A Novel Real-time Multipath Mitigation Algorithm for BeiDou GEO Satellites based on the Spacecraft Reflection Model

The geostationary orbits satellites work at a high circular orbit 35,786 kilometers above the Earth’s equator, and are nearly stationary relative to receivers on the Earth. Therefore, the geostationary orbits satellites have been applied to broadcast the correction messages for increasing the positioning accuracy at the receiver ends as well as to be used as ranging sources in satellites navigation systems. But, there exists a large biases in geostationary earth orbit satellites code range, which are often assigned to multipath effects. Compare with other type of satellites, such as inclined geosynchronous orbit satellites and medium earth orbit satellites, the amplitudes of code error from the geostationary orbits satellites are also outstanding. Currently, some scholars are inclined to attribute the code error to the satellite. But, few papers made a clear conclusions for the code anomaly of BeiDou geostationary orbits satellites. In this paper, we analyze some BeiDou GEO satellites measurements collected by KARR and JFNG stations, and the results show that the multipath errors vary over time, even show a daily periodicity and an elevation dependent behavior during a sidereal day. In addition, we proposed a new signal reflection model on satellites surface, and design a numerical optimization method to obtain the reflection parameters. Moreover, the continuous wavelet transform is also adopted by the optimization algorithm. The simulation results indicate that the modelled multipath curves have a high similarity with the actual multipath time series after calculating the correlation coefficient of these two curves. Furthermore, we can estimate the variation patterns of multipath error of the near future days by using the reflection model and the effective ephemeris. Lastly, based on the inverse continuous wavelet transform, a real-time multipath error mitigation method is also proposed, and the results show improvements in root mean square of the raw multipath time series.