Real-Time Physics Simulation Forum

Re: Stewart & Trinkle - additional resources

2012-10-18T06:20:10+00:00

Erwin - thanks a lot! It's a lot of information to learn and infer from .

-
Cloned and built without a squeak in vs2k10.

Statistics: Posted by kalesony — Thu Oct 18, 2012 6:20 am

Re: Stewart & Trinkle - additional resources

2012-10-17T19:50:42+00:00

I create a new github repository for the most up-to-date daVincyCode3D version 1.1, an implementation of Stewart Trinkle:
https://github.com/erwincoumans/daVincyCode3D

Most of the setup is in btST2World.cpp and interfaces to the PATH MLCP solver are in SimpleLCP_Path.cpp
It compiles out-of-the-box using Microsoft Visual Studio 2010. You can also download precompiled Windows executables in the Download section:
https://github.com/erwincoumans/daVincyCode3D/downloads
It also includes several papers to the related work, I can recommend our STAR paper of Eurographics 2012 that describes the Stewart Trinkle method too.

daVincyCode3D relies on the proprietary PATH solver. It would be great to have an open source C/C++ replacement for that.
There is some interesting PhD thesis with an open source Matlap implementation called COMPASS that could be a starting point for an MLCP solver to replace PATH:
http://www.mat.univie.ac.at/~neum/software/compass/

By the way, the latest Bullet 2.81 release also includes speculative contact constraints to avoid tunneling, in a similar style to Stewart Trinkle.

Statistics: Posted by Erwin Coumans — Wed Oct 17, 2012 7:50 pm

Re: Stewart & Trinkle - additional resources

2012-10-16T14:39:40+00:00

Steve, thank you for spending your time looking for this .

Answering your question: I was looking for virtually anything related to Stewart-Trinkle time-stepper to help me fully understand it; obviously, an app (not to mention source code) always comes in handy in situations like this.

I found your lcp_applet advertised in this thread by ngbinh: http://bulletphysics.org/Bullet/phpBB3/ ... ?f=4&t=431. Don't really know what to expect from it, I just felt tempted by the line "We've made a simple demo to show the different between Stewart-Trinkle and Anitescu Potra (ODE using this now) time stepper." <- seemed to be promising something fairly informative, that is all I could tell .

Java is no constraint either - I am simply looking for resources to learn and experiment with Stewart&Trinkle t-s and its derivatives.

Statistics: Posted by kalesony — Tue Oct 16, 2012 2:39 pm

Re: Stewart & Trinkle - additional resources

2012-10-15T23:29:08+00:00

@kalesony

I went through my backups and I can't seem to find that code. Really stinks that RPI killed all alumni pages.

What are you looking for? That part-seating applet wasn't doing Stewart-Trinkle timestepping. Are you looking for java code implementations of the Stewart-Trinkle timestepping?

Statistics: Posted by sberard — Mon Oct 15, 2012 11:29 pm

Re: Stewart & Trinkle - additional resources

2012-10-15T16:09:00+00:00

It's there, thanks again .

Statistics: Posted by kalesony — Mon Oct 15, 2012 4:09 pm

Re: Stewart & Trinkle - additional resources

2012-10-15T16:03:22+00:00

Oops, that's who I meant! Please let me know if it hasn't shown up in your inbox.

Statistics: Posted by bone — Mon Oct 15, 2012 4:03 pm

Re: Stewart & Trinkle - additional resources

2012-10-15T14:05:24+00:00

@sberard: great, I appreciate that.
@bone: I hope your proposition includes me too, will PM you right away, thanks.

Statistics: Posted by kalesony — Mon Oct 15, 2012 2:05 pm

Re: Stewart & Trinkle - additional resources

2012-10-11T19:29:39+00:00

sberard- It turns out I saved a local copy of that rigid_body_lcp.pdf. Either give me a place to upload it or send me a PM with your e-mail addy, it's only about 1.1Mb.

Statistics: Posted by bone — Thu Oct 11, 2012 7:29 pm

Re: Stewart & Trinkle - additional resources

2012-10-11T14:33:43+00:00

I had no idea that rpi took down my old website. I'll try and track down that code from some old backups and post it here if I can find it. I don't have any of ngbinh's stuff though.

--Steve

Statistics: Posted by sberard — Thu Oct 11, 2012 2:33 pm

Stewart & Trinkle - additional resources

2012-10-08T15:27:44+00:00

Hi there.

I need to get a proper grasp over "Stewart & Trinkle" time-stepping scheme (and its descendants) and am thus in great thirst for related resources beyond the commonly available publications.
From 2007 to 2010, ngbinh (user of this forum) wrote several threads providing links to resources I could really use right now. The problem is that as of now those links are dead :/ and I cannot contact (pm) ngbinh anyhow - no response (he may have changed the email or something).

Has any of you stored these resources and could possibly share them with me?

The dead links I'm referring to are:

Thanks!

Statistics: Posted by kalesony — Mon Oct 08, 2012 3:27 pm