Bluebird Gymnasium

bluebird-gymnasium is suite of gymnasium environments for air traffic control (ATC). The environments are based on bluebird-dt (an ATC simulator). The environments support research in agent-based learning (e.g. reinforcement learning) for ATC. It supports either single agent or multi-agents scenarios.

Installation

Install from PyPI:

pip install bluebird-gymnasium

Or, if you're using uv, you can add it to your environment:

uv add bluebird-gymnasium

Getting started

Basic usage

bluebird-gymnasium currently supports the following environments/airspace: X sector, Y sector, I sector, Xplus sector and Springfield sector.

To instantiate a X sector environment with the default config, run:

import gymnasium as gym
import bluebird_gymnasium

env = gym.make("SectorXEnv-v0")

Sample agents

Below, an example agent that takes random actions.

import gymnasium as gym
import bluebird_gymnasium

env = gym.make("SectorXEnv-v0")
obs, info = env.reset()
done = False

while not done:
    action = env.action_space.sample()
    obs, reward, done, truncated, info = env.step(action)

Various examples of using the bluebird-gymnasium package can be found in the form of Jupyter notebooks in the examples directory or in the documentation for the AI(r) Traffic Controller Challenge.

Documentation

The documentation of the latest release is available at https://docs.projectbluebird.ai.