| As a new global positioning technology developed in the late 1990s, precise point positioning (PPP) integrates the technical advantages of GPS standard point positioning and GPS differential positioning. The observation information can be multiplied and the redundancy of parameter estimation can be greatly increased with multi-system global navigation satellite system (GNSS, including GPS/GLONASS/BDS/GALLIEO) combination. As a result, the positioning accuracy and reliability can be significantly improved. For ambiguity-fixed PPP, the integer property of carrier phase observation could be considered and employed and then greatly impove the positioning accuracy, especially in east component. Additionly, more information for solution quality check can be provided by PPP ambiguity resolution (AR). For these reasons, multi-system combined PPP and PPP ambiguity resolution have been a hot research topic in recent years. After a rapid development for about ten years, the GNSS system construction has witness major progress and the basic theory and technology program of PPP AR has been well studied. However, the studies on multi-system PPP and PPP AR are conducted individually. For GPS PPP AR, currently, only the GRGS integer-recovered satellite clock products provided by CNES can be used. There are no other related products provided with open access for PPP user. The lack of integer satellite products has greatly limited the research and application of PPP AR. Furthermore, at user end, the time to first fix (TTFF) for a PPP user is generally longer than 30 minutes, which is much longer than that of network-RTK. Such a long TTFF has become a bottleneck of PPP AR application. Therefore, along with the rapid development of PPP technology, it is urgent to exploit a GPS FCB server system in server end and study the algorithm on accelerating the TTFF of PPP AR at user end.Aiming at the problems mentioned above, this theies is intended to:(1) at server end, fully and thoroughly study the estimation of GPS/BDS FCB with high-accuracy; analyse the factors affecting the accuracy of FCB estimation; seek for a systematic quality assessment methodology for FCB products; finally exploit a GPS FCB server system which can provide the FCB products with open access for PPP users all over the world. This server system is expected to promote the further study and application of PPP AR, (2) at used end, develop a PPP partial ambiguity resolution algorithm to increase the probability of achieving ambiguity-fixed solution when full AR fails. The performance on TTFF and epoch fixing rate is aimed to be improved with PPP PAR. In addition, we expect to extend the GPS PPP AR to GNSS PPP AR from two aspects. The first one is realizing dual-system PPP AR with combined GPS and BDS observation and the second one is aiding the single- and dual-system PPP AR with GLONASS observation. Fixing GLONASS PPP ambiguity is still considered difficult because there are satellite/frequency/receiver-type-specific inter-frequency biases in GLONASS pseudorange and carrier phase measurements. The performance of GNSS PPP AR would be fully analysed.The main work and contributions of this thesis are listed as follows:(1) The basic theory of GPS PPP and AR methods are summarized systematically. The detailed error component of FCB is deduced from the original functional model. The FCB estimation method based on PPP is improved with iterative least squares. A GPS FCB server system is exploited with this improved FCB estmation method. The quality of GPS FCB is fully assessed by internal as well as external indexes. For FCB estimation, the average usage of GPS WL and NL ambiguities is 93.7% and 89.9%, respectively. The percentage of the WL and NL ambiguities residuals within 0.2 cycles is 94.2% and 93.7%, respectively. The comparison of our FCBs with those produced by CNES indicates that our FCBs have a good consistency with the CNES. For wide-lane FCB, almost all the differences between the two products are within +/-0.05 cycles while for narrow-lane FCB,97.4% of the differences are between +/-0.075 cycles. We demonstrated that PPP AR using these FCB products is applicable for GPS measurements collected under various environments. Three examples were provided. Compared with conventional ambiguity-float PPP, the ambiguity-fixed PPP using these FCB products improves the mean position RMS of static PPP results by 44.4%, 28.6% and 25.0% in the east, north, and up, respectively. A kinematic positioning test with observation of 80 minutes achieved a position accuracy of better than 5 cm at the one-sigma level in all three coordinate components, with an improvement of 78.2%,20.8%, and 65.1% in east, north, and up, respectively, over the ambiguity-float results. The RMS of LEO PPP improved by about 23%,37% and 43% for GRACE-A and GRACE-B, in radial, tangential and normal directions, respectively, when the AR is applied to the same data set.(2) Considering that for conventional PPP full ambiguity resolution (FAR), it is not easy to fix all ambiguities during the initialization period because that the ambiguities are strongly correlated with each other and the accuracy of some ambiguitis may be not high enough, a partial ambiguity (PAR) resolution strategy for PPP is proposed. This PAR method can automatically try to select an ambiguity subset from the decorrelated ambiguity vector which can satisfy both the requirement of bootstrapping success rate and ratio-test. When FAR fails it is still possible to obtain an ambiguity-fixed PPP solution with PAR. Resutls demonstate that, compared with those of FAR, the TTFF can be significantly shorten and the fixing rate can be greatly increased with the proposed PPP PAR method. In static PPP, the average TTFF can be reduced by 19.8%, from 25.2 minutes for FAR to 20.2 minutes for PAR. In kinematic PPP, the average TTFF is reduced by 20.9% from 39.1 minutes for FAR to 30.9 minutes for PAR. The average fixing rate of PAR is increased from 83.4% to 97.7% in static PPP and from 77.6% to 94.7% in kinematic PPP.(3) The system-related error processing strategy for BDS is summarized and their influence on BDS PPP and FCB estimation is analysed. Without BDS code bias variation correction, the average BDS WL ambiguity usage would decereace from 91.8% to 80.4% and the RMS of WL ambiguity residuals would increace from 0.106 cycles to 0.113 cycles. The BDS FCB estimation based on GPS+BDS PPP solution is proposed to increase the usage rate of BDS sites, improve the accuracy of BDS PPP ambiguity-float ambiguity, and finally improve the BDS FCB accuracy. The time stability of BDS WL/NL FCBs is analysed. Experimental analysis shows that the change of the daily BDS WL FCBs is less than 0.1 cycles for about 30 days. The difference between two adjacent 15min sessions is usually not over 0.1 cycles; 91.5% of the differences are between -0.075 to 0.075 cycles for NL FCBs. Therefore, we estimated the BDS WL FCBs on a daily basis and the BDS NL FCBs for a 15-min time interval. A poisoning test also shows that our BDS FCBs can be used in BDS kinematic PPP to improve the accuracy of position.(4) An adaptive fusion of the single-differenced GPS and BDS ambiguities are realized with the proposed partial ambiguity resolution. Using the GPS+BDS FCB estimations, GPS-only, BDS-only and GPS+BDS combined PPP AR are performed. The improvement of AR performance, including TTFF and fixing rate, brought by dual-system PPP AR is analyzed in detail. For BDS PPP-AR, several hours are required to achieve the first fixed solution, and the fixing rate is usually less than 30% in both static and kinematic PPP tests. For GPS PPP-AR, the TTFF is approximately 20 min and 30 min, with fixing rates of 98.6% and 97.0%, in static and kinematic PPP, respectively. Additional observations from the other system are beneficial to the single GPS- or BDS-only PPP AR. In general, GPS+BDS combined PPP with ambiguity fixing shows the best performance in terms of both TTFF and the fixing rate. The average TTFF reaches 16.9 min and 24.6 min, with fixing rates of 99.5% and 99.3%, in static and kinematic GPS+BDS PPP, respectively.(5) A GLONASS-aid PPP AR strategy is proposed to further improve the proforamance of PPP AR. Results show that with the aid of GLONASS, the TTFF of PPP AR in various modes (single- or dual-system, static or kinematic) can be further reduced by about 10%-20%. The fixing rate can alwayse be signicicantly increased. The best PPP AR performance can be achieved with GLONASS-aid and GPS+BDS ambiguity fixing. |
PDF Full Text Request