A double fusion localization method of GNSS&IMU based on graph-based optimization

Abstract

Aiming at the situation that the positioning accuracy of GNSS satellites decreases in the absence of satellites, an IMU integrated system proposed to add virtual constraints for GNSS positioning, and because of the situation that the historical information fading with time growing up when traditional filtering is used, the GNSS&IMU fusion positioning method of traditional filtering is replaced by graph-based optimization. Firstly, the pseudo-distance of the satellite at the previous time is transformed into the virtual constraint of the satellite at the current time by combining IMU information, then the IMU information and the fusion information from the virtual constraint solution are used to optimize the graph. Simulation results show that this method can effectively improve the accuracy degradation caused by insufficient number of satellites.