SEA-Nav: Efficient Policy Learning for Safe and Agile Quadruped Navigation in Cluttered Environments

Project Website: https://11chens.github.io/sea-nav

Create a new Python virtual environment with Python 3.8:

conda create -n sea_nav python=3.8
conda activate sea_nav

• Download and install Isaac Gym Preview 4 from NVIDIA Developer.
• Install the python package:

cd isaacgym/python && pip install -e .

• Clone this repository
• Install the package:

cd training/rsl_rl && pip install -e .

cd training/legged_gym && pip install -e .

To start training in headless mode:

python training/legged_gym/legged_gym/scripts/train.py --headless

To visualize and test a trained policy:

python training/legged_gym/legged_gym/scripts/play.py

For instructions on deploying to real-world robots, please refer to the deployment README.