SparkFun Forums

[cloud9]

ive been reading up on kalman filters too and this project is more or less what i want to accomplish. however im stuck with determining the state equations. perhaps you could be so kind as to tell me what was the state equation matrix you used.

im currently planning my project on the information i read from "Attitude Determination and Bias Estimation Using Kalman Filtering" by Roni Yadlin. Its a very good resource however he doesnt state how he got his "angular momentum" term which is essential to his estimates.

[blackjack_304]

Hello mpthompson,

I hope you are still active on this thread.


Thanks for the insightful info you provided here. I saw the youtube video which you have uploaded for two wheel self balancing robot.

I was going through your code implementation of IMU_3 (the last code uploaded by you). I see you are using tilt.c/tilt.h AS WELL AS kalman.c/kalman.h in your code. As far as I know, tilt.c/tilt.h is the kalman filter implemented in autopilot project and there is no need of kalman.c/kalmn.h. Could you please let me know why did you use two different filter implementations in the same project?

If its possible for you, can you upload the latest version of the kalman filter code &control loop you implemented, and succeeded as shown in youtube video?

I also have doubt on how to use tilt.c/tilt.h as my sensor HW is completely different than yours. I have mounted the accelerometer on top of the balancing pendulum (Z axis is perpendicular with ground i.e. shows Az = -1000mg due to gravity, X axis in direction of forward/backward movement of cart). My Gyro can log the pitch rate as falling/tilting of the pendulum.
Is this mounting configuration correct for the given implementation of kalman filter in tilt.c?


Thanks