go2_rl_robotlab

Overview

Train Unitree Go2 with MoE-CTS.

This is a reproduction version of go2_rl_gym on RobotLab/IsaacLab.

Installation Guide

1. Install IsaacLab

Install Isaac Lab by following the installation guide.

Notice that we use certain version of IsaacLab packages, make sure:

isaacsim <= 5.1.0.0 # tested on 5.1.0.0
isaaclab <= 0.53.1 # tested on 0.53.1
isaaclab-rl <= 0.4.7 # tested on 0.4.7

2. Install customized RSL-RL and RobotLab

We uses a customized version of rsl_rl and robot_lab. To install it, run the following commands:

python -m pip install -e source/robot_lab
python -m pip install -e source/rsl_rl

Try examples

Use the following commands to train and play:

# Train
python scripts/reinforcement_learning/rsl_rl/train.py --task=Robotlab-Go2-v0 --headless --num_envs=4096

# Play
python scripts/reinforcement_learning/rsl_rl/play.py --task=Robotlab-Go2-v0 --num_envs=1024

Configuration

1. Modify source/robot_lab/robot_lab/tasks/go2/env_cfg.py for environment config.

2. Modify source/robot_lab/robot_lab/tasks/go2/rsl_rl_cfg.py for algorithm config.

3. Modify source/robot_lab/robot_lab/tasks/go2/__init__.py to add your own task with new config.

Acknowledgements

This repository would not exist without the following open-source projects:

• isaac_lab: Unified framework for robot learning built on NVIDIA Isaac Sim.
• rsl_rl: Reinforcement learning algorithms.
• robot_lab: RL Extension Library for Robots, Based on IsaacLab.

Related publications implemented in this repo:

• CTS: Concurrent Teacher-Student Reinforcement Learning for Legged Locomotion