Research On LEO Real-time Precise Orbit Determination Based On Onboard GNSS

The use of onboard GPS receivers in low Earth orbit(LEO)spacecraft systems has been quite common to provide real-time ephemeris for onboard processing such as accurate time synchronization,autonomous navigation,attitude determination and control,and accurate relative ranging between spacecrafts.However,the traditional autonomous orbit determination method is to adopt standard single point positioning method for real time orbit determination,by the broadcast ephemeris error and the effects of pseudorange measurement noise,the positioning accuracy of this method is 10 m and 0.1 m/s,which can only meet for satellite attitude and orbit control.On the other hand,remote sensing and scientific tasks and applications such as: altimetry gravity measurement,SAR interferometry or atmospheric detection has higher requirements on accuracy,these fields call for the location of satellite reach to decimeter precision and the speed reach to the submillimeter accuracy per second.Real-time onboard navigation using GNSS measurements can thus contribute to an increased autonomy of future space missions,reduced ground operations cost and improved science return.At present,the Kalman real-time track algorithm is widely used for onboard real-time orbit determination field.In this paper,a Kalman dynamic filtering method based on pseudo-distance observation is used to achieve higher accuracy of onboard real-time orbit determination,a data processor program PROD was realized as well.The program mainly realizes the process of data preprocessing,filtering initialization process,calculation of state transfer matrix,time updating and measurement updating.Then,GPS and Beidou navigation data received by the GNOS receiver and GRACE-GPS data are processed by dynamic Kalman filter(PROD)and standard single point positioning(SPP).The results show that with the use of PROD,FY-3C/GNOS-GPS data real-time orbit determination precision is obviously better than SPP's.The filter initialization time is 2 hours,SPP position accuracy(3D_RMS)is 4.5 m,PROD(3D_RMS)can improve the precision to 1.4 m,velocity precision(3D_RMS)is 1.8 mm/s(higher than traditional SPP nominal accuracy of 0.1 m/s speed two orders of magnitude).Another advantage of the algorithm is that even without observation satellite,still can extrapolate from an observation epoch to the next observation epoch,which the data won't lacking during the whole process of orbit determination,maintaining the integrity and continuity of the process.In addition,the extrapolation precision is very high,the precision can reach 3 m(still better than SPP nominal precision 10m).Due to FY-3C/GNOS only has 4 channel,FY-3C/GNOS-BDS data processed show that SPP mode position precision(3D_RMS)is 45 m,and PROD filter initialization time is 4 hours.,all segment of BDS position precision can be up to 22 m(3D_RMS),velocity accuracy for 23 mm/s(speed precision is much higher than traditional SPP nominal accuracy of 0.1 m/s).Without observation data,PROD can still orbit determination,but when satellite number less than four,PROD mainly depend on extrapolation,and the positioning accuracy can reach 30 m,velocity precision(3D_RMS)can reach 23 mm/s(which is much higher than traditional SPP nominal accuracy of 0.1 m/s),and when four satellites constantly positioning accuracy can be increased to 15 m,velocity precision is 10 mm/s(higher than traditional SPP speed nominal accuracy of 0.1 m/s an order of magnitude).Similarly,GRACE-GPS navigation data orbit determination result is very good also,the filter initialization time is 2 hours,SPP mode position accuracy(3D_RMS)is 4.5 m,PROD location precision(3D_RMS)can be up to 2 m,velocity precision velocity is 3 mm/s,extrapolation accuracy is 7 m.In conclusion,a Kalman dynamic filtering method based on pseudorange can effectively smooth positioning results,greatly reduce the positioning error and calculation.This method can also meet the real-time requirements,will be applied to future onboard GNSS orbit determination.