Drake Concepts Tutorial
(Build tested on Drake binaries from 20180604.)
You'll have to install Drake (from binaries or source, your choice) following the instructions on the Drake website, or in this guide. Note that this does not work on recent Drake versions any more; for the best experience, use the Drake binaries from 20180604. For a codebase similar to this maintained against up-to-date Drake revision, see this repo or, better yet, the codebase maintained by MIT's 6.881 course.
You'll also need to use jupyter to view the notebook (.ipynb) files, graphviz, and meshcat-python to view the 3D visualizations in some of the examples. You'll also need some more standard libraries (e.g. numpy). You can install all of these with
apt-get install graphviz pip install jupyter graphviz meshcat numpy matplotlib
And finally, you'll need to have the Drake textbook example code available and on your PYTHONPATH. You can pull it down with
git clone https://github.com/RussTedrake/underactuated ~/underactuated
and add it to your PYTHONPATH with
(you probably want to add that to the end of your
To view the notebook files, use the command
jupyter notebook from a terminal in the same directory (or a parent directory) as the notebook files. Then use the web-browser-based notebook browser to open up the notebook files, and use the notebook interface to play with the notebook. See a guide like this one for info on how to use the Jupyter notebook. Be sure to have both Drake and the textbook code on your PYTHONPATH before launching jupyter!
To run the Kuka simulation, first run
meshcat-server in a new terminal. It should report a web-url -- something like
127.0.0.1:7000/static/. Open that in a browser -- this is your 3D viewer. Then run
python kuka_pydrake_sim.py and you should see the arm spawn in the viewer before doing some movements.