which type of kalman filter do i have to implement for unknown dynamic model's system?

Hello, i want to implement kalman filter to predict and correct errors of a position encoder outputs... but i don't have the dynamic model of the encoder .... which type of kalman filter do i have to choose to do that ? thanks in advance

Answer 1

[Honglei Chen]

Each Kalman filter is tied to a certain dynamic model. If you don't have knowledge of your dynamic model, one possible approach is to setup multiple filters that correspond to different dynamic models and see which one gives you the best result compared to your measurements.

HTH

Answer 2

[El Mehdi FADEL]

Hello thank you for your answer... could you give some examples of filters that works with unknown dynamic model systems

[Honglei Chen]

Here is a good reading on this topic.

http://www.jhuapl.edu/techdigest/TD/td2204/Genovese.pdf

HTH

[El Mehdi FADEL]

thanks but for this topic we have the matrice of covariance R and Q ...i don't have any information about these matrices and i don't have the dynamic model of my system... i just have the outputs ... so what is the solution in this case ? thanks in advance