About anisotropic friction coordinate system

Post Reply
argez
Posts: 6
Joined: Wed Jul 27, 2011 5:18 pm

About anisotropic friction coordinate system

Post by argez » Wed Jul 27, 2011 5:25 pm

Hello,

I have been experimenting with bullet for a while now and i really love it. So congrats to Erwin and other contributors for this great library.

I would like to know if the btCollisionObject's anisotropic friction setting has to be applied in local object space or transformed world space? Is it just the x,y,z friction coefficient for the original coordinates of the rigid body?

Thanks :D

Flix
Posts: 456
Joined: Tue Dec 25, 2007 1:06 pm

Re: About anisotropic friction coordinate system

Post by Flix » Thu Jul 28, 2011 11:08 am

I'm intereseted in this topic too (I still haven't used anisotropic friction yet, but I found very little about it in the Bullet documentation and forum posts). I would like to know:

1) Is there a Bullet demo that shows its usage ?
2) Does btCollisionObject::setAnisotropicFriction(...) work out of the box, or needs something like:

Code: Select all

	m_dynamicsWorld->getSolverInfo().m_solverMode |= SOLVER_USE_2_FRICTION_DIRECTIONS;
or other settings ?
3) [Answer to argez] I suppose the coords are in the local frame (i.e. (0,0,1) = Z direction of the rigid body), but a more reliable opinion is welcome. Is this vector a coefficient that is applied to the default friction value, or it's a valid friction vector itself ?
4) Is it possible to override this friction vector in the custom material callback ? How ? (I would like to model a sledge with bidirectional friction in its lower part and normal friction in its upper part. Is it possible?)

Thanks for any feedback about it in advance.

argez
Posts: 6
Joined: Wed Jul 27, 2011 5:18 pm

Re: About anisotropic friction coordinate system

Post by argez » Thu Jul 28, 2011 1:28 pm

Good points Flix and thank you for the reply. Not to really answer your questions as i have the same ones but

1. I havent't found any demo or place explicitly using the setAnisotropicFriciton() or the m_anisotropicFriction directly.

2. I do suppose that it works out of the box as i can set setAnisotropicFriction(btVector3(0,0,0)) and then the setFriction(1.0f) of the collision object is null'ed out. So looking at the code, they seemed to multiply each other i think.

3. Local frame 0,0,1 would be best. I will do more test tonight and see if i can visualize the effects

4. Good question. I would like to know also if we can customize it in the custom callback.

argez
Posts: 6
Joined: Wed Jul 27, 2011 5:18 pm

Re: About anisotropic friction coordinate system

Post by argez » Sat Jul 30, 2011 5:13 pm

Well, anisotropic friction works great. It is in object coordinates and works in conjunction with the setFriction().

User avatar
Erwin Coumans
Site Admin
Posts: 4228
Joined: Sun Jun 26, 2005 6:43 pm
Location: California, USA
Contact:

Re: About anisotropic friction coordinate system

Post by Erwin Coumans » Tue Aug 02, 2011 2:49 am

Anisotropic friction is in local coordates indeed.

There is no demo indeed, so if you have some spare time, please consider creating some demo.

Anisotropic friction is used for the Bullet integration into the Blender 3d modeler.
Thanks,
Erwin

Post Reply