From: Adam Paul Coates <acoates@st...>  20021228 22:00:57

Okay  I have the paper in a box somewhere  I'll double check that. I'm using GJK, which doesn't mind too much when the objects penetrate, and the contact plane computation doesn't worry about penetration either, so whatever it is, penetration shouldn't be a problem, so long as it's small. I've implemented an impulse based system before and found the same slowdown (Yuck!) I'm looking more along the lines of Baraff's constraint schemes  having glanced at Tokamak, it looks like you're handling resting contact very well  is this based on Baraff's QP formulation? Right now, I'm solving things iteratively, but it's not the most robust scheme; just curious :) Thanks for the notes, I'll check it out! Adam C. On Sat, 28 Dec 2002, David Lam  Tokamak wrote: > I haven't look at Stephane Redon's work so I cant comment on that.<br> But I have worked on the Impulse scheme by Brian Mirtich. If you haven't<br> seen it already here's the link<br> http://www.cs.berkeley.edu/~jfc/mirtich/impulse.html<br>; It describes a scheme for estimating the Time of Impact (TOI) by using <br> the following information:<br> objects' dynamic information (linear and angular velocity)<br> shortest distance between the 2 objects (using CD algoritm called LinCanny)<br> bounding radius of the 2 objects<br> In practice, when use in games, this scheme suffer a few problems.<br> In particular, when 2 objects are in near static contact, you get a series of<br> very small timesteps, causing your simulator to slow down alot. And precision<br> problem could also cause your objects to have penetrated when at the TOI.<br> So I would recommend that when using Impulse, have a minimum timestep<br> plus a backstepping scheme as well.<br> David Lam<br> http://www.tokamakphysics.com<br>; 