Satellites maneuvers

[gwbres]

Hello @ChristopherRabotin,

I thought you had means to describe satellite maneuvers somewhere in your ecosystem, if I recall correctly. But i can't seem to find it.
Is that correct ?

I'm asking because I am upgrading my "SP3" library, and in very rare cases, SP3 may describe satellite maneuvers and I'd like to see if we can propose a bridge method:

[ChristopherRabotin]

Hi Guillaume,

Yes, Nyx supports modeling maneuvers: https://docs.rs/nyx-space/latest/nyx_space/dynamics/guidance/struct.Maneuver.html . These must be specified in a supported frame with the duration and the pointing angles, or with the exact change velocity used by the operator. Are these published by the SP3 providers? If not, then Nyx can try to solve for the maneuver that happened using the differential corrector.

[gwbres]

Hum, so it will be a little more difficult than I thought.

All we have is a flag indicating the thruster was activated at some point between the previous epoch and current epoch. We will always be limited by the epoch interval of the text file.

For each epoch, we have (x_ecef_km, y_ecef_km, z_ecef_km) and (vel_x_km_s, vel_y_km_s, vel_z_km_s) is either directly given or deduced.

So the first question is how to obtain the Local Frame.

From SP3 I have Orbits in ECEF (always), and unlike RINEX we do not have information about the satellites antenna to serve as an NEU reference point.

Then we need to deduce the thrust characteristics from the past states, which will have larger errors with large sampling intervals

[ChristopherRabotin]

How does other software handle maneuvers? Would it make sense to report the trajectory as invalid during the times when there was a maneuver?

[gwbres]

How does other software handle maneuvers?
Would it make sense to report the trajectory as invalid during the times when there was a maneuver?

It would make sense to discard vehicles being maneuvered, as similarly to vehicles being eclipsed.

RTKLib supports true ppp, so

• it supports SP3 input
• I seriously doubt they go this far but I can't say for sure. Only @jmfriedt may be able to answer that question
• They have their own implementation of the eclipse filter, it is said to work only for "block IIA" (I presume GPS? I'm not advanced enough) vehicles. While we have a more powerful calculus which applies to any spacecraft

[gwbres]

So can you help me design the main principle ? the idea is to be able to retro analyze the maneuver from the text file.
Obviously this will be limited, mostly by the sampling rate ( I only get a snapshot of the trajectory from time to time)

I presume I should construct a propagator from all previous "normal" waypoints in the trajectory.
How many waypoints are needed to build a reliable propagator ?
Assuming I have this propagator, I can propagate the previous trajectory, and we can compare to the actual trajectory during the maneuver.
Finally, not being an astro guy, how do you guys characterize a maneuver ?

[gwbres]

Hello, thank you very much for your help. I will keep you updated, I'm not asking you to code the features don't worries ;)
But not being very knowledgeable in astrodynamics, I always appreciate your help that you never fail to provide.

Yeah I read the launch went well, wish you all the best for the misssion 👍
I'm particularly interested in the progress that will be made towards a lunar constellation and the lunar time. Apparently the mission integrates a first (ever) module to link GPS to the module on the moon (italian module me thinks?)

Just so you know, the next version of SP3 is coded and already validated in the RINEX/develop branch which will be merged very soon

[ChristopherRabotin]

That's right! We're flying LuGRE, a collaboration between the Italian Space Agency and NASA Goddard. I'm curious to see the results, but I still think that there will be work for flight dynamics and orbit determination engineers for many years to come because of the requirements of such a payload. But it's very exciting nevertheless!

Is the SP3 validation new? I thought you had it working for many months, but I may be wrong.

[gwbres]

Is the SP3 validation new? I thought you had it working for many months, but I may be wrong.

No it's simply an upgrade, that might be tagged v2.0 - we'll see. That integrates these new tiny features (basically everything but a maneuver characterization method). The interpolation ecosystem is also upgraded. And we propose an ANISE option to create a bridge around the Orbit object

[gwbres]

A quick first notes

Only the state before the maneuver is needed to solve the maneuver, if you know the state after the maneuver

We will always know both, unless in extreme rare case where the file starts or terminates on a maneuver, which we can imagine to forget about

(best is to take one that's 30-45 minutes after the maneuver to you know you're out of the maneuver).

In typical/standard SP3 that would be 2 to 3 data points after the end of maneuver, considering 15' sampling period and considering the maneuver halts soon enough before the end of the text file

[ChristopherRabotin]

Since you know both states, you can set up the targeter to find the maneuver that connects the state before the maneuver to the state after the maneuver.