Given both initial position, end position, IMU data, predict the path of the robot... Kalman filter?

[AudioVideoDisco]

Hello forums,

We have a number of robots for measuring sewer/storm pipes. These robots are fitted out with VectorNAV VN-100 IMU/AHRS sensors which are continuously recording 12 parameters: Yaw, Pitch, Roll, Magnetic field X, Magnetic field Y, Magnetic field Z, Acceleration X, Acceleration Y, Acceleration Z, Gyro X, Gyro Y, Gyro Z. This data is saved to a text file (PRH.dta - attached, VI attached for parsing the data).

I would like to find the path of the robot using the IMU measurements together with the starting and end points (determined from GPS) and plot this as a 3D graph or point cloud, and display this on a map.

I have been looking up how to approach this problem and have come across the Kalman filter however I am not sure how to implement it in this case.

Any help is greatly appreciated!

~AVD

P.S. I had to change the file format for the PRH data from .dta to .txt for uploading - change back before running the PRH read function.