Hi~, thank for mentioning our works on GitHub, I am Lin Jiarong, the author of "Loam-livox" (https://github.com/hku-mars/loam_livox). The scenario you mention is exactly true. Since the FOV of livox-lidar is about 38.4° (actually the effective FOV is 34 °, we abandon the point near the edge of the circle). When the lidar rotates, the overlapping area is getting smaller, the case of degeneration might happen (especially for those surrounding feature is less), make the state estimation not so reliable. However, if the feature (edge and planar) in the overlapping area is enough, we can still estimate the state well.