Feature-Based Tightly-Integrated RTK/INS/LiDAR Fusion Positioning Algorithm in Ambiguity Domain

摘要

To realize robust high-precision positioning in outdoor diverse environments, the prevailing solution is to integrate multiple complementary sensors like GNSS RTK, INS and LiDAR. In the cases where there is no prior LiDAR map, the global accuracies of fusion positioning and mapping depend on the only global positioning technology, RTK, whose cm-level fix result availability, however, plunges greatly in GNSS-Difficult areas. To improve RTK performance, we propose a new RTK/INS/LiDAR fusion positioning algorithm base on LiDAR features in ambiguity domain, in which we tightly integrate the measurements of the repeated-observed LiDAR features with the ones of RTK. Both theoretical analysis and simulation experiment results prove that the addressed method can provide a higher RTK fix rate to maintain the global accuracy and stability of the fusion positioning in GNSS-Difficult areas.