MuJoCo model for the JVRC1 (virtual) humanoid robot #684
rohanpsingh
started this conversation in
Show and tell
Replies: 2 comments 3 replies
-
This is cool! Can you make some kinematic videos from the CMU dataset, like in this video (starting at 2:16)? The code you need is, I believe, here. |
Beta Was this translation helpful? Give feedback.
1 reply
-
Another question: what are you using for visualisation? (specifically those collapsible interactive overlays) |
Beta Was this translation helpful? Give feedback.
2 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
We introduce 🤖 jvrc_mj_description 🤖
jvrc1.mp4
jvrc1_walk.mp4
What is this?
JVRC1 is a virtual humanoid robot created for the Japan Virtual Robotics Challenge (held in 2015). The original VRML model files exist here. We have ported the model to MuJoCo along with the textures and minor modifications to the grippers: jvrc_mj_description.
The robot has a total of 44 (34 actuated + 10 passive) joints making it a 50-DoF robot (
nq=51
,nv=50
,nu=34
). Each gripper has 1 actuated joint and 5 "mimic" joints. Each end-effector (feet and hands) has a 6-DoF force-torque sensor. There is one IMU sensor placed at the root body.Why is this important? This is just a virtual robot?
While there exist several other models such as the mujoco Humanoid and the Humanoid_CMU widely used in the research community, we need a robot model that is more closely modeled after a humanoid robot than an actual human. This is important for creating and evaluating controllers meant to be deployed on real humanoid robots. In my opinion, it is rare to find open-source model files for physical humanoid robots of this size.
Yes, JVRC1 is a virtual robot with no corresponding physical form but it is close to one of the real humanoid robots of the "HRP" series, that some of you may already be familiar with, in terms of - kinematic structure, link masses and inertias, armature values, joint ranges, etc. We can even introduce actuator torque limits in the future.
I hope this can be used to show the generality of new controllers or learned policies and ideally, be used by researchers to make their research more accessible and reproducible.
Cheers to the new year~! 🥂
Beta Was this translation helpful? Give feedback.
All reactions