How to set Q and R in a Kalman filter?
169 views (last 30 days)
Hi, I have a system where I estimated the parameters using RLS procedure. Now I need to implement a Kalman Filter in continuos time but I don't know how to choose Q and R. Do you have some suggestion? Thanks.
Wilson A N on 11 May 2017
As I understand the parameters Q and R are referring to the process noise covariance and sensor noise covariance matrices.
I believe this should be provided by the plant manufacturer. However, for simulation purposes, you can assume any value greater than 0 depending on how much you think the process and sensor noise affects the plant. You can refer to the following link which shows an example:
You can also refer to the example given below where the Q and R are calculated from zero-mean, independent random variables with known variances: