GPS/BDS Combined Precise Point Positioning with Geostationary Satellites Offset

The combined utilization of GPS/BDS has become prevalent to improve the performance of GNSS navigation and positioning. However, the near-static viewing geometry, degraded orbit determination and satellite code bias cause the BDS geostationary (GEO) satellites pseudorange contain more errors and cannot make effectually contribution like other satellites for GPS/BDS combined precise point positioning (PPP), even putting a brake on the convergence. For BDS GEO, the final precise orbit is decimeter level and the satellite-include code bias cannot be modeled effectively. The common method for GEO satellite is to down its weight in GPS/BDS PPP, which causes observations are underutilized. In this paper, we purpose a new method to estimate the unmodelled errors of GEO satellites with all GPS/BDS pseudorange observations. Since all these errors are satellite- dependent, we set separate GEO-offset parameter for each GEO satellite. Coordinates, receiver clock and GEO-offset parameters are estimated together. Many experiments are conducted to present the benefit of the GEO-offset method. At first, we estimate GEO-offset with receiver coordinate unknown or fixed and similar GEO-offset are obtained by two strategies. After corrected by GEO-offsets, most errors can be offset and the variation of corrected multipath observable becomes stable significantly. At last, we apply the GEO-offset into GPS/BDS precise point positioning avoiding a down weight of BDS GEO satellite. The convergence time to the error level of 0.5 m, 0.3 m and 0.15 m are shorted from the average of 16.2 minutes, 23.7 minutes and 51.6 minutes to 5.8 minutes, 11.3 minutes and 20.1 minutes, respectively.