| Network Real-Time Kinematic(NRTK)is a ground-based enhanced positioning technology that provides real-time centimeter-level positioning services within the netw ork coverage composed of Continuously Operating Reference Stations(CORS).Compared with traditional Real-Time Kinematic(RTK)technology and Precise Point Positioning(PPP)technology,NRTK has the advantages of a broad range of services without the requirement for users to establish a reference station,rapid convergence speed,and wide receiver compatibility.In recent years,in addition to traditional basic mapping and scientific research services,NRTK has started providing location services with varying precision requirements for other fields,such as UAV logistics,autonomous driving,high-precision mobile phone positioning and power grid inspection,etc.The high-precision positioning information provided by NRTK will become an indispensable requirement for human beings.On the other hand,with the comprehensive development and operation of China’s BDS and Europe’s Galileo and other systems,as well as the modernization of the United States’ GPS and Russia’s GLONASS,and the continuous construction of a large number of reference station networks in China and the world.The new-generation distributed NRTK system with the ability to rapidly process large-scale CORS network multi-frequency and multisystem GNSS data has become a current research focus.The development of a new generation of NRTK systems is primarily challenged by two theoretical and technical problems.The first is to solve major theoretical and technological problems in the shared technologies of the new generation(large-scale)and conventional(local/regional)NRTK in a more efficient way.Currently,the theoretical system of NRTK technology is relatively complete,and its practical application is relatively mature;however,some practical application issues and bottlenecks,such as the applicability of the ambiguity resolution(AR)method of long baseline between reference stations in the regions with varying ionospheric activity,still require extra research.Due to the substantial elevation difference in the reference station network with complex terrain,the interpolation accuracy of tropospheric delay region modeling is deteriorated,and the positioning accuracy of users is decreased.Moreover,the single-differenced baseline resolution model can provide the single-differenced atmospheric delay,but there is no research report discussing the effect and characteristics of its application in NRTK,and few apply the FDMA integer estimable theory to NRTK to extend the compatibility of GLONASS and other CDMA systems.The second is to investigate the NRTK architecture and comprehensive processing platform for efficient,distributed and unified operation of a large-scale(such as a country-wide)network based on the collaborative operation requirements of the server and the terminal of the large-scale real-time dynamic precision positioning technology.The CORS networks of provinces and cities in China operating autonomously,and the lack of uniform administration and data exchange of decentralized services,which restricts the number of concurrent users with increasing service requests.Thus,it is important to conduct research on a large-scale NRTK data processing system based on a distributed grid,and discuss and design the preliminary framework of the system platform.Therefore,based on GNSS NRTK theory,rank deficiency theory and integer estimation theory,this study focuses on the AR method of medium and long baseline between reference stations,tropospheric modeling interpolation in complex terrain areas,single-differenced baseline resolution model,and the application of FDMA integer estimable theory in NRTK.The preliminary design and development of a NRTK software platform,as well as the assumption of large-scale distributed grid data processing,are being investigated.The main work of this paper are as follows:1.An ionosphere-weighted three-carrier AR(TCAR)method for the long-baseline network reference stations is proposed,and the effect of different combination observations applied to the AR in the active and inactive areas of the ionosphere is exploredFast and reliable ambiguity resolution between reference stations is indispensable in network RTK.Multi-frequency GNSS observations are very beneficial to the AR of NRTK.However,in ionospheric active regions and long baseline situations,the uncertainty of ionospheric delay fluctuations will lead to a decline in the efficiency of reference station TCAR.In order to improve the reliability of the sub-optimal ultrawide lane AR of the north-south and east-west baselines,the ionospheric amplification coefficients and the success rate of AR of different three frequency combination observations in BDS.GPS and Galileo were analyzed in this study,and the combination observations with the minimum influence of ionospheric delay were excavated.And a TCAR method with additional ionospheric constraints is proposed.Experiments are conducted with two sets of three frequency GPS/Galileo/BDS data collected from active and inactive areas of the ionosphere.The results indicate that,in comparison to the conventional method,the proposed ionosphere-weighted TCAR can effectively increase the AR efficiency of NRTK reference stations in both active and inactive areas of the ionosphere.2.An improved linear tropospheric delay interpolation and an improved low-order tropospheric delay interpolation method are proposed,which are more applicable and easier to implement to the NRTK triangular unit for the reference station network with complex terrain and large elevation difference.The improved methods eliminates the tropospheric sy stem error caused by the height difference and enhances the precision of the tropospheric delay interpolation.Because of the configuration of the reference station network,especially in areas with complex terrain,the height difference between the reference stations will inevitably fluctuate substantially,resulting in a considerable elevation difference between the user station and the reference station.The traditional NRTK interpolation method only considers the distribution of tropospheric delay on the plane and ignores the strong correlation between tropospheric delay and height factors,which leads to an unmodeled systematic error in the tropospheric interpolation at the user side.To improve the accuracy of tropospheric interpolation and make it easier to implement in NRTK,two tropospheric interpolation models are proposed,modified linear interpolation(MLIM)and modified low-order surface interpolation(MLSM),which are based on the traditional linear interpolation model and the traditional low-order surface interpolation model,respectively.MLIM employs the tropospheric global prior model in advance to eliminate the influence of systematic bias in the elevation direction on the interpolated baseline,and then executes user tropospheric linear interpolation.MLSM adds a set of constraint equations to the low-order surface interpolation coefficient resolution model in order to obtain accurate interpolation model coefficients that eliminate tropospheric system errors and restore the actual distribution characteristics of the user station troposphere in space.Both methods are simple to implement in the Delaunay Triangulated Irregular Network(DTIN),which is widely used in network RTK,and have a high practicability.3.Based on single-differenced Common-clock(CC),Decoupled-clock(DC)baseline resolution models and FDMA integer estimable theory,a singledifferenced NRTK method compatible with FDMA integer estimability is proposedConsidering that the basic model of traditional NRTK is typically based on doubledifference observation equations,and the double-difference form of atmospheric delay has issues such as strong correlation,unified datum,and differences with actual atmospheric physical characteristics,it will not be favorable to regional atmospheric modeling.Therefore,this study investigates the effects of applying the CC and DC single-difference baseline resolution models to NRTK,and in-depth study of the features of CC and DC single-difference models.By analyzing the receiver bias results extracted from the model,it is discovered that there is a temperature-related intraday time-varying characteristic,and by comparing and analyzing the results of the atmospheric delay resolution,it is found that the DC single-difference model,which considers the receiver bias as a time-varying parameter,is proven to be more suitable for NRTK baseline atmospheric delay extraction,area modeling,and user positioning.On the other hand,the unique FDMA signal system of GLONASS makes it difficult for AR,and it is incompatible with the CDMA system during data processing.This study will learn from FDMA integer estimable theory and apply it to NRTK to realize the integration of FDMA integer estimable theory and CDMA ambiguity fixing.With the use of key indicators such as position dilution of precision,ambiguity dilution of precision,user positioning error,and time to first fix,it is demonstrated that the addition of GLONASS to NRTK positioning services can be beneficial.4.Design and development of a post/real-time multi-frequency and multiGNSS network RTK software platform(APM-NRTK),and conduct preliminary exploration on large-scale NRTK related technologiesGNSS data processing software is an important platform for the research and application of NRTK theory and algorithm.Based on the research of NRTK basic theory algorithm and data processing strategy proposed in this study,a set of multi-frequency and multi-GNSS post/real-time NRTK software platform APM-NRTK is designed and developed.Moreover,in view of the shortcomings of the conventional NRTK system,such as the difficulty of cross-region service and the limited number of concurrent users,the large-scale NRTK service architecture based on distributed grid is preliminarily discussed and designed,and the large-scale CORS service system combining distributed network and grid VRS technology is proposed.The APM-NRTK software platform has high scalability and positioning application capabilities,and can also provide technical support for BDS/GNSS large-scale,one-network,rapid and precise positioning-related theoretical research,which creating a solid foundation for the future research work. |
