I have this simple skel file, which has four identical boxes floating in space (no gravity). The middle two are free, while the other two are weld to world frame.
If I give the middle two boxes a same initial x velocity but with opposite direction, (say (1,0,0) and (-1,0,0)), and set the restitution to 1.0, then in theory the middle two boxes will keep bouncing back and forward between the two weld boxes. And that is indeed what I saw in DART.
However, if I replace all the "0.000" above with pi/2 ("1.570796") - that is, I let all 4 boxes pitch pi/2, in theory that should change nothing. But when I ran the simulation, I saw the middle two boxes quickly losing symmetry after 2 bounces, bypassing the weld boxes, and flying away.
As another example to show this unsymmetrical behavior, I tried to let two spheres free fall, one with no initial velocity and another starting with spinning (say, w_0 = 10 rad/s). In theory spinning should not affect free fall at all (and that is what I see in Mujoco). But from DART output I observed that the spinning ball started to gain +x velocity and position quickly.
I tried to dig into the code base a little bit. I am not sure whether this phenomena is related to the Lie Group operators, which I have absolutely no knowledge of.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.
Hello everyone,
I have this simple skel file, which has four identical boxes floating in space (no gravity). The middle two are free, while the other two are weld to world frame.
If I give the middle two boxes a same initial x velocity but with opposite direction, (say (1,0,0) and (-1,0,0)), and set the restitution to 1.0, then in theory the middle two boxes will keep bouncing back and forward between the two weld boxes. And that is indeed what I saw in DART.
However, if I replace all the "0.000" above with pi/2 ("1.570796") - that is, I let all 4 boxes pitch pi/2, in theory that should change nothing. But when I ran the simulation, I saw the middle two boxes quickly losing symmetry after 2 bounces, bypassing the weld boxes, and flying away.
As another example to show this unsymmetrical behavior, I tried to let two spheres free fall, one with no initial velocity and another starting with spinning (say, w_0 = 10 rad/s). In theory spinning should not affect free fall at all (and that is what I see in Mujoco). But from DART output I observed that the spinning ball started to gain +x velocity and position quickly.
I tried to dig into the code base a little bit. I am not sure whether this phenomena is related to the Lie Group operators, which I have absolutely no knowledge of.
Thanks, Yifeng