search

Noiredd / PEGAS

Powered Explicit Guidance Ascent System - a KSP & RO autopilot using the Space Shuttle guidance algorithm, UPFG
http://forum.kerbalspaceprogram.com/index.php?/topic/142213-pegas-powered-explicit-guidance-ascent-system-devlog
MIT License
114 stars 31 forks source link

Adding a new ascent mode to PEGAS #35

Closed Aram1d closed 2 years ago

Aram1d commented 3 years ago

Hi,

I had some problems with using PEGAS on my SOYUZ rocket:

The initial ascent mode using a single pitch angle followed after an amount of time to then track the prograde vector was not sucessful on my craft, it swings a lot betweens the different steps, and I could'nt find a pair of values working good enough to get at set of data (initial position and velocity) that UPFG could easily manage when it triggers.

That's why I coded some new routine locking the pitch angle as a linear function of altitude. You can use multiple linear regression, just give their beginning and ending altitudes and the specific pitch values for them.

You can have a look to the extended documentation i wrote for more details, and i keep available for questions.

Thanks, Wilfried

Patrykz94 commented 3 years ago

I only really tested it on one vehicle so far but it seems to work really well (I had some serious trouble tuning the original ascent mode to work with this rocket). I like the idea of having a list of pitchControl settings. Makes the ascent profile very customisable.

Noiredd commented 3 years ago

@Aram1d @Patrykz94 Sorry guys for coming late to the party, my notification settings were off and I did not receive any news of this PR. Anyway, I have always wanted to have something like this, so I would absolutely like to help with integrating your feature. Could I request a little more patience from you? I'm currently working on some 1.3 features that I would like to finish first, while I'm focused on them. Then, when these are done and pushed, we would tackle this one. If you still want to put in some effort, I would signal you when I'm done with my stuff, then you would rebase, we'd make a few more tests and have this in 1.3 - please let me know.

Noiredd commented 2 years ago

Since merging this would have required resolving a ton of conflicts and reverting changes in some places, I decided that it would be easier to manually rewrite Wilfried's code (retaining his commit authorship of course via --author and --date). So I created #44 to do just that.

However, I still want @Aram1d to get his contributor badge, and I'm not sure if GH will award it if I just include his commit in my own PR, so I will retarget this PR to another branch, only so it can be merged as-is. master will, of course, contain my modified version. Hope nobody minds what I've done here, it's the last thing on my v1.3 todo list and I'd really like to have it done by the end of the month.