Research On Key Technologies Of Precise Data Processing For GNSS Networks

In this thesis,the GNSS network is defined as a tracking and measurement network related to GNSS satellites,including tracking measurements of satellites at static or dynamic monitoring stations(receivers),(two-way)ranging between satellites and anchor stations(or others),and(two-way)ranging between GNSS satellites.It is of great significance for GNSS itself and for various scientific and engineering applications,by processing the measurement data,to determine the(high precision)relative position and time offsets among the network nodes(satellites and stations),i.e.orbit determination,positioning and time synchronization.This thesis focuses on the research of several key technologies of GNSS network data processing,including the following aspects: observation and satellite orbit modeling,integer ambiguity resolution,large-scale network solution and data processing for more complex GNSS networks with inter satellite link(ISL)or dynamic/kinematic stations involved.Major contributions are as follows.(1)A "one-step" strategy is proposed to estimate the antenna phase center offset(PCO)of BDS satellites by jointly processing BDS and GPS observation data and fixing the GPS satellite orbits and clocks to accurate known values,i.e.IGS products.Based on the obtained PCO estimates,the orbit accuracy of the IGSO and MEO satellites is increased by 10.5% and 70.5% respectively.(2)It is proposed that the SIMP correction model should be expressed with the nadir angle as the independent variable,and correction models for two types of satellites,MEO and IGSO,were constructed respectively.It's shown that,after SIMP correction,the quality of satellite wide-lane FCB estimates is significantly improved;so is the precision of relative positioning with BDS-2,if only the double-difference(DD)ambiguities belonging to non-GEO satellites are fixed.(3)GNSS network solutions considering the integer property of ambiguities is regarded as a generalized mixed integer optimization problem,where ?integer? is defined as a linear combination of integers.Based on the rank-deficit network adjustment theory,it is pointed out whether between-satellite single-difference(BSSD)or zero-difference(ZD)ambiguity parameters have integer property depends on the datum used to eliminate the rank deficits.In order to achieve integer-recovery clocks(IRCs)to support integer ambiguity resolution(IAR)in precise point positioning(PPP).A slightly improved algorithm for ZD ambiguity fixing is proposed and verified.Data processing with the proposed algorithm is quite similar to that using the ?DD approach?,except for the ambiguity fixing procedures themselves.(4)A new IAR method,i.e.to resolve integer BSSD ambiguities is proposed,which can also achieve satellite IRCs and is validated with experiments for GPS and multi-GNSS(GPS/Galileo/BDS)network solutions.GNSS network solutions with different IAR approaches are uniformly described as the least-square problem with fixed(DD,ZD or BSSD)ambiguities as constrain conditions.It is pointed out that in terms of improving the accuracy of solutions,the proposed BSSD IAR approach is theoretical equivalence with the existing DD and ZD approach.(5)Considering the difference of adjustment datum in subnet solutions,a method of satellite clock estimation based on the principle of block adjustment is proposed.This method can overcome the contradiction between the number of tracking stations and the computation efficiency in integrated multi-GNSS data processing,and improve the precision of clocks by combining the results of multi-subnet solution.(6)An approach of generating carrier pseudorange by fixing BSSD ambiguities in PPP is proposed.Based on the adjustment model with constrain conditions,it is pointed out that network solutions with carrier pseudorange observation generated by different methods is theoretically equivalent to that with original data.(7)With clock offsets expressed by piecewise polynomial with appropriate piecewise intervals,a general observation model that can be used for both simultaneous and non-simultaneous observed data is proposed.Based on the proposed model,without the need for deriving observations at selected times and forming geometry-free or clock-free combined observations which are necessary in the existing methods,the original one-way inter-satellite link(ISL)pseudorange data can be directly processed for orbit determination and time synchronization(OD&TS)in various mode,including: 1)using only ISL observation data,2)joint processing of ISL and anchorage station ranging data,3)joint processing of monitoring station and ISL observation data,4)joint processing of monitoring station,anchorage station and ISL observation data.Hardware delay biases of the ISL equipment at satellites(and anchor stations)are estimatable with mode 3(and mode 4).Experiments with the ISL data of BDS-3 satellites were carried out,which validate and demonstrate the feasibility and flexibility of the proposed method.For example,with data at 7 monitoring stations and ISL data,the averaged orbit accuracy of 8 BDS-3 satellites is 2 cm in radial,0.2 ns for satellite clocks,and the repeatability for ISL equipment hardware delay biases is better than 0.5 ns.(8)An alternative method to enhance OD&TS of navigation satellites using observation data onboard low-earth-orbiting(LEO)satellites is proposed,in which the motion of the LEO satellite is represented by a kinematic orbit rather than a dynamic orbit.Compared with the existing "dynamic" approach,the ?kinematic? approach is more computationally efficient,and would neither be affected by the force model errors such as high-order earth's gravity,atmospheric drag,solar radiation pressure,nor by satellite attitude maneuver and attitude instability.(9)A novel system for monitoring and tracking GNSS satellites is proposed,which consists of a number of static monitoring stations built on land and dynamic monitoring stations,e.g.mounted on anchored buoys placed on the international high seas.These marine dynamic monitoring stations can also enhance the OD&TS of navigation satellites.Based on this idea,any country,regardless of its territorial size,can build a globally distributed monitoring and tracking network to continuously track and monitor the GNSS constellation.The above technical achievements have been implemented in the SPODS software developed at Xi'an Institute of Surveying and Mapping.Some technologies can be used for data processing of high-precision GNSS network.Some technologies and concepts can be applied to the construction,operation or augmentation of a GNSS.