Estimator error with Hybrid Arc Coupled Translational and Rotational Dynamics

[z-afzal]

Hi all,

I am currently working on Covariance analysis for coupled translational and rotational dynamics for Jupiter in a hybrid arc configuration. The spacecraft dynamics are modeled using a multi-body, multi-arc configuration. The propagation proceeds without any issues. However, when I attempt to move on to the estimation phase using the numerical_simulation.Estimator , I encounter the following error:

RuntimeError: Error when making multi-arc propagator settings from single arc. Input not recognized.

Given this, I have a few questions:

• Is it possible to perform hybrid arc multi-type propagation (translational and rotational dynamics for Jupiter) with multi-body, multi-arc propagation for the spacecraft, and subsequently transition to the estimation phase without encountering issues?
• Are there specific configuration steps I might be overlooking in terms of setting up the propagation or estimation for multi-body, multi-arc setups?

Code repo Link

Thanks

[DominicDirkx]

Hi,

This error occurs since, at the moment, the hybrid-arc estimation only supports translational dynamics. It's been a while since we wrote this, but it does not look like any kind of fundamental limitation. The (a?) missing link is the implementation of the book-keeping of the coupling terms in the variational equations between the single- and multi-arc dynamics.

We can implement this function, it shouldn't take long, but we don't have time right now to do the testing properly. If you have time to set up a test for this, we can implement the code that should make it possible. Let us know if you'd have the time for this!

[z-afzal]

Thanks for the quick reply and clarification. I do have the time to set up a test for this, and I'd be happy to help with the testing process once the implementation is in place. However, I'm not entirely familiar with the specific steps required for testing, so feel free to share the necessary steps when ready.

[DominicDirkx]

What we can do, is add the missing missing functions to allow multi/hybrid-arc, multi-type estimation. From the looks of it, this is just a few book-keeping steps that are missing, but since the whole procedure for the hybrid-arc specifically is rather complex, I'm not 100% confident making just these small changes will be sufficient.

The main complication in setting up the hybric-arc multi-type estimation is to make the solution to the variational equations come out correctly (https://www.sciencedirect.com/science/article/pii/S0032063322001179 for details). Luckily, testing if this is done correctly is fairly straightforward, you can do this by numerical differencing. Specifically: perturb a parameter (like an initial rotational state element) up, then perturb it down, and compute the derivative from this numerically. Comparing this against the relevant column of the solution of the variational equations will show whether the implementation is correct or not. This is the step we would ask you to do. Would that work for you?

[z-afzal]

Yes, I'll do that and compare the results to verify the implementation.

[DominicDirkx]

I'll let you know when we make the changes, and which branch we do it to. Hopefully we can get to it in the coming week. You'll have to build the new code locally to test the changes.

[z-afzal]

Sure, I will do that. I look forward to hearing from you once the changes are made.