I am currently working on APM 2.0 to implement Kalman filter using GPS and INS data using 9 by 9 state matrix,  [P_x,V_x,Pl_x,P_y,V_y,PI_y,P_z,V_z,PI_z]', where PI_x,PI_y, and PI_z represent platform tilt about x, y, and z axis.

A problem is that newly updated global matrices X and P are not transferred to next iteration although I defined those as global matrices.

As far as I tested, a single global variable which is defined before void setup() works well as a global variable in that the previous value of the variable is successfully transferred into next iteration.

Is there anyone who could shed some lights on me?

Views: 1116

Reply to This

Replies to This Discussion

Hi Sunghun,

Did you get solution to your problem ? If so could you tell me what as I am also working on Kalman filter using this board. Moreover, I wanted to know where in the code is kalman filter implemented in ArduPlane (or any Ardu code)..

Seeking your reply.

Hello Sunghun. 

Nice to see your interest on implementing Kalman filter on APM 2.0. Have you solved the problem of updating global variable?  I too have implemented 9 state Kalman filter for navigation and i did not faced any problem in updating global variables. I am getting good prediction of attitudes but if i enable measurement update part of filter the estimation give NAN. I guess that there is some problem while computing Kalman gain. Would you like to discussion with me to sort out this problem?

Sudesh K Kashyap

Hi Dr. Sudesh,

I think that post was over a year ago. Might be a good idea to create a new forum, as I think an alternative to DCM as Kalman might be interesting to people out there, who can reply to you.



Reply to Discussion


© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service