an env for learning human motor models and exoskeleton control
In order to keep it experiment friendly, this codebase uses local scripts with very less modules to install/maintain. The only requirement is the sucessfull installation of mujoco and mujoco-py.
-
Download the MuJoCo version 2.1 binaries here: Mujoco.
-
Extract the downloaded
mujoco210directory into~/.mujoco/mujoco210.
If you want to specify a nonstandard location for the package, use the env variable MUJOCO_PY_MUJOCO_PATH.
To install mujoco-py, run the following
pip3 install -U 'mujoco-py<2.2,>=2.1'
After testing the sucessfull isntallation of mujoco-py,to test the codebase,run
cd hmm-ec/
python3 test_env.py
Since it is a evolving code base, at present kindly follow the usage and implementations in demos and utils.
Make a directory named data and a directory for your data with the following structure,
data
|- your_data
|- marker_data
|- c3ds # keep all your c3d files inside this folder
|- id_solns
|- ik_solns
To generate an end to end demo, simply run
python3 utils/compute_all.py --static_c3dfilepath data/your_data/marker_data/c3ds/static.c3d --trial_c3dfilepath data/your_data/marker_data/c3ds/trial.c3d --plot_id_solns --plot_ik_solns
- data directory addition in read me.
- update commands.md.
- COP point of application to be fixed.