"constraint inconsistent after World Space constructor"

"constraint inconsistent after World Space constructor"

I am loading a rotation matrix and translation data from file to position and orient a hinge constraint. The data looks fine, but a breakpoint is trigged by Havok:

This is the code at the breakpoint:

hkpLimitedHingeConstraintData *cnt = new hkpLimitedHingeConstraintData;
cnt->setSolvingMethod( hkpConstraintAtom::METHOD_STABILIZED );
cnt->setMinAngularLimit( min_angular_limit );
cnt->setMaxAngularLimit( max_angular_limit );
cnt->setMaxFrictionTorque( friction );
 cnt->setInWorldSpace( body_transform, IDMatrix, pivot_position, axis_vector );  <---- BREAKPOINT HERE

This is the breakpoint message: Constraint\Bilateral\LimitedHinge\hkpLimitedHingeConstraintData.cpp(87): [0x3A0A5294] Assert: Members of LimitedHinge constraint inconsistent after World Space constructor..

These are the values of the data passed to setInWorldSpace:

body_transform:
   rotation: 0.99997663,0,0 - 0,0.99997663,0 - 0,0,0.99997663 (X,Y,Z)
   translation: -288.00000,559.99890,863.94623,0.99997455 (X,Y,Z,W)

IDMatrix:
   rotation: 1,0,0 - 0,1,0 - 0,0,1 (X,Y,Z)
   translation: 0,0,0,1 (X,Y,Z,W)

pivot_position: -288.00000,512.00000,890.14362,0.99997455 (X,Y,Z,W)
axis_vector: 0,0.99997663,0,0 (X,Y,Z,W)

min_angular_limit: -1.5393804
max_angular_limit: 1.5393804
friction: 0.0

Any help would be greatly appreciated. If any more information is needed, please let me know and I'll post it.

Thank you

 

 

 

 

 

 

3 post / 0 nuovi
Ultimo contenuto
Per informazioni complete sulle ottimizzazioni del compilatore, consultare l'Avviso sull'ottimizzazione

I believe I may have figured out what the problem was. After normalizing the rotational axes of "body_transform", it did not trigger a breakpoint. It seams like 0.99997663 would be close enough to unit-length, but I guess not?

Correct me if I'm wrong. It may just be that this fixed something else related to the matrix unit lengths.
Thank you again for any help.

Hi Bobby M.--

The un-normalized rotation axis was probably at fault. The verification logic tests for orthonormal transforms with a tolerance of 0.00001, which would cause 0.99997 to fail. If the issue crops up again, please let me know.

Thanks,
--Tim

Accedere per lasciare un commento.