Research On Performance Improvement Methods Of Precise Orbit Determination For Low Earth Orbiters Using GNSS Observations

Precise orbit determination(POD)is a prerequisite for low Earth orbit(LEO)satellites and their formation to complete specified mission targets.It is prospective to improve the POD performance for the successful implementation of their missions and expand the field of their applications.As the LEO missions gradually develop toward refinement,diversification and complexity,new requirements and challenges are put forward for the conditions,modes,accuracy and reliability of POD.This thesis mainly focuses on the key issues in LEO satellite POD using onboard GNSS measurements,and proposes some mathematical methods for improving orbit determination performances from the perspectives of fine-scale modeling and compensation of perturbation forces,modeling and calibration of system errors,and efficient processing and estimation of parameters.The contributions are summarized as follows.(1)The fine-scale modeling and calibration method of orbital perturbation forces for LEO satellite is studied.The effects of refined atmospheric drag modeling and atmospheric density model optimization on the POD and orbit prediction(OP)performances of large-scale spacecraft are first analyzed.Then,the influences of orbital maneuver modeling on the POD for LEO satellites are investigated.The results show that,after using a macro-model to refine atmospheric drag modeling,the 3D RMS of estimated empirical accelerations in the reduced-dynamic POD of Tiangong-2 spacecraft is substantially reduced by 37%.The POD accuracy validated by satellite laser ranging(SLR)data is improved by 5%,and the OP accuracy after 24 h is improved by 35%.Meanwhile,the OP accuracy after 24 h is significantly improved by using the optimized empirical atmospheric density model.In addition,with the proper thrust modeling of orbital maneuver using constant acceleration parameters,the effects caused by orbital maneuver on POD for Tianhui-2 formation-flying satellites can be largely eliminated.(2)The accuracy of single-frequency(SF)orbit determination for LEO satellite is affected by the systematic errors existed in the group and phase ionospheric correction(GRAPHIC)observations.In order to solve this problem,the estimation method of GPS receiver antenna GRAPHIC residual variations(GRVs)is proposed.In this method,an azimuth and elevation dependent high-resolution grid correction model is constructed,and the influence of some systematic errors on SF orbit determination is eliminated by using the characteristics of GRAPHIC residuals.The data from GRACE-A,GRACE-B,CHAMP and HY-2A satellites is used to assess the potential of receiver antenna GRVs in SF orbit determination.The results show that,after correcting the receiver antenna GRVs,the accuracies of SF orbit determination validated by SLR data are improved by 11.1%,4.6%,14.6% and 4.7%,respectively.(3)The method of pass-by-pass single-difference phase ambiguity resolution by using GPS integer phase clock and bias products is studied.In this method,the integer phase clock and bias products are used to eliminate the effects of GPS satellites decimal deviations on single-receiver phase ambiguity resolution.While the effects of GPS receiver decimal deviations are eliminated by forming pass-by-pass single-difference ambiguity.Thus,the constraints of single-difference phase ambiguity resolution are used in singlesatellite POD.Furthermore,the performances of different integer phase clock products provided by different agencies on single-receiver phase ambiguity resolution and POD are compared and analyzed.The results of the data from GRACE-FO mission show that,after phase ambiguity resolution with the products provided by CNES,Wuhan University and CODE,respectively,the POD accuracies validated by SLR data are all improved by about 30%,and the accuracies of dual-satellite relative orbit determination validated by K/Ka band ranging(KBR)data are improved by 63%,63% and 72%,respectively.(4)In order to solve the problem of high dynamic between satellites and difficult to form stable baseline in multi-satellite formation system,the united estimation method for multi-satellite parameters with double-difference integer ambiguity constraint is proposed.In this method,the observation intensity of relative orbit is enhanced by adding double-difference integer ambiguity constraints after single-satellite absolute orbit determination,and the orbit dynamic parameters of multi-satellite formation are united estimated iteratively.The results show that,in the multi-satellite formation system composed of GRACE-C/-D and SWARM-A/-C satellites,the accuracy of the relative orbit determination for GRACE-C and SWARM-A satellites is improved by 27.6 % after using the double-difference phase ambiguity integer constraints between GRACE-C and SWARMA satellites,and the absolute orbit accuracies for all satellites are all improved to different degrees.