Real-Time Physics Simulation Forum

The finger bones of your hand would be driven by the data glove, but would they be driven kinematically or dynamically? Kinematic objects are guided by external logic (keyframed path, script, or perhaps real-time data acquisition) and can bump other objects that are dynamic, but cannot be influenced...

Yes, I believe Bullet could perform planetary simulations with decent accuracy, however if you're using true solar system units then you definitely need to compile Bullet with double precision otherwise floating point errors will be too big.

Directly setting the velocity IS how you change it. I don't see how that "overwrites the collision" as you put it. If you want to constantly reorient the velocity according to the direction of the ship I would suggest implementing an Action that does just that and then give the action to t...

Beware: if the mass discrepancy between jointed parts is too high then you can get instabilities. Best if the two connected masses are not more than a factor of 10 from each other. The closer to unity the better, especially when simulating more joints. I once hacked the Bullet RagdollDemo to use btF...

(1) Yes, I think it is possible to perform custom handling of certain collisions. Such handling would be mostly custom code that you would need to write. There is probably more than one way to do it -- it is possible to catch and modify contact points in Bullet, but Bullet also supports soft-body co...

btDiscreteDynamicsWorld derives from btCollisionWorld and correctly calls updateAction() on actions that have been added. Use that.

That is a bug in the btKinematicCharacterController, which should have been Demo code rather than put in the library. Best practice is to copy the btKinematicCharacterController implementation into your own project and then modify it heavily to suit your own needs. That said (IANAL), the btKinematic...

After looking at the code it appears that they way to get the profiling information is to call CProfileManager::dumpAll() and it will print to stdout the times spent in the various "scopeName" instances of the BT_PROFILE("scopeName") macros that are sprinkled about in the btDiscr...

The easiest way to do this would be to modify the restitution and/or friction of the objects that are colliding. The restitution is a coefficient in the range [0,1] that modifies the amount of kinetic energy that is preserved in the relative frame during impact and can be set on a per-object basis. ...

The constructor for btBoxShape epxects the half extents NOT the full extents. Similarly for the btCapsuleShape. Check the documentation for the box and cylinder shapes.

The KinematicCharacter just takes a velocity, which means that if you supply a diagonal velocity then the controller will go that way. The solution lies in changing your own code that captures the "commands". BTW I notice that your functions are incorrectly named. PlayerObject::addToSpeedF...

In short, I pretend that I wrote btKinematicCharacterController from scratch and then I modify it. The longer story: In my case I've got a large project with several files that were written from scratch. These live in directories that are completely independent of Bullet and its Demos (I use Bullet ...

http://www.bulletphysics.org/Bullet/phpBB3/viewtopic.php?f=9&t=10458 I think I encountered this same problem when using the btKinematicCharacterController . When I looked into it I discovered that there is a bug in the getNormalizedVector() helper function in btKinematicCharacterController.cpp f...

Not in my experience, especially when there are transformations involved. Yes, your box probably has identity rotation but I'll bet its not located at the origin. When vectors are transformed into the local-space of the box shape and back out floating point errors creep in. Put your box 1000 times f...

All of those values are approximately the same. You're seeing floating point error.