Quadcopter Project

This project is an exercise in reinforcement learning as part of the Machine Learning Engineer Nanodegree from Udacity. The idea behind this project is to teach a simulated quadcopter how to perform some activities. I have chosen to teach it two tasks: Take off and maintain position (or hover).

Concept

The agent is based on the theory in the Deep Deterministic Policy Gradient (DDPG), specifically the concepts from this paper: https://arxiv.org/abs/1509.02971. This method is a special variant of Actor-Critic learning.

Running the code

To run the code, you can use anaconda. Create an environment as follows:

conda create --name quadcopter python=3

To use the environment, execute the following:

For mac / linux:

source activate quadcopter

For windows:

activate quadcopter

Afterwards, install the requirements as follows:

conda install numpy matplotlib jupyter notebook

Notes

The base code provided for this project is taken from the Nanodegree sessions.