电离层延迟是制约导航和定位精度的误差源之一，实践证明高精度的电离层模型对于电离层延迟误差改正具有重要意义。随着北斗三号完成全球组网以及Galileo系统日益完善，卫星数达到甚至超过GPS，并且Galileo和BeiDou地面跟踪站超过200个，利用Galileo和BeiDou建立全球电离层模型已成为可能。并且不同系统卫星数量、星座类型和信号频率都存在差异，有必要对不同系统建立的全球电离层模型的一致性进行验证。鉴于此，本文研究了Galileo和BeiDou系统单独建立全球电离层模型的可行性，并在此基础上建立了多系统GNSS高精度全球电离层模型。此外，将新建模型应用于PPP中进行分析评估，具体研究工作和成果如下：

（1）建立了GPS、GLONASS、Galileo和BeiDou各单系统全球电离层模型，发现不同系统间的VTEC具有较高的一致性，均能较好地反映全球VTEC的时空分布。其中Galileo和GPS系统的一致性最高，Galileo相对于GPS VTEC的平均Bias为0.02 TECU，2019~2020年STD均小于1.5 TECU。而BeiDou受限于西半球跟踪站数量偏少，全球范围内平均偏差为-0.11 TECU，STD在1~2.5 TECU之间，与GLONASS相当。

（2）建立了GPS、GLONASS、Galileo和BeiDou四系统组合的全球电离层模型，结果表明多系统较各单系统的VTEC差异主要存在于海洋和两极等跟踪站稀少的地区。多系统较各单系统RMS均不同程度的降低，其中较GPS降低0.3 TECU，较GLONASS降低1.6 TECU，较Galileo降低0.6 TECU，较BeiDou降低2.5 TECU，特别在海洋等单系统观测值缺失地区尤为明显。

（3）利用事后GIMs产品、GNSS dSTEC观测值和CAS DCB，从多角度对新建模型精度进行验证，结果表明单系统模型已经具有很高的精度，BeiDou系统在东半球尤其亚太地区精度与其他系统相当。多系统融合在提高模型精度的同时，可显著增加模型的可靠性。本文将建模产品与IGS各分析中心产品的对比，结果表明建模产品的精度与各个分析中心的产品一致，估计的各系统卫星DCB与CAS DCB产品非常接近且多系统DCB的一致性更好，验证了本文建模方法和数据处理策略的正确性。

（4）以标准单频PPP为参考，评估了本文不同电离层建模产品的PPP定位效果。在定位精度方面，本文建模得到的五个电离层产品定位误差的平均RMS值相差不大，与国际上CODG、ESAG、IGSG和WHUG结果相当。在收敛时间方面，多系统产品（GREC）和Galileo较标准单频PPP平均收敛时间缩短47.5%和46.7%，略差于CASG（57.1%）和WHUG（48.6%），但优于CODG（46.2%）、ESAG（44.7%）、IGSG（39.3%）、JPLG（21%）和UPCG（37.2%）。BeiDou的收敛时间改善效果略差（35.3%），但仍优于JPLG（21%）。总体上，本文建模产品性能与ESAG、IGSG、JPLG和UPCG相当。

Ionospheric delay is one of the error sources that restrict the accuracy of navigation and positioning. Practice has proved that the high-precision ionospheric model is of great significance for the correction of ionospheric delay errors. With the completion of the BeiDou-3 global network and the increasing improvement of the Galileo system, the number of satellites reached or exceeded GPS, and there were more than 200 ground tracking stations of Galileo and BeiDou, it is possible to use Galileo and BeiDou to build a global ionospheric model. In addition, there are differences in the number of satellites, constellation types, and signal frequencies of different systems. It is necessary to verify the consistency of the global ionospheric models established by different systems. In view of this, this paper studies the feasibility of the Galileo and BeiDou systems to establish a global ionosphere model separately, and builds a multi-system GNSS high-precision global ionosphere model on this basis. In addition, the new model is applied to PPP for analysis and evaluation. The specific work and results are as follows:

(1) The global ionospheric model of GPS, GLONASS, Galileo and BeiDou each single system has been established, and it is found that the VTEC of different systems has high consistency, and they can better reflect the temporal and spatial distribution of global VTEC. Among them, Galileo and GPS systems have the highest consistency. The average Bias of Galileo relative to GPS VTEC is 0.02 TECU, and the STD from 2019 to 2020 is less than 1.5 TECU. Due to the small number of tracking stations in the Western Hemisphere, average deviation between GPS VTEC and BeiDou VTEC is -0.11 TECU, and the STD is between 1~2.5 TECU, which is equivalent to GLONASS.

(2) A global ionospheric model combining the four systems of GPS, GLONASS, Galileo and BeiDou has been established. The results show that the VTEC difference between multi-systems and single systems mainly exists in areas where tracking stations are scarce, such as oceans and poles. The RMS of multi-systems is reduced compared with each single system, which is 0.3 TECU lower than GPS, 1.6 TECU lower than GLONASS, 0.6 TECU lower than Galileo, 2.5 TECU lower than BeiDou, and the decline is particularly obvious in areas where single-system observations are missing, such as oceans.

(3) The accuracy of the new model was verified from multiple aspects using post-event GIMs products, GNSS dSTEC observations and CAS DCB. The results showed that the single-system model has high accuracy. The accuracy of BeiDou system in the eastern hemisphere, especially in the Asia-Pacific region, is comparable to other systems. Multi-system fusion can significantly increase the reliability of the model while improving the accuracy of the model. The comparison between the modeling products in this thesis and the products of IGS analysis centers shows that the accuracy of the modeling products is consistent with the products of various IAACs, the estimated DCB of each system satellite is very close to the CAS DCB product, and the consistency of the multi-system DCB is better, which verifies the correctness of the modeling methods and data processing strategies in this article.

(4) Taking the standard single-frequency PPP as a reference, the PPP positioning effect of different ionospheric modeling products in this dissertation is evaluated. In terms of positioning accuracy, the average RMS value of the positioning errors of the five ionospheric products obtained in this paper is close, which is equivalent to the results of CODG, ESAG, IGSG and WHUG. In terms of convergence time, the average convergence time of multi-system products (GREC) and Galileo is 47.5% and 46.7% shorter than standard single-frequency PPP, slightly worse than CASG (57.1%) and WHUG (48.6%), but better than CODG (46.2%), ESAG (44.7%), IGSG (39.3%), JPLG (21%) and UPCG (37.2%). BeiDou's convergence time improvement effect is slightly worse (35.3%), but it is still better than JPLG (21%). In general, the performance of the modeled products in this thesis is comparable to that of ESAG、IGSG、JPLG and UPCG.