Mujoco210 is required for simulation and should be placed in ~/.mujoco . To build the mujoco simulation, make in ./cassie-mujoco-sim folder. Then copy libcassiemujoco.so to arm-cassie/arm_cassie_env/cassie_m
Stable-baselines3 in required for rl algorithm. Create conda env:
conda env create -f environment.yml
conda activate lg
Run python train.py to train the agent defined in cassieRLEnvStable.py.
Run python play.py to test the policy.
Run python arm.py to run the task.
XML Models for possible use are in ./models
https://github.com/osudrl/apex