Process Noise “Q” covarience matrix in a kalman filter
12 views (last 30 days)
I am trying to implement a Kalman filter on a Phasor Measurement Unit (PMU) values. I meaured the signal from PMU and give those meaurement as input to Kalman filter to get best estimate. I do not have a Process model. I assume A, B, C and D matrices.
My question is while calculating Q covarience matrix (process noise) in MATLAB, should i give the whole measurement as input to "cov" function in MATLAB or instead of whole measurement i should give the error(actual- measurement) to "cov" function to calculate Q?
Please guide me? Thanks in advance.