imuPreintegration.cpp issue

[LHZ-code]

Excuse me, why do we need to transform the pose of the lidar odometry into the IMU coordinate system in the odometryHandler at this node, and then transform it back to the lidar coordinate system in the IMUHandler? What is the purpose of doing so?

[JokerJohn]

@LHZ-code This is because the built-in IMU factor in GTSAM requires integration operations in the IMU body frame. However, in the backend optimization of mapOptimization, it is carried out in the LIDAR frame. The IMU pre-integration node needs to use the bias from the optimized previous frame to predict the IMU rotation part before the current LIDAR frame. Therefore, a conversion back to the LIDAR frame is necessary.