AnsweredAssumed Answered

A great drift of YAW using Sensor fusion library with FRDM-K66F

Question asked by AMAR BENRAIS on Feb 18, 2019
Latest reply on Aug 19, 2019 by Smith Clarkson
I would like to integrate the NXP Sensor Fusion Library 7.00 to my applicaiton, whre I'm working on an algorithm of IMU/GPS-based positioning of an outdoor mobile robot. 
To do so, I use the FRDM-K64F-AGM01 board with the software NXP Sensor Fusion Toolbox, to evaluate the performance of the MEMS and different kinds of available filters.  
When using the 6DOF Kalman filter (accelerometer+gyroscope), I obtained satisfaying results (with a drift of less then 6°/min). Note that I'm only using a  6DOF Kalman filter, instead of a 9DOF, because I have huge magnetic perturbations due to the electric motors of my robot. 
I also used the NXP Sensor Fusion Library with the FRDM-K66F board, as it has the same MEMS as the FRDM-K64F-AGM01, with an additional 2 CAN output ports,  in order to use the software NXP Sensor Fusion Toolbox  with it. I made the necessary changes using the following functions : ApplyhalAccel () and ApplyhalGyro ().

Unfortunately, I got very bad results. Indeed, the yaw drift is between 15° and 20°/min now !
I was wondering if you can help me solve this issue, by answering the following questions : 
1) Could it be because of a misalignement of the MEMS on the FRDM-K66F ? 
2) Does the Sensor Fusion Library take into account the distance and the position of the MEMS on the board?
3) Do you think I can have the desired performance just by tuning the Kalman filter parameters?
4) Has anyone ever used this library with the FRDM-K66F board? 

Thank you in advance for your kind help.