5 ms·
I'm a big fan of orbiting problems. There are some cool references on your gitlab - I'll need to dig through them sometime. Now some of my "tooting": I did a
by musgravepeter 10y ago
I'm a big fan of orbiting problems. There are some cool references on your gitlab - I'll need to dig through them sometime.
Now some of my "tooting":
I did a quick JS Kepler evolver (http://nbodyphysics.com/blog/2016/05/29/planetary-orbits-in-javascript/ http://nbodyphysics.com/blog/2016/05/29/planetary-orbits-in-...) and some very funky regularizing integration for the three body problem (http://nbodyphysics.com/blog/threebodyv2/ http://nbodyphysics.com/blog/threebodyv2/).
Lots more side projects queued up - latest is a Unity Asset, Gravity Engine.
Under "hugely ambitious" is trying to grok http://www.cds.caltech.edu/~marsden/volume/missiondesign/KoLoMaRo_DMissionBook_2011-04-25.pdf http://www.cds.caltech.edu/~marsden/volume/missiondesign/KoL... and do real three-body trajectories.
- exDM69 10y agoGood stuff and interesting projects. I spent a while going through your blog posts and other stuff. Do you mind if I drop you an email to exchange some ideas? Your Kepler's equation solver looks like a pretty standard textbook Newton iteration. I'm sure you're aware of its shortcomings, look at libtwobody for a more advanced solver (Laguerre-Conway) and the bibliography for the papers describing it. There's also a universal variable solver which doesn't have issues for near-parabolic orbits and handles elliptic/parabolic/hyperbolic orbits with no special cases. I've also got an interest in 3-body and n-body stuff but so far I've concentrated on the two body mechanics and dabbled a bit with simulating rocket launches. Your Unity Asset looks pretty interesting. Has it been popular? I've also considered making (code) assets for Unity, so please do share any experiences you've got.
- kartikkumar 10y agoI've developed a lot of open-source space simulation code and got a few projects in the pipeline. Maybe worth having a chat? My email is in my profile. Anyone interested, feel free to ping me.
- musgravepeter 10y agoThe Kepler equation solver is the standard Newton iteration, which is ok until e gets larger. The actual N-body simulator in gravity engine allows for better orbits and the integrator is configurable (Hermite, Leapfrog) and easy to extend. Colwells book has LOTS on this subject too (https://www.willbell.com/MATH/mc12.htm https://www.willbell.com/MATH/mc12.htm) There are about 20 GE users so far, so it seems somewhat "niche". Offering an asset forces me to think a lot more about usability and documentation. I have recently added support for any solar system object and real world units - with tunable scaling. Not things I would have done if I was not making it an Asset. It's gratifying to get some uptake and the people who use it seem to like it. If you like Three Body stuff, checkout my iOS/Android three body app. LOTS of solutions to explore. (https://itunes.apple.com/us/app/threebody-lite/id951920756?mt=8/ https://itunes.apple.com/us/app/threebody-lite/id951920756?m... or https://play.google.com/store/apps/details?id=com.nbodyphysics.threebodylite&hl=en https://play.google.com/store/apps/details?id=com.nbodyphysi...)
- exDM69 10y agoThanks for all the good links. Get in touch if you want to chat about projects.