Multipath error mitigation method considering NLOS signal for high-precision GNSS data processing

The non-line-of-sight (NLOS) signal has different signal characteristics and error from multipath, and it may lead to significant errors in the Global Navigation Satellite System (GNSS) positioning at the centimeter to millimeter range. However, the current multipath error mitigation methods for carrier phase observations do not distinguish them effectively. The multipoint hemispherical grid model (MHGM) is a method to model the multipath error in the spatial domain. We detect the NLOS carrier signal at a static station by the three-dimensional environment data and evaluate the reduction effects of the MHGM on multipath and NLOS errors. In our experiments, the MHGM reduces both multipath and NLOS errors significantly for the double-differenced residuals of carrier phase observations. But after using MHGM, the accuracy of NLOS observation is still significantly lower than that of multipath. We determine the accuracy difference between multipath and NLOS observations of the station, and propose two improved strategies based on the original MHGM which is also called strategy 0: strategy 1 of eliminating NLOS signal and strategy 2 of reducing its weight in the GNSS data processing. Our kinematic positioning test indicates that with multiple satellite systems, strategies 1 and 2 improve the mean RMSE of positioning results by about 10% and 15–17%, respectively. With fewer satellite systems, strategy 1 of eliminating NLOS signals may lead to poor satellite geometry and cannot improve effectively, while strategy 2 of reducing the weight of NLOS signals is more robust and recommended.