Research On Model And Algorithm Of Real-time Precise Point Positioning Combined With BDS-3/BDS-2/GPS

BDS-3(Bei Dou-3 Navigation Satellite System),opened on July 31,2020,has been widely used in many fields,such as national defense-security,transportation,disaster reduction and relief,and chip industry chain.It has become an important part of national infrastructure.This study focuses on researching products and data processing algorithms of BDS-3/BDS-2 for real-time PPP(Precise Point Positioning),including positioning error modelling and correction,data preprocessing,pseudorange and fractional cycle bias estimation,real-time PPP solution,and fixing ambiguity.The main research contents and conclusions are summarized as follows:(1)The data preprocessing technology is studied from the aspects of multi-GNSS(Global Navigation Satellite System)satellite selection,pseudorange noise reduction,clock jump and cycle slip processing.The reason why the low speed of optimal DOP(Dilution Of Precision)satellites selection algorithm is inapplicable for real-time data processing is analyzed.A fast-rotating partition satellite selection algorithm based on equal sky distribution is proposed,which significantly improved the satellite selection speed.The precision of pseudorange is analyzed,and a noise reduction algorithm based on the least square is presented and extended to multi-epoch for improving the precision pseudorange and initial coordinates for real-time PPP.For observation discontinuity caused by clock jump and cycle slip,a clock jump processing model is proposed based on clock drift difference between the predicted value by MPF(Measurement-based Polynomial Fitting)and the calculated value.A step cycle slip detection and repair algorithm based on ephemeris and error double constrain is proposed to effectively protect the continuity of phase observations for ground tracking stations under an active ionospheric environment.(2)Aiming at the problem of slow convergence caused by lots of parameters as estimating real-time satellite orbit,a data processing model based on long-time postprocessing high-precision POD(Precise Orbit Determination)integral parameters to predict satellite orbit as the real-time orbit is established.The influence of fixing ambiguity on POD is analyzed.The BDS satellite orbit is estimated based on ECOMC(Empirical CODE models 1 and 2 Combined)solar pressure model,and the model's applicability to BDS-3 satellite orbit is verified.The result shows that fixing ambiguity technology can improve the precision of BDS satellite orbit by about 28%.Compared with the ECOM(Empirical CODE Orbit Model)5 parameter model,the ECOMC model improves the POD precision of BDS by about 5%.Compared with the ECOM 9parameter model,ECOM 5 parameter model improves the POD precision of BDS by about 16%.The 4 hours real-time orbit prediction results show that the ECOMC model is slightly better than the ECOM 5 parameter model,while the ECOM 5 parameter model is better than the ECOM 9 parameter model.(3)Aiming at the problem of decreasing the solution speed and efficiency caused by a large number of constant parameters for estimating real-time clock offset,a clock offset solution strategy to process ionospheric-free observations based on SRIF(Square Root Information Filter)with good numerical robustness and the quick operation speed is established.The post-processing results show that the IGSO clock offset precision of BDS-2 and BDS-3 IGSO is 0.19 ns and 0.21 ns,respectively,and that of BDS-2 and BDS-3 MEO is 0.14 ns and 0.12 ns,respectively.The real-time results show that the clock offset precision of BDS-2 IGSO and MEO are 0.27 ns and 0.24 ns,respectively,and that of BDS-3 IGSO and MEO are 0.42 ns and 0.19 ns,respectively.(4)Aiming at the problem of prolonging the convergence time of real-time PPP caused by the difference of pseudorange reference,this paper established the spherical harmonic ionospheric model and estimated the DCB(Differential Code Bias)of different signal pseudorange of BDS.The results show that the B1I/B3 I and B1I/B2 I DCB of BDS-2 satellite are within ±15ns and ±20ns,respectively,and the difference of BDS-3 satellite B1I/B3 I DCB is up to 70 ns,ranging from-45 ns to 25 ns.The estimated DCB residuals of BDS satellite are less than 0.6ns taking CAS product as a reference.The long-term stability of satellite DCB is good,deducing that there is no need to estimate DCB in real-time.(5)To realize real-time PPP-AR(Precise Point Positioning Ambiguity Resolution),this paper establishes a pseudorange multipath at the satellite end for BDS and then avoids the rank deficiency of the normal equation by constraining the sum of the satellite terminal's fractional cycle bias(FCB)as zero.The model for estimating realtime NL(Narrow Lane)FCB based on post-processing WL(Wide Lane)FCB,realtime orbit,and clock offset is constructed,and then the real-time estimating of NL FCB is realized.The results show that the WL FCB in a few days can meet the precision requirements for post-processing and real-time PPP solution,deducing that there is no need to estimate WL FCB in real-time.There is strong time variability of NL FCB,concluding that it can only realize short-term prediction and should be corrected by short-term or real-time estimation results.The stability of real-time NL FCB is good within 4 hours,and the real-time products are in good conformity with post-processing products.(6)Aiming at the low ambiguity resolution success rate and slow convergence rate of traditional PPP-FAR(Full Ambiguity Resolution)caused by a large number of estimating parameters and their correlation,a PAR(Partial Ambiguity Resolution)algorithm based on the maximum change of check factor is proposed,including Ratio,ADOP(Ambiguity Dilution Of Precision),and Bootstrapping success rate.The PPP solution program,product availability,and PPP-F(P)AR algorithm are verified by static,simulated kinematic,and true kinematic data under post-processing and real-time model in five different combinations of BDS-3/BDS-2/GPS.The results show that the algorithm and program in this paper realize the high-precision PPP-AR solution under post-processing and real-time.The proposed PPP-FAR method can shorten the convergence time,speed up the ambiguity fixed speed,and improve the ambiguity epoch fixed rate and positioning accuracy for PPP.