nhz2
commented
4 years ago Overall, the propegation logic and everything looks great. I am wondering, however, if we'd want the
bestestimatefunction to ever return an orbit that wasn't caught up to the present. I'm not sure if this will ever be useful information for the rest of gnc/fsw/ Thoughts?
It probably won't be useful for the rest of GNC if the best_estimate isn't caught up. Hopefully this will only happen on startup, or if the GPS time and teensy clock get off by a huge amount (like more than 1 second) for some reason. GNC and flight software can check by calling est.best_estimate().numgravcallsleft() and est.best_estimate().valid() I'm not sure, but hopefully that will just be inlined and won't actually copy the current Orbit into a temporary object and then call numgravcallsleft().
Ground Propagator
Fixes #181
Summary of changes
Orbit::numgravcallsleft()that it returns 0 if the Orbit is invalid.Ptest Effects
None
Testing
Unit tests are in
test_groundpropagator.cppLet me know if there is some edge case I'm missing that might break the estimator.Under normal conditions, this estimator just propagates the most recently uplinked Orbit.
The main goal of the testing is to ensure there is no input that will break the estimator so it can't recover if ground does eventually uplink a good Orbit, and the sat is given enough time to propagate it to the current time.
Also the testing should ensure that too many Orbits getting uplinked won't overload the estimator and prevent it from making progress.
Documentation Evidence
Inline documentation is sufficient.