Closed teubert closed 1 year ago
(For context, I was asked to take a look at this PR by Chris since the changes directly impact the project I'm working on)
I can't say I quite understand the changes (and like before, I can't test them on my end or anything) but I see some familiar code regarding getting the median SoC. Out of curiosity, is there a particular reason we get the median of the sample predictions and not, say, the mean?
Otherwise, I don't see anything of immediate concern. Hope this fixes the problems we've been encountering!
This PR ports over an improvement from ProgPy that updates SOC estimation to include voltage when voltage approaches the VEOD threshold. When this takes effect is configurable (VDropoff
This fixes a phenomenon where RUL is 0, but EOD is not 0.
See https://github.com/nasa/prog_models/pull/45