5)电离层总电子含量(Total Electron Content,TEC)是影响卫星导航定位的主要误差源之一。在详细分析GNSS定位模型和定位误差源后,介绍了一种简单的电离层TEC计算方法。为验证该方法的计算精度,分别利用太阳活动低年和太阳活动高年低、中、高维度的四个时段IGS数据进行了实验分析,结果表明,该方法计算的TEC与IGS发布的TEC相差较小,约90%的残差值在±3 TECU内,并且该方法在太阳活动低年的适用性更好。另外,与普通内插方法相比,该方法计算的TEC精度更高。
[Abstract]:With the development of the global navigation satellite system (Global Navigation Satellite System (GNSS)), GNSS has been widely used in all fields of the world. These applications are becoming more and more demanding for GNSS accuracy and reliability, and some applications need near real-time or real-time positioning. The fast and accurate solution of integer ambiguity is one of the key factors to achieve and ensure the location of GNSS in a fast and high-precision way. Therefore, this paper focuses on the estimation of integer ambiguity in GNSS positioning. The determination and search strategy of the search space in the integer ambiguity resolution are analyzed and discussed in detail. On the basis of GNSS integer ambiguity estimation problem of equivalence theory and the closest vector problem, the search algorithm in the lattice closest vector problem is improved and optimized, and applied to the ambiguity solution, solves fast ambiguity resolution problem. The ambiguity resolution of precision single point positioning is discussed in detail, and the performance and application effect of static precision single point positioning based on Fuzzy floating-point solution are analyzed. The calculation of the ionospheric electron content which affects the GNSS positioning is also introduced in this paper. The main research contents and achievements are as follows: (1) the size of fuzzy search space is one of the main factors that affect the efficiency of ambiguity resolution. In view of the fact that the traditional method of determining ambiguity search space is conservative, resulting in too large search space, an improved method based on least squares fuzzy degree correlation correlation adjustment (LAMBDA) is proposed. First, on the basis of introducing the mixed integer least squares method, taking the LAMBDA method as an example, the three methods of determining the search space are analyzed and compared in detail, and the advantages and disadvantages of these classical methods are evaluated. Then, an influence factor of the search space is defined, and a modified formula for determining the ambiguity search space is proposed with the LAMBDA method. Experiments based on simulated and measured data show that the proposed algorithm ensures smaller number of integer points in the fuzzy search space, while ensuring the expected number of fuzzy sets. It ensures that the number of 90% actual fuzzy sets is close to the expected value. (2) in the method of searching in the fuzzy domain, the efficiency of the Bootstrapping method is high, but the success rate is low. The theory of integer least squares is strict, and the success rate of ambiguity resolution is high, but the efficiency of calculation is relatively low. In order to ensure the efficiency and success rate of integer ambiguity resolution, a method of ambiguity resolution based on partial search is proposed, which takes Bootstrapping success rate as constraint condition and R-Ratio test as the recognition principle by combining the LAMBDA method and the Bootstrapping method. The experimental analysis shows that compared with the original algorithm, the algorithm can more quickly fix the fuzzy degree under the condition that the fixed rate of success of the ambiguity satisfies the preset value, and it is more suitable for the fast solution of high dimension ambiguity. (3) the fuzzy estimation is a mixed integer least square problem, and in fact it is also the nearest vector problem in the lattice. The basic search algorithm to solve the closest vector problem in brief, aiming at the existing ambiguity search method are still not very good to meet the rapid positioning needs, M-VB search algorithm has been improved in two aspects, and applied to the ambiguity: one is to optimize the implementation of the algorithm update the upper bound of the problem in the process; two is presented by using Bootstrapped estimation method to determine the radius of the search space. Based on the simulated data and measured data, the above improved methods are compared with the LAMBDA method and the MLAMBDA method respectively under the condition of correlation and no drop correlation. The results show that the improved M-VB algorithm can fix the whole week ambiguity faster than the other two methods. (4) based on the decimal (Fractional Cycle Bias, FCBs) based on correction of ambiguity fixing method, introduces the static precise point positioning (Precise Point Positioning, PPP) the basic process of fixing ambiguity, ambiguity and systematic analysis of the advantages and disadvantages of PPP positioning solution based on. On this basis, the accuracy and convergence time of PPP floating point parameter estimation are quantitatively analyzed by using Bernese 5 software and CSRS (Canadian Spatial Reference System) PPP software. Finally, aiming at the problems of high cost and unequal accuracy in the traditional measurement method, a precise wellbore and ephemeris product and PPP technology are used to establish near wellbore points. The results show that the point position accuracy is better than 5 cm, and the static PPP technology can be applied to the near well point measurement. (5) the total electron content of the ionosphere (Total Electron Content, TEC) is one of the main error sources affecting the navigation and positioning of the satellite. After a detailed analysis of the GNSS location model and the location error source, a simple method for calculating the ionosphere TEC is introduced. For the accuracy of the method is verified, using solar activity four hours IGS data in low and high solar activity is low, medium and high dimensions were analyzed. The results show that TEC and IGS calculated by this method is the release of TEC is small, the residual value of about 90% in 3 TECU, and this method has better applicability in the years of low solar activity. In addition, compared with the common interpolation method, the proposed method has higher TEC accuracy.
【学位授予单位】：中国矿业大学(北京)
【学位级别】：博士
【学位授予年份】：2016
【分类号】：P228.4

【相似文献】