Covariance matrix frame

Hello to everyone!

I’m quite noobie in orekit usage, so I struggle some conceptual misunderstanding. Hope someone to be kind to help me with this.

Here the point. I’m trying to perform OD with pair of ground stations offsets estimation. Optimization process converges well with BatchLSEstimator, so I retrieve physical covariance matrix from it. So here l’v got some trouble. As I understand, orbital parameters are estimated in it’s SpacecraftState initial frame, and stations offsets are given in it’s local topocentric frames (individual for each station).
So my question is: Are my suggestions correct? And if so, how it connects with covariance matrix parameters frames? Am I getting diagonal blocks of parameters in different frames? And how I should interpret off-diagonal blocks in this situation? And is it possible to transform full covariance matrix to, say, some ECEF frame?

Thanks a lot for your attention!

Hi @an-gorb

Welcome to the Orekit forum.
Sorry for the late answer.

Yes, they are.

Elements in the covariance matrix depend on the frame used to estimate the different parameters. Could I recommend you the paper: Covariance Transformations for Satellite Flight Dynamics Operations It is very useful to understand covariance transformations. It is available here.

Best regards,
Bryan

Thanks a lot @bcazabonne! I suppose now I’m get it.