Reinforcement learning with musculoskeletal models in OpenSim
Reinforcement learning with musculoskeletal models in OpenSim
Design artificial intelligent controllers for the human body to accomplish diverse locomotion tasks. Participate in the NeurIPS 2019 challenge to win prizes and fame.
Learn more about the challenge →Use our musculoskeletal reinforcement learning environment for other projects in computer science, neuroscience, biomechanics, etc.
Learn more about osim-rl →Quick-start Instructions
~ $ conda create -n opensim-rl -c kidzik opensim python=3.6.1
~ $ source activate opensim-rl
~(opensim-rl) $ conda install -c conda-forge lapack git
~(opensim-rl) $ pip install git+https://github.com/stanfordnmbl/osim-rl.git
~(opensim-rl) $ python
from osim.env import ProstheticsEnv
env = ProstheticsEnv(visualize=True)
observation = env.reset()
for i in range(200):
o, r, d, i = env.step(env.action_space.sample())
In 2017 we used osim-rl in a challenge at NIPS were participants were asked to build controllers for running. They did great :)