Reinforcement Learning applied on a Tello Drone to follow human faces using keras-rl. For this project DJI Tello Edu drone was used.
Agent is trained on a continuous state simulator to adjust velocity depending upon the position of the target using Deep Deterministic Policy Gradient (DDPG)
Click below image to redirect to youtube demo video
Clone or download the project to your system
Python
Install Python 3 or Anaconda
Install Dependencies
pip install -r requirements.txt
Train the agent on a simulator and save weights
python train.py
Load the saved weights and test it on the drone
python drone_test.py