Research On Multi-GNSS Precise Point Positioning Technology

Precise point positioning(PPP)can get the high accuracy location information with one single receiver all over the world,but there are two certain bottlenecks,including the limited number of available satellites and the long initialization time.With multi-GNSS era coming,the considerable increasing number of navigation satellites and the new multi-frequencies observation bring possibilities to resolve these two problems,meanwhile some new challenges are brought by the way.Based on the analysis about some certain key problems of multi-GNSS PPP technology,systemic and deep researches have been done focusing on the GNSS cycle-slip detection,the processing of multi-GNSS PPP inter-system bias and the parameter estimation etc.(1)Un-difference cycle-slip detection methods were researched,which were suitable for multi-GNSS PPP.Firstly,a new cycle-slip detection algorithm based on TurboEdit was proposed.In this algorithm,cycle-slip scalars were accurately estimated by elevation angle weighted recursive smoothing,and more realistic thresholds,which related to the error distribution of cycle-slip scalars,were calculated by introducing the sliding window method.At last,the cycle-slip searching model based on difference between epochs was introduced,which could determine the quality of cycle-slip by judging ratio and repaire the observation with the information of cycle-slips.(2)Inter-system bias(ISB)in multi-GNSS PPP technology had been deeply studied.Using the MGEX data,four varied schemes were designed,in which the processing methods of ISB for both carrier phase and pseudorange observations were different.Comparing the results,it was found that ISB effects on carrier phase and pseudorange observations were obviously different and should be considered.Form the analysis results,we suggest to do the same ISB process for these two kinds of observations to get higher accuracy results with the least number of parameters.(3)A parameter estimation method with the constraint of velocity based on sequential filtering algorithm was proposed.The velocity series was inversed from the location information for a period of time before the current epoch,so the random model could be improved by changing the process noise with the average and standard deviation of the velocity series.Testing by the MGEX data,it was found that the new method could produce more reasonable results which matched the actual status with the help of the previous positioning information.(4)A multi-GNSS PPP software called MGP was programmed by FORTRAN and was verified by testing.Utilizing the observation data from Australia and Europe area to test the multi-GNSS PPP ability for GPS,GLONASS,BeiDou,Galileo etc.,it turned out that GPS PPP results were still the most precise and stable;although PPP based on GLONASS could be done all over the world,its robustness was the weakest;the stability of BeiDou PPP was high,following GPS in Asia pacific region,because lacking the relevant documents for accuracy phase center correction the precise of BeiDou position was affected;with limited number of satellites,Galileo could not do PPP by itself.Combining GPS,GLONASS and BeiDou,the number of available satellites and the stability of PPP could be improved obviously;because of the limitation of our PPP algorithm,the participation of current Galileo would influence PPP seriously for its limited number of satellites.