[ODE] ODE Message 2: vector has zero size in dNormalize4()

[dvir schirman]

Hi

This is a sketch of my robot:

||||||||||||||||||||||||||||||||||--|||||||||||||||||||||||||||||||||||--||||||||||||||||||||||||||||||
        ^                           ^                          ^
        ^                           ^                          ^
    body1                     body2                  body3

The bodies are connected using a hinge joint.
The robot is layed on the ground, and trying to stand up.
While trying the robot falls down several times, I suspect that one of the
falls causes the mentioned problem.
After I get "ODE message 2:...", the angles between the bodies become NaN'
and therefore' my simulation is ruined.

Does anyone have an idea how can I solve it?

Thanks,
Dvir

On 7/17/06, Remi Ricard <remi.ricard at simlog.com> wrote:
>
> Hi,
>
> >
> >     > I'm running a simulation of a robot at ODE.
> >     > My robot is made out of three rectangular boxes connected with
> >     two hinges.
> >     > My simulation is working. But, when I run it I get the following
> >     > message repeatedly:
> >     >
> >     > "ODE Message 2: vector has zero size in dNormalize4()"
> >     We had this message when collision points where very close
> >     together on 2
> >     different geometries.
> >                                 | |
> >     ___geom1______||______Geom2_____
> >                  *
> >                  ^-- Collision point
> >
> >
> >     Remi
> >
> >
> > I think I get this message because of the same reason.
> > Do you have any idea how to solve it?
> I was able to solve my problem by moving one of the geom a little bit
> or changing how they touch
>
> I had previously something like this
>
>   |  | <- geom1
>   |  |
>   |  |
>   |  |
>   |  |
>   |  ||-------------
>   |  ||              Geom 2
>    -- |-------------
>      *
>      ^-- This point was causing problem
>
> and it was changed for
>
>   |  | <- geom1
>   |  |
>   |  |
>   |  |
>   |  |
>   |  |
>   |  |
>    --
>    |-----------
>    |
>    |-----------
>
>   ^
>   |-- The geom are not align here so they can not touch another geom
>       at the same time.
>
> Remi
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://q12.org/pipermail/ode/attachments/20060717/cd39d1b5/attachment-0001.htm