[ODE] Convex Volume Geom

[Rodrigo Hernandez]

Well, I already have convex-sphere and convex-plane collision detection 
implemented and I am studiying "Computing the penetration depth of two 
convex polytopes in 3D 
<http://biogeometry.duke.edu/pubs-pankaj/papers/penetration.ps.gz>"(http://www.cs.duke.edu/~pankaj/papers.html) 
by Agarwal in order to get the penetration depth, whether or not the 
shapes collide is determined by using Seidel's method for
solving linear programming problems.

Bullet might do a great job at solving GJK numerical stability problems, 
but the fact is that GJK, although matematically correct is inherently 
unstable due to floating point limits on
current hardware, using doubles instead of floats might somehow aleviate 
the issues, but they dont go away, and might pop out when you least 
expect it, I rather avoid that.

But, thats me, if it fits your needs and works for your goals, thats 
great :).

Cheers!

Nguyen Binh wrote:

> I don't really feel like Bullet's implementation has serious numerical 
> instabilities.
> Also, what's your other options? Lin-Canny doesn't work so well, 
> SAP-based method need alot of special code (I think Novodex uses this 
> approach).
> I can live happily with Bullet . I don't know why not many ODE's users 
> used it though.
>
> On 12/21/05, *Rodrigo Hernandez* < kwizatz at aeongames.com 
> <mailto:kwizatz at aeongames.com>> wrote:
>
>
>     No, I want to stay clear from GJK (any implementation), due to its
>     numerical instability,
>     perhaps that's an option for the original poster though.
>
>     Nguyen Binh wrote:
>
>     > You can use
>     >
>     http://www.continuousphysics.com/Bullet/phpBB2/viewforum.php?f=11&sid=fe8071fb3e129323b78b000f65954697
>     <http://www.continuousphysics.com/Bullet/phpBB2/viewforum.php?f=11&sid=fe8071fb3e129323b78b000f65954697>
>     > <
>     http://www.continuousphysics.com/Bullet/phpBB2/viewforum.php?f=11&sid=fe8071fb3e129323b78b000f65954697
>     <http://www.continuousphysics.com/Bullet/phpBB2/viewforum.php?f=11&sid=fe8071fb3e129323b78b000f65954697>>
>     >
>     > It's GJK and proved working.
>     >
>     > On 12/21/05, *Rodrigo Hernandez* < kwizatz at aeongames.com
>     <mailto:kwizatz at aeongames.com>
>     > <mailto:kwizatz at aeongames.com <mailto:kwizatz at aeongames.com>>>
>     wrote:
>     >
>     >     I am working on that as a personal project, but I am having
>     problems
>     >     with convex-convex collision detection,
>     >     once I am satisfied with the results I will commit the code
>     to the
>     >     UNSTABLE branch, I would commit it as-is if anyone
>     >     shows interest in helping out, but I see no use in commiting
>     a half
>     >     working solution.
>     >
>     >     Cheers!
>     >
>     >     Rodrigo Luque wrote:
>     >
>     >     >Is there any plan to support convex volume in ODE?
>     >     >Actually, I was using TRIMESH but a notice that no
>     collision is
>     >     detected
>     >     >when a geom is inside the TRIMESH, I mean not intersecting any
>     >     triangle of
>     >     >the TRIMESH. This is a big problem when using small hi-speed
>     >     geoms. This
>     >     >problem won't happen in a convex volume because it's easy
>     to find
>     >     out if the
>     >     >geom is inside or not.
>     >     >So, is there any plan?
>     >     >
>     >     >Thanks,
>     >     >
>     >     >_______________________________________________
>     >     >ODE mailing list
>     >     >ODE at q12.org <mailto:ODE at q12.org> <mailto:ODE at q12.org
>     <mailto:ODE at q12.org>>
>     >     >http://q12.org/mailman/listinfo/ode
>     >     <http://q12.org/mailman/listinfo/ode>
>     >     >
>     >     >
>     >     >
>     >
>     >     _______________________________________________
>     >     ODE mailing list
>     >     ODE at q12.org <mailto:ODE at q12.org> <mailto: ODE at q12.org
>     <mailto:ODE at q12.org>>
>     >     http://q12.org/mailman/listinfo/ode
>     >     <http://q12.org/mailman/listinfo/ode>
>     >
>     >
>     >
>     >
>     > --
>     > --------------------------------------------------
>     > Binh Nguyen
>     > Computer Science Department
>     > Rensselaer Polytechnic Institute
>     > Troy, NY, 12180
>     > --------------------------------------------------
>
>
>
>
>
> -- 
> --------------------------------------------------
> Binh Nguyen
> Computer Science Department
> Rensselaer Polytechnic Institute
> Troy, NY, 12180
> --------------------------------------------------