Theory And Methodology Of Ionospheric TEC Modelling And Differential Code Biases Estimation With Multi-GNSS

Ionosphere is an important error source in GNSS navigation and positioning,which can cause several meters to several hundred meters positioning error and seriously affect the navigation satellite system applications and services,especially for the huge number of single frequency users.But at the same time,GNSS provide a new technique for remote sensing of the ionosphere.Compared to traditional ionospheric monitoring technology(e.g.ionosonde,incoherent scatter radar and satellite radar altimeter),GNSS-based ionospheric monitoring technology has become the focus of current ionospheric monitoring research because of its high accuracy of ionospheric total electronic content(TEC)inversion,wide coverage,continuous mon.itoring of all-weather and measurable electronic content of 2000 km or more.Although GNSS ionospheric monitoring technology has been developed for nearly two decades and has made a large number of research results in ionospheric model and related applications,there are still a series of problems that need further study.In particular,with the development of global satellite navigation systems such as GPS,GLONASS,BDS and Galileo,it can bring unprecedented opportunities to the GNSS-based ionospheric research.But at the same time,it also can bring more challenges to ionospheric modeling because of more satellites,more signal types,complex constellations and massive of data.This thesis aims to study the theory and method of GNSS-based ionospheric TEC monitoring,and strive to establish a high precision ionospheric TEC model based on existing research to improve the application range of GNSS ionospheric TEC products.Firstly,the methods of improving the accuracy of ionospheric modeling are studied from two aspects:precisely extracting GNSS ionospheric observables and ionospheric modeling with multi-GNSS.Secondly,the estimations of GPS,GLONASS,BDS and Galileo differential code biases(DCBs)with high-accuracy based on the multi-GNSS-based ionospheric TEC model are realized.Then,in the application of ionospheric research,a new method(two-step method)for detecting the anomalous perturbation of ionosphere based on the time series method(ARIMA model)is proposed,which solves the problem that the detection accuracy of the traditional ionospheric anomaly detection methods are low.And on the basis of that,a set of GNSS ionospheric data processing and anomaly monitoring platform(GIMP)are developed independently by the author,which achieved daily generation of multi-GNSS-based ionospheric TEC products and multi-GNSS DCBs products,and routine monitoring of ionospheric anomalies.The main work and contributions of this thesis are listed as follows:(1)The basic principles and methods of GNSS ionospheric monitoring are expounded in detail.The main problems of current GNSS ionospheric research are summarized,and the influence of error of ionosphere monitoring by GNSS technique is analyzed.Finally,the paper summarizes the expression and calculation methods of the commonly used ionospheric mapping functions,and deeply analyzes the differences of different projection functions with satellite elevation and single layer height.(2)The method of extracting ionospheric observables based on ambiguity-fixed phase observables is proposed,and the extraction process and feasibility of the proposed method are described in detail.The precision and reliability of the proposed method and the traditional methods(eg.phase levelling code method and un-difference precise point positioning method)are analyzed by short-baseline/zero-baseline experiment and single frequency precise positioning experiment.Based on the analysis of single differences of the ionospheric observations obtained from pairs of co-located dual-frequency GNSS receivers,the results show that the accuracy of extracting ionospheric observable using PL-C method,UD-PPP method and PPP-Fixed method are 1.81,0.59 and 0.15 TECu,respectively.The maximum of the deviations for three methods are 5.12,1.68 and 0.43 TECu,respectively.Compared with PL-C method,the accuracies of the ionospheric observations using UD-PPP and PPP-Fixed method can improve by 67.3%and 91.7%,respectively.According to the positioning result of single-frequency PPP,the ionospheric TEC correction performance of PPP-Fixed method is best not only in terms of the convergence rate and positioning accuracy but also the reliability.In the meanwhile,no anomaly such as negative TEC observation occurs with PPP-Fixed method,and high precision of extraction can be guaranteed even for short-arc observations,which has important implication for the extraction of precise TEC based on short-arc observations obtained from low earth orbit(LEO)satellites in future.(3)The effects and contributions of multi-GNSS on the globally ionospheric TEC monitoring using the datasets achieved from the multi-GNSS experiment(MGEX)and the international GNSS service(IGS)tracking networks.The analysis and evaluation of ionospheric TEC modeling precision by means of multi-GNSS are conducted from several aspects.The results show that the accuracy of global ionospheric modeling with multi-GNSS has not been significantly improved recently except for few areas,which indicates that the contribution on the globally ionospheric modeling is limited for BDS and Galileo.Main reasons for this are that the number of visible satellites of BDS and Galileo is limited recently and the tracking stations of MGEX network mainly distribute in only a few areas though the number of stations is over 100.(4)The Chinese regional ionospheric TEC model is established based on the observation of three satellite systems(GPS/GLONASS/BDS)obtained from the Crustal Movement Observation Network of China.The result shows that BDS has the equivalent ability in the aspect of ionospheric monitoring as GPS and GLONASS.Thanks for the combination of multi-constellations satellite systems,the ionospheric modeling precision has been improved,particularly,it is highly improved in the area of regional edge.In the aspect of estimation precision of the DCB,the accuracy of the DCB estimation for satellites is significantly improved using multi-GNSS observations.By adding the observations of only GLONASS,only BDS and GLONASS+BDS to GPS,the precision of GPS satellites DCB estimation can improve by 12.3%,29.4%and 32.2%,respectively.Furthermore,by adding the observations of only GPS,only BDS and GPS+BDS to GLONASS,the precision of GLONASS satellites DCB estimation can improve by 13.8%,23.4%and 31.2%,respectively.Also,by adding the observations of only GPS,only GLONASS or GPS+GLONASS to BDS,the precision of BDS satellites DCB estimation can improve by 10.6%,10.4%and 13.6%,respectively.(5)Targeting at the present situation that the GLONASS inter-frequency code biases(IFCB)of the receiver has not been taken into consideration in ionospheric modeling with GLONASS observations,the effects of the GLONASS IFCB on ionospheric modeling and DCB estimation are studied.The results shows that with or without IFCB can cause a mean bias of 0.53?1.13 TECu(standard deviation:0.98?1.75 TECu)on ionospheric modeling.With the consideration of the GLONASS IFCB,the residuals of the GLONASS DCB estimation for C1C-C1P and C1P-C2P decrease from 0.193 and 0.434 ns to 0.045 and 0.153 ns,respectively.(6)In order to improve the accuracy of DCBs estimation,the DCBs of GPS,GLONASS,BDS and Galileo with high-accuracy are estimated based on the multi-GNSS-based ionospheric TEC model.The multi-GNSS DCBs estimated by the proposed method are compared with the DCB products from the German Aerospace Center(DLR)and Institute of Geodesy and Geophysics(IGG)over a 2-year span(from January 2015 to December 2016).The precision of inner-frequency DCBs estimation for GPS and GLONASS satellites are 0.1?0.2 ns and 0.2?0.5 ns,respectively.The precisions of intra-frequency DCBs estimation for GPS,GLONASS,BDS and Galileo are 0.2,0.3,0.3 and 0.5 ns,respectively.The monthly stability indices of inner-frequency DCBs for GPS and GLONASS determined by the proposed method are better than 0.032 ns and O.11ns,respectively.The monthly stabilities indices of intra-frequency biased for GPS,GLONASS,BDS and Galileo are better than 0.12,0.16,0.15 and 0.14 ns.In general,in the aspect of monthly stability,the DCBs estimated using the proposed method is comparable with that of DLR and IGG,or slightly worse than DLR and a little better than IGG.(7)A new method(two-step method)for detection of pre-earthquake ionospheric anomalies based on the time series method(autoregressive integrated moving average model,ARIMA model)is proposed,which solves the problem that the detection accuracy of the traditional ionospheric anomaly detection methods is low.The accuracy in predicting the TEC background values using the proposed method as well as the traditional ones is analyzed in details.The results show that the relative and absolute precisions of reference background values using ARIMA method are both better than the traditional ones.During perturbed period,ARIMA method can still guarantee high prediction accuracy.Finally,the detection performance is analyzed and compared between the proposed method and the sliding window method taking the Sumatra earthquake(January 10,2012)as an example.The results show that the anomalous days detected by sliding window method are obviously more than ARIMA method,and once an anomaly appears,the anomalies will always appear in the following days.What's worse,the positive and negative anomalies occur irregularly,however,the new method can successfully deal with it.