This is a fork + my own stuff. Old Project, could be useful for someone. I have successfully trained a PPO agent on the simulator and run on the real hardware for swinging up Quanser Rotary Pendulum. I have created simulators for it seperately, I beleieve they are not in this repo but will be uploaded when I find time.
Happy hacking
Simulated environments and real robot interfaces for a set of Quanser platforms.
This package is compatible with Python 3.6.5. To install the it, execute
pip3 install -e .
To confirm that the setup was successful, launch a Python3 console and run
import gym
import quanser_robots
env = gym.make('Qube-v0')
env.reset()
env.render()
If installation worked well, proceed to the robot-specific documentation
If you are getting errors during installation that you do not know how to fix, check below whether the requirements are satisfied, and if necessary follow the detailed installation instructions.
The main requirement comes from the graphics library vpython
which is used for rendering 3D environments.
It requires Python >= 3.5.3 (preferably Pyton 3.6.5).
Note that the default version of Python on Ubuntu 16.04 is Python 3.5.2,
so visualization will not work with it.
You can still use the environments though, just don't call env.render().
Awesome stuff developed by
- Boris Belousov
- Fabio Muratore
- Hany Abdulsamad
- Jonas Eschmann
- Robin Menzenbach
- Christian Eilers
- Samuele Tosatto
- Michael Lutter
Add your name if you contributed, so that we know whom to blame.