Real-Time Physics Simulation Forum

Indeed, that's going to be more difficult. Particularly the edge/edge case (when an edge of one cube touches the edge of another cube, yet no verts have penetrated the faces of the other cube). I believe some people just skip that case altogether.

My previous post should have said "connected" not "converted". Anyway ... Your original post said: "it seems that edges code doesnt work for by cube." I'm asking what isn't working. Does the cube collapse? Blow up? In other words, in what way does it not work? I'm guess...

I can't figure out where you're coming up with 18 edges. If every vert in the cube is converted to every other, it should be 28, no? What's not working for you?

If it looks good and the CPU usage isn't causing any problems, I'd say you are done. That being said, yes this could be solved more efficiently with a lot of work (note that I personally don't even consider a timestep of 0.003 seconds to be particularly small; presumably you'd never even go above th...

No, you are only computing the collision response (an impulse) once, based on the current states of both objects. Once you have the impulse, you apply it positively to one object and negatively to the other, and the order is irrelevant.

You might do a search on this forum for the same subject on CFM and ERP; I think Dirk and I discussed it previously. IIRC, Box2D uses one derivation (including the h in the coefficients), Bullet does the opposite. Or maybe vice versa. But it can be made to work either way, so if you're deriving it i...

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

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.

jarno wrote:The edge versus edge case can be converted to a vertex versus plane problem by collapsing one edge to a point, and extruding the other edge in the same direction to form a quadrilateral.

A-ha! That's the trick I was missing. Thanks for the idea!

Do you mind sharing how you handle the edge-edge case? My version is atrocious, but basically works.

This of course all depends on the spectral radius of the problem and the nature of the LCP, but my general experience is that it sucks for anything but perhaps large contact systems, where few other methods can be used at all. I don't know that it necessarily "sucks", but it's fairly well...

Erin Catto's demonstration uses velocity-based constraints; I usually think of Lagrange multipliers as used with acceleration-based constraints, but I'm not sure of the exact definition and it could be the same thing. I think if you've modeled constraints, you've either used Lagrange multipliers wit...

... or you can basically have the best of all worlds and just use doubles.

If there's no velocity, those impulses are presumably going to be zero. Most game physics engines depend on solving contact impulses to resist penetration by clipping the penetration velocity to zero. I don't know what your application is, but it sort of sounds like you want to estimate the force ba...

You might want to go back to the main forum page and read the red text regarding this sub-forum.