A gymnasium environment for autonomous decision-making in fixed-wing aircraft
env = gymnasium.make("flyer-v1")In this task the aircraft must navigate to a specified 3D goal point in space in the shortest possible time.
env = gymnasium.make("trajectory-v1")
Trajectory Following
In this task the aircraft must follow a trajectory created by a moving target, maintaining on target maximizes the reward. The possible trajectory primitives are as follows:
sl, maintain straight and level flight.climb, climb to a specified level at a fixed climb angle.descend, descend to a specified level at a fixed descent angle.lt, turn left to a specified heading at a fixed rate.rt, turn right to a specified heading at a fixed rate.
env = gymnasium.make("runway-v1")
Runway Landing
In this task the aircraft needs to navigate to a runway to land.
env = gymnasium.make("forced_landing-v1")
Forced Landing
In this task the aircraft needs to find a suitable landing spot in an unprepared location.
pip install flyer-env
import gymnasium as gym
env = gym.make("flyer-v1", render_mode="human")
done = truncated = False
while not (done or truncated):
action = ... # Your agent code goes here
obs, reward, done, truncated, info = env.step(action)Read the documentation: https://aos55.github.io/FlyerEnv/
If you use the project in your work, please consider citing it with:
@misc{flyer-env,
author={Quessy, Alexander},
title={An Environment for Autonomous Fixed-Wing Guidance, Navigation and Control Tasks},
year={2023},
publisher={GitHub},
journal={GitHub repository},
howpublished={\url{https://github.com/AOS55/flyer-env}},
}