This repository store supplementary materials for a paper, "Muscle Activation Analysis from Gait Kinematics and Reinforcement Learning".
All the walkthrough here are tested on Ubuntu 18.04.
sudo apt-get install libtinyxml-dev libeigen3-dev libxi-dev libxmu-dev freeglut3-dev libassimp-dev libpython3-dev python3-tk python3-numpy virtualenv ipython3 cmake-curses-gui
We strongly recommand that you install boost libraries from the source code (not apt-get, etc...).
-
Download boost sources with the version over 1.66.(https://www.boost.org/users/history/version_1_66_0.html)
-
Compile and Install the sources
cd /path/to/boost_1_xx/
./bootstrap.sh --with-python=python3
sudo ./b2 --with-python --with-filesystem --with-system --with-regex install
- Check yourself that the libraries are installed well in your directory
/usr/local/
. (or/usr/
)
If installed successfully, you should have something like
Include
/usr/local/include/boost/
/usr/local/include/boost/python/
/usr/local/include/boost/python/numpy
Lib
/usr/local/lib/libboost_filesystem.so
/usr/local/lib/libboost_python3.so
/usr/local/lib/libboost_numpy3.so
apt-add-repository ppa:dartsim/ppa
apt-get install libdart6-all-dev
You should first activate virtualenv.
virtualenv /path/to/venv --python=python3
source /path/to/venv/bin/activate
Then, install.
pip3 install torch torchvision
pip3 install numpy matplotlib ipython
cd learning
mkdir build
cd build
cmake ..
make -j8
- Run Training
source /path/to/virtualenv/activate
cd learning/python
python3 main.py -d ../dataFAT/metadata_walking.txt
All the training networks are saved in training/nn folder. The training can be continued by
python3 main.py -d ../dataFAT/metadata_walking.txt -m current
- Render 3D visualization of the learning in GUI.
source /path/to/virtualenv/activate
cd learning/build
./render/render ../dataFAT/metadata_walking.txt ../nn/xxx.pt ../nn/xxx_muscle.pt
In GUI, you can press 'o' to show the bone alignement.
- Export muscle activation
source /path/to/virtualenv/activate
cd learning/build
./actExport/actExport ../dataFAT/metadata_walking.txt ../nn/xxx.pt ../nn/xxx_muscle.pt
- Metadata file The metadata file is used to tell the program where is the motion capture file, skeleton, and muscle details. All these resources can be generated by source code in data_preporcessing directory by running 01, 02, 04, and 05 python script in order.