This Repository is an Example of using Airsim with Deep Q Learning for Reinforcement learning.
Requirements:
AirSim-1.2.
UnrealEngine-4.18.
Tensorflow.
Using learned weights:
Use main-deploy.py for testing the Agent. The agent tries to find a fixed goal (white cone) from a fixed staring point.
you need first to download the custom airsim blocks environment from here :
https://drive.google.com/open?id=1aBkrsqiG1t4KJbSlHjhvoowYPVWOKfB6
Training from scratch:
Use main.py for training from scratch.