[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Orekit Developers] Writing an OEM file after an ephemerispropagation



Hi all,

 

I’m back trying to figure out how to compute the covariance and, even with your two useful links, I can’t yet understand how to do it…

 

According to [2], the set of additional equations is automatically added to the numerical propagator in order to compute partial derivatives of the orbit along with the orbit itself. And it also says this is useful for Orbit Determination, which would be nice since I’m also trying to do that.

 

The problem is that, if it is indeed automatically added to the numerical propagators, I have no idea how to define an initial covariance and how to get the computed one.

 

For example, in the case of orbit determination, I’m following the tutorial OrbitDetermination.java (modified for my data but it doesn’t change the algorithm). In that example, the estimated orbit is given by the line « Orbit estimated = estimator.estimate().getInitialState().getOrbit(); » where :

  • Estimator.estimate() is a numerical propagator
  • estimator.estimate().getInitialState() is the last estimated Spacecraft state

I thought that the partial derivatives could be obtained from one of the two previous elements or event from the estimated orbit but it doesn’t seem to be the case.

 

Do you know how to compute that covariance and how to get it afterwards ?

Is there really a need for an initial covariance in the case of OD since the orbit doesn’t really use the initial state other than for initializing the least square problem ?

 

Thanks in advance for your help !

 

Have a good day, everyone.

Best regards,

Nicolas

 

De : Ward, Evan
Envoyé le :mercredi 8 mars 2017 15:08
À : orekit-developers@orekit.org
Objet :Re: [Orekit Developers] Writing an OEM file after an ephemerispropagation

 

Hi Nicolas,On Wed, 2017-03-08 at 11:28 +0100, contact@nicolasrol.fr wrote:

> First of all, great job on OREKIT, it’s a really nice tool!

 

Welcome to the mailing list. We appreciate the feedback.

 

> Is this covariance matrix even available after a

> propagation ?

 

I think Hank answers your other question so I'll just try to answer

this one. The State Transition Matrix (STM) can be computed during

propagation. Take a look at [1] and [2]. Then you can use the STM and

your a priori covariance to compute the covariance at latter times.

 

Best Regards,

Evan

 

[1] https://www.orekit.org/forge/projects/orekit/wiki/Propagation#Propa

gating-both-equations-of-motions-and-additional-equations

[2] https://www.orekit.org/site-orekit-8.0/apidocs/org/orekit/propagati

on/numerical/PartialDerivativesEquations.html