Ah yes, that works nicely. I didn't know you could define contact pairs like that, that seems to be the trick. I can't think of anything else.
MuJoCo uses the convex hull to compute contact: http://www.mujoco.org/book/computation.html#Collision So your hole is simply ignored. You can...
It seems that dry friction is missing (or too small) between the ball and plates, hence a small velocity is needed for the friction force. I don't...
I made a little example model to test this (see bottom). It's fairly easy to make the ball almost stick, but it seems to always trickle down very...
You could use a height field for this, it's probably easiest. However, the cylinder will always be 'pixelated'. Alternatively you could use a...
Does it work if you tilt it only slightly? E.g. at 10 degrees, 30, 50, etc. up to 90. Also, how do you apply the pressure to push them together?
You could search through d->contact directly to look for the contact force. If you use mj_contactForce() to find the contact force it will be in...
I wanted to come to this to complete it. You can use mju_derivQuat() to compute the quaternion derivative based on the angular velocity. You...
I don't think there is a built-in feature for this. Instead, I now determine it numerically. Which is okay because I only need to do it once,...
Nevermind, it's body-fixed. This was easily tested with a simulation of a floating body and setting an initial orientation and angular velocity....
Title pretty much says it all. If I have a model with a freejoint, which has some rotation around all axes, and I prescribe some angular velocity...
Do any of the non-GUI MuJoCo sample applications work? Like ./derivative for example? Because it sounds like a graphics error. Maybe this is...
Yeah, though what exactly is your question? You can find my inverse kinematics solution here:...
There is no official package for MuJoCo integration in ROS. But there is nothing keeping you from making such an integration. See also this reply...
Separate names with a comma.
