I am using an IMU to measure the yaw rotation angle of a car. I need to translate the rotation measured to COG of the vehicle. I can ignore roll and pitch as I am driving it on a flat surface. So, does yaw rotation is same at origin of IMU and at COG? If not, how to translate the measured rotation to COG of vehicle?
I went through some research papers, I could not quite follow things. Any help is appreciated.
The best way to solve this is to position your IMU onto the Car's COG, so you can associate directly the IMU information to the COG.
In case that's not possible (design choices, space, etc), two things need to be done:
1) Find the transformation from the IMU to the COG. You need to find the translation and rotation from the IMU to the COG so you can compute any information given by the IMU in relation to the COG. If you're using ROS, TF2 solves that for you once you have the URDF.
2) Using the transformation, relate the changes in the IMU with changes in the COG. Every time you read information from the IMU (Force, acceleration, etc.), you have to transform that information in relation to the COG. The angular rate will change accordingly to the transform, for example. It's a challenging problem where there is not one single solution. Most solutions I found use an EKF for it.
Some papers I found:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4208239/
https://ieeexplore.ieee.org/document/6728637
https://journals.sagepub.com/doi/abs/10.1177/0954410014550052
