An improved joint modeling method for multipath mitigation of GPS, BDS-3, and Galileo overlapping frequency signals in typical environments.

The new generation of BDS-3 broadcasted open service signals B1C and B2a, which are compatible and interoperable with GPS and Galileo overlapping frequency signals, are suitable for multi-constellation global navigation satellite system (multi-GNSS) precise point positioning ambiguity resolution (PPP-AR). However, multipath errors caused by an actual complex environment can affect the ability of ambiguity resolution, thereby restricting the positioning performance of multi-GNSS. Due to different orbital repeat periods of GNSS systems, the implementation complexity of a multipath correction method based on time-domain repeatability is relatively high, while that based on spatial-domain repeatability are research hotspot at present, thanks to the advantages of simple algorithms, easy implementation, and real-time correction. Based on the original multipath hemispherical map (MHM) and trend-surface analysis MHM (TMHM) methods, four multipath processing schemes, namely, the independent modeling and correction (I-MHM, I-TMHM), together with the joint modeling and correction (C-MHM, C-TMHM) of different GNSS systems are proposed in this paper. We find that the residuals of GPS, BDS-3, and Galileo overlapping frequency show a strong correlation at the same spatial position after considering the GNSS inter-system biases in static PPP-AR modes, while the multipath joint modeling and correction method can improve the positioning performance more than the independent modeling and correction. This can be attributed to the ability of multi-GNSS to improve the space coverage within grids, making the modeling results more explanatory. Compared to C-MHM, the C-TMHM derived positioning accuracy and convergence time of combined GCE in 3D component can be improved by up to 29.3% and 40.7%, respectively. In addition, through using multi-GNSS data for multipath modeling, the modeling time can be shortened by more than half to obtain a correction effect similar to that of full orbit period modeling, specifically, 3-day data for GC modeling, while 4-day data for GE, CE, and GCE modeling. Finally, the performances of our improved multipath modeling method were verified and evaluated by using the observation data in environment with fewer blind areas. Compared with the uncorrected cases, the positioning accuracies of GC, GE, CE, and GCE in 3D component improve by 51.7, 63.8, 59.7, and 65.7%, after correcting the multipath error by the proposed C-TMHM method, while the convergence time can also be shortened by 55.3, 51.0, 52.2, and 64.2%, respectively. This research has significant applicability for mitigating multipath errors in various scenarios to improve positioning accuracy and reliability. [ABSTRACT FROM AUTHOR]