GNSS精密单点定位及非差模糊度快速确定方法研究

发布时间：2018-07-28 08:42

【摘要】：GPS精密单点定位(Precise Point Positioning,简称PPP)是上世纪九十年代末发展起来的一种新的GPS定位技术,它集成了GPS标准单点定位和GPS相对定位的技术优势,是GPS定位技术中继RTK/网络RTK技术后出现的又一次技术革命。历经十多年的快速发展,GPS精密单点定位的基本理论与实践问题已经得到了比较好的解决,目前正在朝工程化应用方向迈进。但是,现阶段能工程化应用的PPP技术仍然以PPP浮点解为主,PPP固定解尚处于研究和发展阶段。特别是实时PPP中的快速模糊度固定问题仍然是当前的一个研究难点。本论文系统深入地研究PPP的数学模型,分析传统PPP模糊度无法固定为整数的本质原因,进而提出一套新的非差模糊度整数解方法。详细研究了PPP模糊度无法快速固定为整数的原因,提出了实时PPP快速初始化和快速重新初始化的新方法。融合网络RTK与PPP技术的优势,建立一套完整的统一的精密定位服务系统,为实时PPP技术的广泛应用奠定理论与实践基础。论文的主要工作和贡献如下： (1)详细介绍了精密单点定位的常用数学模型、主要误差项以及参数估计方法。为了满足实时PPP的需要,提出并开发了高采样率精密卫星钟差快速估计和更新算法。讨论了其解算效率及精度,估计的钟差与IGS最终钟差吻合较好,其相对偏差大多数在0.1ns以内；开发了一套基于因特网的实时精密单点定位服务系统,实验结果表明：该实时系统可取得平面方向5厘米,高程方向10厘米左右的实时定位精度。同时,针对非差相位观测值数据预处理中小周跳难探测、难修复的问题,提出了一种新的实时PPP中的周跳固定与修复方法,该方法精细分离各项误差,并引入LAMBDA方法搜索周跳值。 (2)从GNSS观测的基本数学模型出发,分析了传统PPP模型中模糊度无法固定为整数的原因,基于此提出了非差模糊度整数解的新模型与方法。大量的实验结果表明PPP模糊度固定可以显著提高精密定位与定轨的精度。同时,提出了一种利用预报大气层延迟辅助PPP模糊度快速固定的新方法,该方法充分利用首次初始化完成后生成的精密大气信息,可实现短时信号中断后模糊度的单历元固定,避免了信号中断所引起的PPP重新初始化的问题。大量实验结果表明：使用本文所提出的方法,即使在所有的卫星全部短时失锁的情况下,仍然可以实现PPP单历元模糊度固定。 (3)提出了区域增强的PPP方法。充分利用从区域参考网提取的精密大气改正数,实现了实时PPP的快速初始化。该方法可以将网络RTK和实时PPP两大技术融为一体,集成为一个统一的定位服务。即在没有区域增强信息的情况下,可全球范围内实现几个厘米精度的实时PPP服务,在有区域信息增强的区域,则可提供更高精度的快速精密定位服务。同时,基于此开发了一种分层的GNSS精密数据处理策略,可以实现大规模且疏密不均的参考网的PPP区域增强。精密轨道和钟差的生成作为基本的两层,另外的两层分别为相位小数偏差估计和区域增强改正数。 (4)针对PPP模糊度初始化时间较长的难题,提出了顾及电离层特征的PPP模型,基于该模型开发了新的相位小数偏差估计算法。新模型不仅能改进相位小数偏差产品的质量,而且显著缩短PPP模糊度的初始化时间。实验结果表明：新方法能缩短实时动态解收敛时间30%,由超过30分钟缩短为不超过20分钟；首次模糊度固定时间减少25%,由20分钟缩短为15分钟。处理静态测站,新方法首次固定时间约为10分钟,在测站坐标固定为已知值情况下固定时间仅需5分钟。 (5)对各种定位模型与算法进行深入分析,并经过提炼与综合,提出了一种统一的顾及大气约束的非差非组合的精密定位模型与算法。即采用非差原始观测值,将电离层和对流层延迟作为参数进行估计,并依据其时空特征以及实时得到的大气模型或区域大气改正数合理的施加适当的约束。这对于使用稀疏参考网增强PPP尤为重要,合理的约束大气延迟参数可以补偿由大的测站间距引起的残余大气误差。这个统一的算法不仅能将PPP和网络RTK集成为无缝的定位服务,且可将这两个技术融合为一个统一的模型和算法。基于实时产品流和统一的处理策略,我们可以提供如下实时动态定位服务：5-10厘米精度的全球PPP；2-5厘米精度的固定解PPP；1-3厘米精度的区域增强PPP。
[Abstract]:GPS precision single point positioning (Precise Point Positioning, abbreviated as PPP) is a new GPS positioning technology developed at the end of the 90s of last century. It integrates the technical advantages of GPS standard single point positioning and GPS relative positioning. It is another technical revolution after GPS positioning technology relay RTK/ network RTK technology. After more than 10 years of rapid development. The basic theoretical and practical problems of GPS precision single point positioning have been solved well and are moving towards engineering applications. However, at present, the PPP technology that can be applied in engineering is still based on PPP floating-point solutions, and the fixed solutions of PPP are still in the research and development stages, especially the fast fuzzy fixed questions in real-time PPP. The problem is still a research difficult point at present.
This paper systematically studies the mathematical model of PPP, analyzes the essential reason that the traditional PPP ambiguity can not be fixed as an integer, and then puts forward a new set of non differential integer solution method. The reason why the PPP ambiguity can not be fixed quickly to the integer is studied in detail, and a new method of rapid initialization and rapid re initialization of the real-time PPP is proposed. Combining the advantages of RTK and PPP technology, a complete and unified precision positioning service system is set up, which lays the foundation for the wide application of real-time PPP technology. The main work and contribution of this paper are as follows:
(1) the common mathematical model, main error term and parameter estimation method of precision single point positioning are introduced in detail. In order to meet the needs of real time PPP, a fast estimation and updating algorithm of high sampling rate precision satellite clock difference is proposed and developed. The calculation efficiency and accuracy are discussed. The estimated clock difference is in good agreement with the final clock difference of IGS, and its relative deviation is relatively biased. Most of the difference is within 0.1ns, and a set of real-time precision single point positioning service system based on the Internet is developed. The experimental results show that the real-time system can obtain real-time positioning accuracy of 5 cm in plane direction and 10 centimeters in elevation direction. At the same time, it is difficult to repair and repair the small and medium cycle hops according to the number of non differential phase observations. In this paper, a new cycle skip fixing and repairing method in real-time PPP is proposed. The method extracts the errors accurately and introduces the LAMBDA method to search cycle slips.
(2) starting from the basic mathematical model of GNSS observation, the reason that the fuzzy degree of the traditional PPP model can not be fixed as an integer is analyzed. Based on this, a new model and method for the integer solution of the non difference ambiguity is proposed. A large number of experimental results show that the precision of the precision localization and the orbit determination can be greatly improved by the PPP fuzziness. At the same time, a kind of use is put forward. A new method for predicting the fast fixation of PPP fuzzy degree aided by atmospheric delay is used. This method makes full use of the precise atmospheric information generated after the first initialization, and can realize the single epoch of the ambiguity after short time signal interruption, avoiding the problem of the re initialization of PPP caused by signal interruption. A large number of experimental results show that the use of this paper is used in this paper. The proposed method can achieve PPP single epoch ambiguity even if all satellites are short lost.
(3) a region enhanced PPP method is proposed. The rapid initialization of real-time PPP is realized by making full use of the precision atmospheric correction number extracted from the regional reference network. This method can integrate the network RTK and the real time PPP technology into a unified positioning service. A real-time PPP service with a few centimeter precision can provide a higher precision positioning service in a region with enhanced regional information. At the same time, a hierarchical GNSS precision data processing strategy is developed to achieve the PPP enhancement of a large and unevenly distributed reference network. The generation of precision orbit and clock difference is made. For the basic two level, the other two layers are phase decimal deviation estimation and region enhancement correction.
(4) aiming at the problem of the longer initialization time of PPP ambiguity, a PPP model considering the ionosphere characteristics is proposed. A new phase decimal estimation algorithm is developed based on the model. The new model can not only improve the quality of the phase decimal deviation product, but also significantly shorten the initialization time of the PPP ambiguity. The experimental results show that the new method can be reduced. The short time dynamic solution convergence time is 30%, which is shortened from more than 30 minutes to less than 20 minutes. The first fuzzy fixed time is reduced by 25% and the 20 minute is shortened to 15 minutes. The static station is treated with a fixed time of about 10 minutes for the first time, and the fixed time of the station coordinates is fixed to the known value only 5 minutes.
(5) in depth analysis of various positioning models and algorithms, and after refining and synthesis, a unified model and algorithm for non differential non combinatorial positioning, taking into account of atmospheric constraints, is proposed. That is, using the non differential original observation values, the ionosphere and tropospheric delay are estimated as parameters, and based on their temporal and spatial characteristics and real time results. The atmospheric model or the regional atmospheric correction number is properly constrained. This is particularly important for the use of sparse reference networks to enhance the PPP. The reasonable constraints of the atmospheric delay parameters can compensate for the residual atmospheric errors caused by large station spacing. This unified algorithm can not only integrate PPP and network RTK into seamless location services, but also can be used as a seamless location service. The two technologies are integrated into a unified model and algorithm. Based on real time product flow and unified processing strategy, we can provide the following real-time dynamic positioning services: 5-10 cm precision global PPP; 2-5 cm precision fixed solution PPP; 1-3 cm precision regional enhanced PPP.
【学位授予单位】：武汉大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：P228.4

【参考文献】