Files
bullet3/examples/pybullet/gym/pybullet_envs/minitaur/envs
Krishnan Srinivasan 0d4108f307 Fix call to convert_to_leg_model
Fix call to convert_to_leg_model which is defined as a static method without an underscore in MinitaurExtendedEnv
2019-10-29 16:00:55 -07:00
..

Simulated Minitaur Environments

This folder contains a number of simulated Minitaur environments implemented using pybullet.

The following two environments are used in the RSS paper "Sim-to-Real: Learning Agile Locomotion For Quadruped Robots":

  • Galloping environment: minitaur_reactive_env.py
  • Trotting environment: minitaur_trotting_env.py

The rest are experimental environments.

Prerequisites

Install TensorFlow

Install OpenAI gym

pip install gym

Install ruamel.yaml

pip install ruamel.yaml

Examples

To run a pre-trained PPO agent that performs the galloping gait

python minitaur_reactive_env_example.py

To run a pre-trained PPO agent that performs trotting gait

python minitaur_trotting_env_example.py 

Authors

  • Jie Tan
  • Tingnan Zhang
  • Erwin Coumans
  • Atil Iscen
  • Yunfei Bai
  • Danijar Hafner
  • Steven Bohez
  • Vincent Vanhoucke