| With the widespread promotion and application of global navigation satellite system(GNSS)technology in the mass market,emerging applications such as unmanned aerial vehicle navigation,automatic driving,and precision agriculture have put forward higher requirements for reliability,accuracy and availability of spatiotemporal information services.Precise point positioning(PPP)technique has the advantages of portability,flexibility,low-cost and wide operation range,which can provide high-precision location services for the majority of GNSS users.However,the traditional PPP technique still takes a long convergence time to achieve centimeter-level positioning accuracy,which is difficult to apply to some time-critical applications.Recently,with the development of multi-frequency and multi-constellation GNSS and the deployment of ground-based enhanced stations,a regional augmentation PPP(RAPPP)technique has emerged which utilizes the atmospheric corrections provided by regional reference networks to effectively shorten the PPP convergence time,thus making it possible to realize the instantaneous ambiguity resolution.Therefore,this article focuses on the key technologies of GNSS rapid precise positioning with regional augmentation in urban environments,and is devoted to the construction of multi-system RAPPP model,multi-frequency PPP rapid ambiguity resolution,single-frequency and dual-frequency data fusion processing of low-cost devices,and positioning performance optimizations in complex scenarios.This thesis establishes a set of high-precision location service solutions for urban environments based on the RAPPP technique.The main works and contributions of the thesis are as follows:(1)This article systematically describes the RAPPP mathematical model,the source of un-differenced observation errors,the classification of un-differenced observation correction and the estimation method of parameters.Starting from GNSS observations and models,this thesis summarizes the sources of various un-differenced observation errors and their processing strategies in the RAPPP method.The classification principle and interpolation method of un-differenced observation corrections,as well as the Kalman filter parameter estimation methods in GNSS kinematic positioning are introduced.(2)Multi-system RAPPP algorithm models are deduced based on GPS/Galileo/BDS(GEC)raw observations.Relying on the abundant observation resources of ground-based reference stations,an un-differenced correction generation method using fixed single-differenced ambiguities is designed,wherein the generated phase correction accuracy is better than 3 cm in a 40 km scale network.The positioning results show that the PPP fixed performance and epoch availability of the only-GPS RAPPP scheme gradually decrease with the increase of the satellite cut-off altitude angle.For the GEC scheme,a sufficient number of visible satellites and a good spatial configuration can guarantee the continuity and reliability of user positioning,the fixing rate and epoch effectiveness can be stably maintained at 99%,exhibiting an improvement of 10% compared to that of the only-GPS scheme.In terms of positioning accuracy,the RAPPP method proposed in this paper can reach the positioning accuracy corresponding to that of existing regional augmentation GNSS positioning solutions,wherein the average positioning accuracy of the GEC fixed solution can be reached at1～3 cm.(3)A multi-frequency single-differenced rapid ambiguity resolution is proposed based on the un-combined RAPPP model.Thanks to the full completion of BDS-3,an un-combined BDS ambiguity resolution model is constructed with the B1 C,B2a new signals and the B1 I,B2I,B3 I legacy signals.The high-precision extra-wide-lane and wide-lane ambiguity constraint equations combined with multi-frequency observation information can significantly improve the convergence and fixed speed of narrow-lane ambiguity estimates.It is found that the BDS five-frequency scheme can achieve an instantaneous ambiguity resolution,and the average positioning accuracy is improved by 9～15% compared with that of the dual-frequency scheme.Moreover,this thesis also analyzes the positioning performance of the GPS/Galileo/BDS multi-frequency RAPPP scheme under different urban scenarios.When under the open-sky conditions,the multifrequency scheme can maintain over 98% of the fixing rate and epoch effectiveness,wherein the positioning accuracy can be stabilized at 1～3 cm.In the complex scenario,the availability,stability and reliability of the RAPPP decrease,but the performance of dual-frequency RAPPP can be improved by utilizing multi-frequency observations.The results show that the fixing rate can be increased from 76.37% to 84.27%,and the positioning accuracy of the multi-frequency fixed solution is 3～8 cm,exhibiting improvements of 26～65%.(4)A set of low-cost RAPPP rapid ambiguity resolution is established.By analyzing the GNSS observation quality of low-cost devices,the observation model and stochastic model of RAPPP method are improved and a hybrid single/dual frequency PPP ambiguity resolution is proposed.On the one hand,for different GNSS chips,the RAPPP performance of ublox-F9 p and Xiaomi Mi8 GNSS chips is discussed,wherein the positioning accuracy of the ublox-F9 p scheme is improved by 38～75% by adopting the new observation model,and it is stable at the centimeter level in an open environment.Meanwhile,a more obvious improvement has been found in the Xiaomi Mi8 scheme,fixing rate is increased by 20%,and positioning accuracy is improved by38～56% wherein 85.13% of the fixed epochs can obtain horizontal accuracy of better than 5 cm.On the other hand,for different low-cost GNSS antennas equipped with the same ublox-F9 p receivers,the RAPPP performance of the patch-antenna and helicalantenna schemes are discussed respectively.The results show that both low-cost antenna schemes can achieve centimeter-level positioning accuracy in a good observation environment.While for complex urban environments,the fixing rate of RAPPP scheme is reduced by about 20～40%,the positioning accuracy of fixed solution drops to 1～6dm.If using a survey-grade antenna(such as the Trimble antenna),a better positioning performance with a fixing rate of 90% and horizontal accuracy within 3 cm is achieved.(5)Due to the poor positioning performance of low-cost GNSS devices in complex environments,a low-cost RAPPP performance optimization strategy is proposed from two aspects of NLOS(non-line-of-sight)observation quality control and INS(Inertial Navigation System)enhanced GNSS positioning.Under the condition of only-GNSS raw observations,the prior information of carrier-to-noise ratio and pseudo-range observation residual can be used to weaken the influence of NLOS observation errors and improve the horizontal accuracy of low-cost fixed solutions from decimeter level to centimeter-level.Especially obvious for the helical-antenna scheme,the fixing rate is increased by 10%,and the positioning accuracy is improved by about 70%.When using additional INS data,the utilization of low-cost GNSS observations and the error proportion of ambiguity resolution can be improved by the RAPPP/INS combination.For two low-cost antenna solutions,the fixing rates are increased to 84.57% and 87.32%,respectively,wherein 90% horizontal accuracy and 70% elevation accuracy of fixed solution can attain centimeter levels. |
