[ODE] Re : Triangle Mesh Creation Time.

Patrick Enoch Hendrix_ at gmx.net
Tue Jun 12 15:23:07 MST 2007


Hi,

what I did in this case, is replace the trimesh with a new one (like  
you) every frame (and keep this constant over the subframes).
This way, I have a "jumping" vertices mesh from frame to frame. For  
the records: I am not writing realtime software.

What gets lost is the "velocity" of a single vertex (imagine the  
trimesh is static = velocity=0), and imagine an object close to the  
trimesh , not moving ATM:
- this frame is doesnt penetrate the mesh - OK
- next frame the next suddenly expands and the objets penetrate - the  
correctional velocity velocity is determined by the "maximum  
correction velocity" and "ERP" solely. NOT by relative velocity of  
vertices that make up the mesh (e.g. the collision-triangle).

So you derived the net-velocity of the collision-point within a  
collision-triangle, and adjusted the max-correction-vel of the  
contact accordingly? Does it work well?

Best,
Patrick

On 12.06.2007, at 18:38, Gonçalo Lopes wrote:

> Hi there,
>
> I'm interested in this problem myself. Me and my team are using ODE to
> integrate dynamic interaction in real-time with image processing data
> and we had the same issues. We got along quite well with dynamically
> creating trimeshes every image frame, but these were low triangle
> count trimeshes.
>
> The ideal solution, one would think, would be to create just one
> single trimesh, but then to allow dynamic manipulation of its
> vertexes? Would this be feasible with the current trimesh
> implementation?
>
> Also, another important question for dynamic interaction is how to
> correctly set up parameters on the ODE side so that the newly created
> entities really behave like continuous objects, and not like
> instantaneously created geometries. We've achieved this by hammering
> the derived velocities for each frame into the ODE bodies, but maybe
> this is not recommended.
>
> Does anyone here has any experience with setting up dynamically
> created interactions with ODE? I feel that the need to incorporate
> this is rising, given the increase in the augmented reality and
> real-time physical user interaction trends.
>
> Best regards,
>
> Gonçalo
>
>
> On 6/12/07, Patrick Enoch <Hendrix_ at gmx.net> wrote:
>> Did I miss something? Why not create one trimesh?
>>
>>
>> On 12.06.2007, at 14:37, Frédéric Brachfogel wrote:
>>
>> > Hello,
>> >
>> >
>> > Implementing dynamic trimesh  collision is a simple solution to the
>> > problem of long data structure creation time ;)
>> >
>> > Concerning the contact points, I found a simili answer in the
>> > documentation, but I have to be sure. Here is the sentence :
>> > "
>> > The convention is that if body 1 is moved along the normal vector
>> > by a distance depth (or equivalently
>> >
>> > if body 2 is moved the same distance in the opposite direction)
>> > then the contact depth will be reduced to
>> >
>> > zero. This means that the normal vector points "in" to body 1.
>> >
>> > "
>> >
>> > Is that true that with the contact point array filled by the user
>> > function, ODE perform the dynamic physic changes on the body ? It
>> > really seems to be the case, but I have to be sure too.
>> >
>> > Moreover, in the case of a point patch (contact along surfaces), do
>> > the Depth be devided by the number of points in the patch ?
>> >
>> >
>> > F.Brachfogel
>> >
>> > --------------------------------------
>> > Chronophage Games
>> > 3-5 rue Lashermes
>> > 43000 Le Puy en Velay
>> > FRANCE
>> >
>> >
>> > > Date: Tue, 12 Jun 2007 08:13:07 +0200
>> > > From: jsinecky at tiscali.cz
>> > > To: fred3d2 at hotmail.com
>> > > CC: ode at ode.org
>> > > Subject: Re: [ODE] Triangle Mesh Creation Time.
>> > >
>> > > Hi,
>> > >
>> > > Frédéric Brachfogel wrote:
>> > > > Anyone for this no ? Is my question clear ?
>> > > >
>> > > > Thank you,
>> > > >
>> > > >
>> > > > F.Brachfogel
>> > > >
>> > > > --------------------------------------
>> > > > Chronophage Games
>> > > > 3-5 rue Lashermes
>> > > > 43000 Le Puy en Velay
>> > > > FRANCE
>> > > >
>> > > >
>> > > >
>> >  
>> --------------------------------------------------------------------- 
>> -
>> > --
>> > > > From: fred3d2 at hotmail.com
>> > > > To: fred3d2 at hotmail.com
>> > > > Subject: RE: [ODE] Triangle Mesh Creation Time.
>> > > > Date: Sat, 9 Jun 2007 14:11:31 +0000
>> > > >
>> > > >
>> > > >
>> > > > Thank you for reply.
>> > > >
>> > > > A simple solution for me is to use custom collision class (?).
>> > > >
>> > > This is not clear to me .... do you think implementing dynamic
>> > trimesh
>> > > collision on your own is a simple solution?
>> > > >
>> > > >
>> > > > I have a question about this :
>> > > >
>> > > > The user collision function fills a contact data array.
>> > > > How will the engine (ODE) perform the pose changes on the
>> > > > geom to simulate the contact so that there is no
>> > > > inter-penetrations ? In other terms, how must I fill the
>> > > > contact data ? In the case of patch ?
>> > > >
>> > > This is covered in ODE documentation, try to search it for
>> > > "dContactGeom". Or even better, read it through before going  
>> on with
>> > > some serious ODE development.
>> > > >
>> > > >
>> > > > F.Brachfogel
>> > > >
>> > > > --------------------------------------
>> > > > Chronophage Games
>> > > > 3-5 rue Lashermes
>> > > > 43000 Le Puy en Velay
>> > > > FRANCE
>> >
>> >
>> > http://www.windowslive.fr/hotmail/default.asp Windows Live Hotmail
>> > _______________________________________________
>> > ODE mailing list
>> > ODE at ode.org
>> > http://ode.org/mailman/listinfo/ode
>>
>>
>> _______________________________________________
>> ODE mailing list
>> ODE at ode.org
>> http://ode.org/mailman/listinfo/ode
>>




More information about the ODE mailing list