The rule to follow is this: Imagine the system with just the bodies and joints (constraints removed). That tree-structured system defines the mass matrix and must be non-singular. So you can weld massless bodies together if they are part of the tree system (i.e. using a weld joint rather than a weld...

No, sorry Nicos. Angular velocity is not the derivative of Euler angles (except in one special case: when all three Euler angles are zero). Otherwise you have to use Euler angle derivatives for u to get N=identity. Angular velocity is in fact not the derivative of any set of q's -- there is always a...

Hi, Nicos. No, that's a very reasonable question! The answer is that FunctionBased Mobilizer is a special case of Custom Mobilizer provided for convenience. The author of that class designed it to work only for mobilizers where qdot=u (that is, N=identity). The general Custom mobilizer supports N !=...

Nice! Congratulations.

You're welcome!

Here is a plot of the three components of the angular velocity of the shaft, expressed in the shaft frame. I set the initial conditions to (10, 1000, 1000) degrees/s. The y & z velocities vary between +/- 1414.21 degrees/s. plot.png The modified code I used is here: #include "Simbody.h" #include <io...

Hi, Naushad. I found a few minor problems with your implementation: - angular velocity in Simbody is given in radians/sec, not degrees/sec - angular velocity for a body B is given and reported as w_GB_G, that is, the angular velocity of body B in Ground, expressed in Ground; what you are interested ...

Hi, Naushad.

Please take a look at the DzhanibekovEffect example, which I think is similar to what you are trying to do. I suggest you run that one to see what it does. They you might want to modify that code to your specific case.

Regards,
Sherm

In the meanwhile you may want to consider replacing your SmoothHeightMap with a triangle mesh that approximates it.

Hi, Andres. The SmoothHeightMap has not been integrated into the contact subsystem yet. Looking through the code it appears that there are a few methods of SmoothHeightMap that are unimplemented that would be needed by the contact subsystem. Also a ContactTracker needs to be selected (or written) an...