I attempted to use LIO to test the building1 dataset. In this dataset, LIO showed a high degree of overlap, but the trajectory deviation was significant and followed a certain pattern. I discovered through testing that this was due to the RTK trajectory and IMU extrinsics. This issue became more apparent when the vehicle was rotating. I tried correcting my LIO trajectory using visually estimated RTK and IMU extrinsics, which significantly reduced the deviation.
