[ODE] brush simulation - next steps..

Patrick Enoch Hendrix_ at gmx.net
Tue May 15 10:04:18 MST 2007


I didnt get the last one:

> Is there anyway to keep the platform, where my bristles are
> connected, stable? Now if I push it down,
> the bended bristles bring it into a tilted position. But I would
> favour a flat platform that keeps it orientation!

Do you mean the "head" where the bristles are all connected to, is  
doing weird stuff because of the feedback from the bristles via the  
joints? You can always set the mass to a very large value. Although  
many ppl told me it is not good to do so, it works fine for me (mass  
= 1e20).

This is 100% correct way. It is like connecting the bristles to the  
moon or so.

Two alternatives:

a) you can unlink the heads geom from the body temporarely / link the  
bristles to NIL instead of the head temporarely - this way the engine  
will not add forces to the "head"-body, but I am not sure if the  
"head"'s velocity is correctly accounted for, because it was  
unlinked! no body - no velocity.

b) you can patch the engine not to apply cforce / self-generated  
torque to objects with (inv_mass<1e-19) or some special flag set (I  
invented the dxBodyImmune-flag for that). that is, the engine  
calculates correctional forces for both attached bodies to the links  
- and then you disregard the one body.

These are not 100% correct ways, but it works.


More information about the ODE mailing list