Issue with create_observation_simulators: Ephemeris Error in LEO Satellite Range Observations

[luigigisolfi]

Question

User:*

Hello all,
I'm using the estimation functionalities in Tudat to simulate range observations from constellations of LEO satellites, and have some question wrt to the observation simulator objects.
After defining my SystemOfBodies and creating a link definition, when I use create_observation_simulators with the observation settings to create my observation simulators, I get the runtime error "Error when requesting state from ephemeris of body , body has no ephemeris" when running the simulators with simulate_observations. This, however, doesn't happen when I define an Estimator and use the observation simulators from that object instead, following the alternative method of the observation model setup in the docs.
Since I'm not doing estimation and I'm only interested in the observation simulation, I'd like to use the first method, but am I misunderstanding the setup? The propagator settings are never introduced to the observation simulators if I follow the guide in the docs, which is probably what I'm missing in my setup. If someone could shed a light on the correct use of create_observation_simulators I'd be very happy.

[luigigisolfi]

Answer

Tudat-team:
Hi Oyvind, the reasons for this are a little hidden, but come down to the following. When creating an Estimator , the propagator settings are automatically modified as follows:
propagator_settings.processing_settings.set_integrated_result = True
See this page: https://docs.tudat.space/en/latest/_src_user_guide/state_propagation/propagation_setup/printing_processing_results.html#updating-eph[…]onment-models for details. In short, setting this to True resets the ephemeris of the propagated body with the numerical propagation results after the propagation is finished. Use the above line of code for your propagator settings should do the trick, also when not using the Estimator class. If you run into issues, please post a link to your code (on a private repo, if you prefer), and we can have a look

[luigigisolfi]

User:
Thanks so much, this works perfectly.