Fig. 1. Overall framework of SLAM graph optimization algorithm with vector map constraints

Fig. 2. Relationship between local maps

Fig. 3. Local factor graph after adding vector factors

Fig. 4. Global factor graph after adding vector factors

Fig. 5. Fitting situation of point cloud and vector map (campus 1). (a) Without adding vector constraints; (b) adding vector constraints

Fig. 6. Comparison of trajectories (campus 1)

Fig. 7. Fitting situation of point cloud and vector map (campus 2). (a) Without adding vector constraints; (b) adding vector constraints

Fig. 8. Comparison of trajectories (campus 2)

Fig. 9. Fitting situation of point cloud and vector map (campus 3). (a) Without adding vector constraints; (b) adding vector constraints

Fig. 10. Comparison of trajectories (campus 3)

Fig. 11. Fitting situation of point cloud and vector map (campus 4). (a) Without adding vector constraints; (b) adding vector constraints

Fig. 12. Comparison of trajectories (campus 4)

Fig. 13. Visualization of position error comparison. (a) Comparison of RPE box plots for single - building survey areas; (b) comparison of APE violin plots for single-building survey areas; (c) comparison of RPE box plots for multi-building survey areas; (d) comparison of APE violin plots for multi‒building survey areas