Autonomous Navigation of drones using Q-Learning

This project implements autonomous navigation of an ARDrone in Gazebo simulator. Q-Learning algorithm was used in a simple gridworld setting. The state space included the current coordinates of the drone. A reward of -1 was given upon not ending up in the goal state and a reward of +100 was awarded for reaching to the goal.In the demonstration below, white spot shows the starting point and the red spot is the goal.

[code]

Q Learning

Autonomously navigating to the goal

Reference paper : Pham, Huy X., Hung M. La, David Feil-Seifer, and Luan V. Nguyen. "Autonomous UAV Navigation Using Reinforcement Learning." arXiv preprint arXiv:1801.05086 (2018).