This week, I focused mainly on determining the specific requirements of implementing a Kalman filter for our IMU data. I looked exhaustively both at articles explaining the functionality of the filter and at possible existing libraries that I could base an implementation off of.

What I learned is that the Kalman filter works to decrease the variance of our position estimation by continuously estimating new position and measuring for a new prediction. This loop continues over and over until we have an estimate for position that has been able to weed out a large amount of statistical noise.

I was unable to find an existing Kalman filter library that takes our specific IMU’s (Sparkfun MPU 9250) data and determines position, but I was able to find some libraries that implement the filter towards determining attitude and believe that I can use these implementations as a starting point.


0 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *