CS 6756 (Robot Learning for Decision-Making) Final Project. This project explores the utilization of deep reinforcement learning to train a vision-aided fixed-wing aircraft autopilot in AirSim (with the JSBSim flight dynamics model) for waypoint navigation and obstacle avoidance.
The integration between AirSim and JSBSim and some other setup is adapted from the Fixedwing-Airsim repository cited below:
@misc{FW-AirSim,
author = {Quessy, Alexander},
title = {Fixedwing-Airsim},
year = {2021},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/AOS55/Fixedwing-Airsim}}
}