Real-Time Physics Simulation Forum

Try giving the keyframed platform a velocity value that agrees with its motion. The velocity won't actually move the platform (it is moved via the keyframed animation of course), but it will make objects that collide with it respond as it if were moving.

Kinematic objects are supposed to be controlled by a custom class derived from btMotionState . The DynamicsWorld will call the MotionState's getWorldTransform() every frame, expecting the MotionState to provide the new transform (and maybe setting the object's correct velocity). Setting the velocity...

The body's active/inactive state is influenced by its world-frame velocity, which never goes to zero when it is sitting on a moving kinematic object -- the DynamicsWorld will never allow the object to settle and it will always bounce and jitter a little bit. You might be able to do what you want usi...

jorrit, I assume you're actually using Vim instead of Vi . There are plugins for Vim that would allow you to jump to function definitions and implementations, much like some of the shortcuts in an IDE. Finding these plugins and figuring out how they work is left as an exercise for the reader. Person...

My experience has been that when I am seeing an inexplicable bug it always sensible explanation when I figure out what the real problem is.

I've had one other idea: does your object happen to have zero mass?

Another idea: when a btRigidBody is created it is set static by default (via the default btCollisionObject constructor from which btRigidBody derives). Perhaps you never cleared that flag and the object actually has both kinematic and static flags set? Try clearing both of flags whenever you try to ...

Did you perhaps clear the object's gravity? Or maybe you didn't set the gravity of the world?

Try giving it some velocity after activating it as a test.

Ah, I think your problem is that forceActivationState() doesn't reset the activation timer, so the object is getting deactivated almost immediately. Here is an excerpt from btCollisionObject.cpp: void btCollisionObject::forceActivationState(int newState) const { m_activationState1 = newState; } void...

So you want bodyA to rotate around bodyB . You can slam bodyA 's position relative to its "parent" bodyB using a method very similar to my advice in this thread . If you want to avoid slamming positions and couple bodyA to bodyB with forces and velocities as johnsonalpha suggested then you...

I don't think there is an simple way to do it, but you can manually compute the transform of each body. First you'd need to compute the relative transform of each body in its parent's local frame. Then after rotating the palm you would compute the new world-frame transforms down each chain of linked...

It sounds like the reverse motion of the forklift is a result of a slight oscillation of the solution of the system of constraints and how that interacts with details of the contact points with the floor. I note that in the problem image the forklift's front wheels appear to be off the floor, so it ...

To clear just the CF_KINEMATIC_OBJECT flag you could do something like this:

No, I don't think there is functionality in Bullet to measure the potential energy of two objects that are being compressed. As far as I know the engine does not track internal stress forces at contact points during the simulation. Instead constraints are solved for new velocities and the internal f...

You need to remove the [/b]CF_KINEMATIC_OBJECT[/b] flag to make the object dynamic. In other words, consider the implementation of btCollisionObject::isKinematicObject(): SIMD_FORCE_INLINE bool isKinematicObject() const { return (m_collisionFlags & CF_KINEMATIC_OBJECT) != 0; } BTW: Kinematic obj...

My advice would be to avoid Euler angles and just compute the new rotation directly: Suppose you know both forward axes of your character projected onto the horizontal plane and normalized : oldCharacterForward (where it is now) and newCharacterForward (where you want it to be). You can compute the ...