Having troubles setting up sliding environment with constant interaction force #1925
-
|
Hello everybody, I am new to the mujoco simulations and I am having some troubles setting up an experiment: Some specifications about the objects:
bridge:
I added a site to the end of the wiper that has a force sensor attached to it, to measure the contact force. In order to make the wiper slide over the bridge, I also defined a motion planning law that acts on a given trajectory and the necessary orientations for the bridge's geometry. The frequency of this motion planning is 30 Hz while I have also specified an impedance controller that acts on the motion plannings action in 600 Hz (same frequency as for the mujoco simulation). In the XML-files of my simulation I specified the objects contact behaviour as follows: wiper: <body name="wiper" pos="0 0 0.107" quat="1 0 0 0">
<site name="ft_frame" pos="0 0 0" size="0.0001" rgba="1 1 0 0" type="sphere" group="1"/>
<inertial pos="0.00 0.001 -0.018" mass="0.709" diaginertia="55e-8 54e-8 53e-8"/>
<geom type="mesh" pos="0 0 0.0" name="wiping_surface_vis" mesh="wipe_eef" contype="1" group="1" solimp="0.05 0.9 0.1 0.1 6" solref="0.05 0.5" priority="1" condim="4" friction="0.0001 0.0005 0.0001" rgba="0.25 0.25 0.25 1" />
<!-- < For visusalization > -->
<body name="eef" pos="0 0 0.0536" quat="1 0 0 0">
<inertial pos="0 0 0" mass="0.0000001" diaginertia="0 0 0" />
<geom type="cylinder" pos="0 0 0" size="0.025 0.00001" rgba="0 1 0 0" group="1"/>
<site name="grip_site" pos="0 0 0" size="0.001 0.001 0.001" rgba="1 0 0 0.5" type="sphere" group="1"/>
...
</body>
</body>
...
<force name="eef_force" site="grip_site"/>bridge: <body name="object">
<geom name="bridge" pos="0.5 0 .97" mesh="bridge_mesh" type="mesh" density="50" group="0" condim="4" solimp="0.75 0.99 0.001" solref="0.02 0.2" friction="0.001 0.01 0.01"/>
</body>Note, that these snippets are only parts of the XML file specifications that specify the contacts. Expected behaviour:I expect the wiper to slide over the bridge with constant measured force. Current outcome:Although I am experiencing a constant velocity over the bridge (due to my motion planning), I am also experiencing a weird skipping motion of the wiper. This skipping motion leads the wiper to lift from the bridge which results in no contact force between the two objects for a slight moment. Then the wiper is getting in contact again leading to a force. This behaviour can be seen in the next image, where I have plotted the measured force of the sensor. In blue I plotted the overall changes of the force in the frequency of the simulation physics dt while I plotted the general reading times of the force sensor (frequency of my motion planning law) in red. My question:Is that problem originating from my simulation specifications? I mean, should I expect this kind of contact behaviour since my specifications of the simulation are bad chosen? Thanks in advance! |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
|
You may want to tune friction. Start by slightly increasing the friction values for both the wiper and bridge to see if that helps stabilize the contact. Your current friction values : Also try adjust Adjust Can you also try Hope it works |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your reply! I was not aware of the effect a too low friction might have. |
Beta Was this translation helpful? Give feedback.
-
|
For your interest: After changing this, I experienced much better behavior of my simulation. |
Beta Was this translation helpful? Give feedback.
For your interest:
I found out the solution to my problem. Above, you can see that I am changing the attributes of the body called wiper. However, since I am incorporating a force sensor, the wiper will never be in contact with the bridge. The body that interacts with the bridge is the geom cylinder which is part of the eef.
Therefore, I have to change those attributes.
After changing this, I experienced much better behavior of my simulation.