You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Models of floating base systems are not correctly loaded using buildModelFromMJCF. For example, in quadruped models, the parent link for all the hips is incorrectly set to 0, instead of being assigned to the base or trunk link (which should be 1). As a result, some computations are wrong (crba, nonLinearEffects).
Expected behavior
The hips' parent link should be the trunk (link 1).
Thank you for reporting the issue.
It has been fixed (#2386) but has not been released yet. It will be in the next release of pinocchio.
In the meantime, if you need to use the mjcf version of go1, i recommend you to install pinocchio from source, using the version on the devel branch.
Hi all,
Bug description
Models of floating base systems are not correctly loaded using
buildModelFromMJCF
. For example, in quadruped models, the parent link for all the hips is incorrectly set to 0, instead of being assigned to the base or trunk link (which should be 1). As a result, some computations are wrong (crba, nonLinearEffects).Expected behavior
The hips' parent link should be the trunk (link 1).
Reproduction steps
This can be reproduced by loading go1 model from mujoco-menagerie and compared to loading from example_robot_data:
Inspecting the models:
Additional context
I only tested for quadrupeds (go1 and anymal), but I believe it should also affect humanoids.
Let me know if I'm missing something. Loading manipulators (e.g. Panda) from mujoco menagerie works fine. Thank you in advance.
System
The text was updated successfully, but these errors were encountered: