Revision history [back]

OK. Thanks a lot! So, from my understanding of how ODE works, the problem here arises when the wheels and the ground plane make contact. ODE creates contact joints and the erp and cfm of those contact joints are calculated from the kp and kd of each of the bodies (wheels and ground plane) the way you explained before. Is this correct?

We can change the kp and kd of the ground plane directly in the world file. However, in the same world file we can also set cfm and erp values for the general case inside a <physics:ode> tag. How are those values related? How do those "general" cfm and erp influence the calculation of new contact joints' cfm and erp?

Besides, if I'm correct, I would also have to set the kp and kd values of the wheels. My wheels are modeled in a urdf file. As far as I know from the urdf documentation and tutorials, links do not have any kp and kd parameter. How could I set those parameters, if they have to be set?

Finally, fixing the kp and kd of the ground plane of the world file, I achieved a reasonable solution (kp=1.0e+9 and kd=1.0e+5) that seems to work fine. However, a "small" portion of the wheels are under the floor. Does it have any effect in the behavior of the platform or should I forget about it?

Just in case it might be useful for you, I put the links to download the files needed to make my platform work in Gazebo (I use dropbox for that purpose, since I didn't see here any other way to do it; hope it's fine for you): http://dl.dropbox.com/u/23291294/platform_description/base.gazebo.xacro http://dl.dropbox.com/u/23291294/platform_description/base.transmission.xacro http://dl.dropbox.com/u/23291294/platform_description/experimental_base.urdf.xacro http://dl.dropbox.com/u/23291294/platform_description/empty.world