Feature Overview
Create a policy to get the humanoid robot to stand up.
Describe the solution you'd like
Use IsaacSim/MuJoCo to train the RL policy to stand the humanoid robot from a position laying down. Perhaps try randomizing a bit further from different positions (seated, etc).
Steps to Implement
Use the same domain randomization seen in locomotion, but change up the rewards and environment initialization for standing up.
Testing
Train in one sim, validate in another, then deploy to the real robot.
Feature Overview
Create a policy to get the humanoid robot to stand up.
Describe the solution you'd like
Use IsaacSim/MuJoCo to train the RL policy to stand the humanoid robot from a position laying down. Perhaps try randomizing a bit further from different positions (seated, etc).
Steps to Implement
Use the same domain randomization seen in locomotion, but change up the rewards and environment initialization for standing up.
Testing
Train in one sim, validate in another, then deploy to the real robot.