Loosely vs Tightly Coupled LIO
In loosely coupled LIO, the IMU and GPS are first fed into error state kalman filter which outputs a global pose estimate T_G. T_G is fed into a lidar odometer (such as LOAM) as an initial relative pose estimate. The lidar odometer will iterative optimize the pose estimate to minimize its residual, and outputs a refined pose T_R. T_R is fed back into the ESKF as an SE3 estimate (just like GPS).
In tightly coupled lio, The Error state kalman filter has an embedded lidar odometer with iterative pose optimization. IMU, GPS, and lidar residuals are added to the ESKF as individual observations.
Therefore, loosely coupled LIO keeps the ESKF and lidar odometer relatively separate, while tightly coupled LIO merges the two together.
Tightly Coupled Lio
FastLIO
Highlights:
- It terms “integration model” as forward prop
- Motion undistort is just the intuitive model that aligns all points to the same timestamp.
