[ODE] Simulator explodes on zero Quaternion normalisation

Garvin Haslett G.A.Haslett at sussex.ac.uk
Tue Jun 1 18:28:31 MST 2004


I'm constructing a body made of two spheres connected by four thin 
cylinders.  I've dropped the body from a height and applied a force at 
initialisation to it to cause it to roll.  Now what happens is that the 
body rolls clockwise before starting to roll anticlockwise back towards 
its equilibrium point.  The moment the  anticlockwise rotation begins 
the body disappears and the the following error message is output:

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

Tracing the error back in the debugger it appears that the call to 
dNormalize4 is being made from within moveAndRotateBody and is being 
done so to normalise the quaternion associated with the body.  Now at 
this point the trail goes dead for me as I don't really understand what 
a quaternion is, why it should not be zero and why, crucially, it is 
zero in this context.

Any help much appreciated,

Garvin.



More information about the ODE mailing list