【作者】 张瑞；

【导师】 姚宜斌；

【作者基本信息】 武汉大学 ， 大地测量学与测量工程， 2013， 博士

【摘要】 电离层是日地空间环境的重要组成部分,认识电离层的结构和活动规律,是人类对自身生存环境认识和利用的重要基础。电离层的研究对于研究日地空间环境及高空大气各层之间的相互关系和作用,提高测速、定位、授时、导航等系统的精度等方面,都具有非常重要的价值。相比较于传统电离层探测技术,GNSS探测技术能够实现连续运行、高密度覆盖。目前,四大全球卫星导航系统(Global Navigation Satellite System, GNSS)并存与发展的局面已初步形成。随着多模CNSS连续运行跟踪站数量增加与观测数据质量不断提升,针对多模GNSS系统不同特点进行深入研究,开展基于多模GNSS系统的电离层建模十分必要。由于电离层模型、伪距观测值精度等因素影响,单频单点定位仅可获得分米至米级定位精度。通过不同频率组合,双频用户可以消除超过99%的电离层误差影响。但观测值组合放大了观测噪声,需要更多的时间实现双频PPP收敛。因此,进一步提高电离层模型精度,可有助于提高单频定位精度、加速双频PPP收敛速度。电离层延迟对单频导航用户影响很大,但导航系统自身提供的实时改正产品效果有限,难以获得高精度定位结果。因此,基于区域或者全球的跟踪站网络提供的实时观测数据流,建立实时电离层模型,能够有效提高服务区域内单频用户定位精度。同时,利用实时电离层模型能够实现高精度的电离层监测,有助于研究电离层扰动及不规则变化。本文针对电离层二维层析中的一些问题,对数据预处理、电离层提取方法等进行深入分析；利用实测GPS/Beidou数据建立中国区域电离层模型,分析多模系统对电离层模型、硬件延迟估计精度影响；分析GLONASS伪距频间偏差特性,提出一种先验改正的方法,可有效消除伪距频间偏差对GLONASS卫星电离层延迟提取的影响；利用区域密集的参考站,建立基于单颗卫星的精化电离层模型,提出一种基于精化电离层产品的双频PPP快速收敛算法；利用中国区域实时数据流开发了一套实时电离层模型系统,为中国区域用户提供实时电离层延迟产品,可应用于中国区域单频定位及电离层变化监测。本文的主要研究内容包括：1.针对观测数据中钟跳问题,采用星间单差MW组合辅助非差MW、GF探测周跳、钟跳；对比分析了采用相位平滑伪距及组合观测值分解方法提取电离层延迟的精度；对中国区域电离层建模的布站选择进行分析,实验结果说明采用60个左右分布均匀的跟踪站即可获得较高的模型精度。2.利用Beidou/GPS双模实测数据建立了中国区域电离层模型。在当前卫星、跟踪站分布状态下,利用单模Beidou系统建立电离层模型在高、低纬度等边缘区域精度较低,在中纬度区域精度略低于单模GPS系统；Beidou卫星硬件延迟比较稳定,月平均变化不超过1.8ns。3.利用GLONASS并址站数据分析伪距频间偏差特性,认为不同卫星的接收机P1P2频间偏差变化稳定,但差异较大,且没有体现与频率有显著的线性关系：鉴于此,本文提出一种顾及频间偏差的多模电离层建模算法,首先估计出全部基准站相对某一测站的频间偏差并修正,再进行电离层建模,该算法可有效提高电离层模型及硬件延迟分离精度。4.为减少建模过程中造成的精度损失,建立基于单颗卫星的区域精化电离层模型。同时,基于精化电离层产品、宽巷UPD及单差残余硬件延迟产品,提出一种PPP快速收敛算法。利用精化电离层模型产品进行单频PPP解算,在N、E、U三个方向的平均RMS值分别为3.8,4.3和9.2cm；采用快速收敛算法,双频PPP用户的收敛时间显著降低,大约50%的流动站可以在5分钟内收敛到2dm。5.基于中国陆太网络实时数据流搭建了电离层实时估计服务软件,对实时数据处理中的定权、抗差、硬件延迟处理策略以及背景场的选择等进行探讨；通过与自估产品、CODE事后产品以及流动站反演等多种手段评估,在中高纬度地区实时模型精度可达2-3TECu,低纬度地区略低约为3-4TECu。利用实时电离层产品进行单频SPP、PPP定位,在中高纬度地区,单频SPP的3D偏差优于1.2m,低纬度地区优于2.2m；单频PPP可以达到平面0.2m,高程0.5m的精度；实时电离层产品有效监测了2013年076日磁暴期间中国区域电子密度的变化。更多还原

【Abstract】 Ionosphere is an important part of the space environment, understanding the structure of ionosphere and the rule of the ionospheric activity is an important basis to better understand human living environment. The research on ionosphere has very important value for the study of the relationship between solar-terrestrial space environment and atmosphere, and improving speed, positioning, timing and navigation system’s precision.Compared to the traditional ionosphere detection technology, GNSS detection technology can achieve continuous operation and high density. At present, the situation of the four Global Navigation Satellite System (GNSS) coexistence and development has been initially formed. With the improvement of multimode CNSS continuously tracking station number and observation data quality, the research on different characteristics of multimode GNSS system is conducted. It is necessary to carry out ionosphere modeling based on multimode GNSS system.Due to the inaccurate ionosphere model and pseudorange observation accuracy, it can only achieve decimeter to meter accuracy for single frequency single point positioning. Through the combination of different frequencies observations, double-frequency users can eliminate more than99%of the ionosphere error. But the linear combination observations magnified observation noises; it needs more time to implement dual-frequency PPP convergence. Improving the ionosphere model accuracy can help to improve positioning precision accuracy and accelerate convergence speed.Ionospheric delay has great influence on single frequency user, but the effect of real-time ionospheric delay correction provided by the navigation system is limited, it’s difficult to get high precision positioning results in this case. Therefore, using real-time observation data streams based on regional or global tracking station network, to build real-time ionospheric model can effectively improve ionospheric model accuracy. At the same time, using real-time ionospheric model can help to monitor the ionosphere, which is useful for the research on ionospheric disturbance and irregular change.In this paper, some problems in the ionosphere two-dimensional chromatography are discussed, such as observation pretreatment, ionosphere extracting method and so on; China regional ionospheric model is built based on observed GPS/Beidou data, and the multimode system impact on the ionosphere model, hardware delay estimation precision is analyzed; Through the analysis of numerous GLONASS observations, GLONASS pseudorange deviations between different frequency are analyzed; we put forward a prior correction method, which can effectively eliminate the pseudorange frequency deviation of GLONASS satellites in ionospheric modeling. For the local area with dense reference stations, we build the ionospheric model based on single satellite, and put forward a double-frequency PPP fast convergence algorithm. Based on Chinese real-time data stream, we develope a real-time ionospheric model system, which can provide real-time ionospheric delay for the regional user products in China. In this paper, the main research contents include:1. Against the clock jump problem, we use satellite single-difference MW combination together with undifference MW and GF to detect cycle slips; we analyze the difference of phase smoothed pseudorange method and combined observations decomposition method and their effects on ionospheric delay extracting; the stations selected strategies for modeling ionospheric delay is analyzed in China, experimental results show that using60uniformly distributed tracking stations can achieve considerable accuracy.2. We use the measured Beidou/GPS dual mode data to establish a Chinese regional ionospheric model; under the current satellites and stations distribution situation, the accuracy of ionosphere model using single Beidou system is low at high and low latitudes, especially for the edge region, and is slightly lower than single GPS system in middle latitudes; Beidou satellites hardware delay are relatively stable, the average monthly variation is less than1.8ns.3. We use numerous GLONASS observations to analyze the pseudorange deviations between different frequencies, the results show that GLONASS pseudorange ICB is very stable but with significant differences from each other, we put forward a new multi-mode ionospheric modeling algorithm considering the impact of inter-frequency bias;, which can effectively eliminate the pseudorange frequency deviation of GLONASS satellites in ionospheric modeling.4. In order to reduce accuracy losses in regional ionospheric modeling, we build the ionospheric model for each satellite in regional area; meanwhile, we propose a new PPP rapid convergence algorithms based on the ionosphere refined products, wide lane UPD and residual SD hardware delay. Using the refined products, the single-frequency PPP users can obtain precision positioning results, for the N, E, U directions, the average RMS values are3.8,4.3and9.2cm; using the fast convergence algorithm, dual PPP users significantly reduce the convergence time, about50%of the rovers can converge to2dm within5minutes.5. We build real-time ionospheric estimated service system based on Chinese real-time data stream, and discuss the weight, robust, hardware delay processing strategy and background field in real-time modeling. Compared with the self-assessment products, CODE afterwards products and other means, in the high and middle latitudes, the real-time model accuracy is about2-3TECu and slightly in lower latitudes, around3-4TECu. Using the real-time ionospheric products, for the single frequency SPP users, the3D deviation is less than1.2m, for the lower latitudes is than2.2m; the single frequency PPP users can achieve0.2m for the horizontal,0.5m for the vertical direction accuracy. We use the real-time products to analyze the electron density magnetic storms impact on the China’s regional ionosphere at DoY076,2013.更多还原