The Filtered GNSS Real-time Precise Orbit Solution

Abstract

The reliable real-time precise satellite orbit is the prerequisite and foundation of real-time precise positioning service for Global Navigation Satellite System. Recently, real-time precise orbit products are usually predicted from the post precise products that are determinate by the batch-processing method based on the history data. This method bears some obvious drawbacks that degrade the performance of real-time orbit products, especially for the Chinese BeiDou Satellite System (BDS). Firstly, the predicted orbit depends heavily on the accuracy of dynamic force model and responses slowly to the model accuracy degradation and variation. Secondly, the predicted precise orbit exhibits uncontinuity between two consecutive orbit arcs, which will affect the position accuracy. In this paper, the Square Root Information Filter (SRIF) method is adopted to update the satellite orbit in real-time and response quickly to orbit maneuver. The strategy is validated for the POD of GNSS satellites using one-month data. The results show that the 3D RMS of SRIF orbit for GPS and GLONASS satellites is 6.7 and 9.3 cm, respectively. The mean value of SLR validation residual for BDS IGSO and MEO satellites is less than 10 cm and for GEO satellite is about 20.7 cm. At last, the real time products were used to for kinematic PPP. The positioning results show that, compared to the predicted real time product, the real time product generated by the proposed real time filtering method improves the BDS-only kinematic positioning accuracy significantly and the accuracy of 3 cm in the horizon and 5 cm in height can be reached for the multi-GNSS positioning.