/cassie-rl

An OpenAI Gym style reinforcement learning interface for Agility Robotics' biped robot Cassie

Primary LanguagePython

cassie-rl

An OpenAI Gym style reinforcement learning interface for Agility Robotics' biped robot Cassie. CassieEnv is a reimplementation of: https://arxiv.org/abs/1803.05580.

Installation

  1. Place mjpro150/ and mjkey.txt in the cassie/cassiemujoco folder (see http://www.mujoco.org/ for info on downloading MuJoCo and obtaining a license).
  2. Place libcassiemujoco.so in the cassie/cassiemujoco folder (see https://github.com/osudrl/cassie-mujoco-sim for info on compiling libcassiemujoco.so). Temporary note: cassie-rl currently uses a different version of cassie-mujoco-sim with a foot position API. For now see my fork of the former here: https://github.com/p-morais/cassie-mujoco-sim.
  3. Run $ pip install -e . in root folder (the one that contains setup.py).

To test installation:

$ python3
> from cassie import CassieEnv
> env = CassieEnv("walking")
> env.render()

TODO:

Add the OpenAI gym environment registration code so that the environment can be created with gym.make().