| The Precision Point Positioning(PPP)technology of Global Navigation Satellite System(GNSS)is an important way to provide high-precision and high-frequency space-time reference information services due to its characteristics of single-machine operation,high positioning accuracy,flexibility and low cost.As one of the important GNSS error sources in positioning,the effective processing of ionospheric delay is of great significance to improve the accuracy and convergence speed of PPP positioning.Based on this,this thesis establishes regional ionospheric delay model for the ionospheric delay error,and conducts a series of theoretical research and algorithm effectuation of precise single point positioning with additional ionospheric constraints.The main research contents are as follows:(1)Based on the research on the mathematical model of precise single point positioning,detailed derivation and comparative analysis are conducted for the functional model and random model of the two combined models of Uofc and IF,and the UC non combined model.By estimating and analyzing the advantages and disadvantages of different models and the differences in parameter settings,the accuracy of each error model and the reasonable parameterization of the parameters to be estimated are studied and determined.To ensure the reliability of positioning results,PPP data preprocessing and quality control methods are introduced,and the PPP quality control process is summarized.(2)Based on the GNSS ionospheric delay modeling theory,the accuracy of extracting ionospheric delay measurements using the traditional carrier phase smoothing pseudorange method is easily affected by the length of the observation arc.In this paper,the ionospheric delay measurements under the satellite path are estimated and solved as unknown quantities using a nondifferential and non combined PPP method,making full use of the effective information provided by the original observation values,This weakens the amplification of observation noise caused by linear combinations and the impact of observation arcs on the accuracy of ionospheric delayed observations.The experiment uses the measured data from 15 stations of the Henan Geological Information Continuous Collection and Operation System in Henan Province for 7 consecutive days to separate the DCB parameters at the receiver end from the ionospheric delay parameters through spherical harmonic functions and polynomial functions,respectively,to establish a regional ionospheric delay model,At the same time,the Global Ionospheric Model(GIM) product released by the Center for Orbit Determination in Europe(CODE)was introduced as an external comparison source.The results showed that the two regional models had good consistency with the final GIM product released by CODE.The difference between the two regional models and the CODE product was mostly distributed within ± 4TECU,with the difference in daytime being greater than the difference in night,The residual RMS is maintained at around 2TECU except for the geomagnetic active period from November to April.During the geomagnetic active period,the maximum difference can reach 10 TECU,and the residual RMS exceeds 2.5TECU.(3)Starting from the mathematical model of non differential non composite PPP,a detailed derivation of the mathematical model of non differential non composite PPP with additional ionospheric constraints for single and dual frequencies was conducted.In response to the difficulty in determining the weights of virtual and actual ionospheric observations of non differential non composite PPPs with additional ionospheric constraints,this paper summarizes the methods for determining the prior variance of ionospheric constraint models,including commonly used prior variance determination methods(conventional constraints,spatiotemporal constraints,and gradual relaxation constraints)and weight search constraint methods.Based on the commonly used prior variance determination methods and weight search algorithm constraint methods,We adopted a weight factor search constraint method that takes into account the spatiotemporal changes of the ionospheric TEC.In order to evaluate the effectiveness of applying conventional constraints and weight factor search constraints that take into account the spatiotemporal changes of ionospheric TEC in PPP,this article uses measured data from five Henan Province geological information continuous collection and operation system stations for 7 consecutive days to conduct PPP experiments in four modes: single frequency with additional GIM products,dual frequency with additional GIM products,single frequency with additional regional ionospheric products,and dual frequency with additional regional ionospheric products.The results show that compared to traditional non differential non composite PPP models,The additional ionospheric constraint PPP model can effectively improve the convergence speed of PPP,and the degree of improvement mainly depends on the accuracy of external ionospheric products;The convergence speed of conventional constraint methods is slightly lower than the weight factor search constraint that takes into account the spatiotemporal changes of the ionospheric TEC,and its positioning accuracy cannot be guaranteed after PPP convergence.However,using the weight factor search constraint that takes into account the spatiotemporal changes of the ionospheric TEC can improve the convergence speed while ensuring the positioning accuracy after convergence;Considering the spatiotemporal variation of ionospheric TEC,the weight factor search constraint was applied to the single frequency additional ionospheric constrained PPP.The convergence time using GIM products was reduced by an average of 27.9 minutes(48.91%),and the convergence time using regional ionospheric products was reduced by an average of 31.26 minutes(54.80%).When using dual frequency additional ionospheric constraint PPP,the convergence time of GIM products was reduced by an average of 10.58 minutes(25.6%),and the convergence time of regional ionospheric products was reduced by an average of 14.12 minutes(34.18%). |
