Duckietown - AI Driving Olympics

Expert	Trained Duckiebot

The “AI Driving Olympics” (AI-DO) is a competition with the objective of evaluating the state of the art in machine learning and artificial intelligence for mobile robotics. The goal of the competition is to build a machine learning (ML) model that allows a self-driving "car", called Duckiebot, to drive on streets within Duckietown.

The AI Driving Olympics competition is structured into the following separate challenges:

• Lane Following - LF
• Lane Following with Vehicles - LFV
• Lane following with Vehicles and Intersections- LFVI
• Autonomous Mobility-on-Demand - AMoD

This project is a solution for the aido-LF challenge:

• Control of a Duckiebot to drive on the right lane on streets within Duckietown without other moving Duckiebots present.

More info about the Duckietown Project and aido-LF challenge here.

Table of contents

• Description
• Prerequisite
• Getting Started
  • Install
  • Run
• Author

Description

The approach is to train a Reinforcement Learning (RL) agent to build an expert that drives perfectly within an environment, then use this expert to collect data. Data are pairs <observation, actions>collected in different maps/environments used to train an agent that imitates the expert's behaviour (Imitation Learning / Behavioural Cloning) . Finally, you have a self-driving car that navigates within Duckietown using only one single sensor, the camera.

Trough this tutorial you will:

• Use the Reinforcement Learning approach (Decision Tree) to build an expert.
• Use the expert to collect data.
• Use the Imitation Learning approach to train the Duckiebot.
• Submit the solution to the “AI Driving Olympics” (AI-DO).

Prerequisite

It's highly recommended to create a virtual environment using virtualenv or anaconda

Before proceeding:

• Make sure you have pip installed.
• This project is tested on Python 3.7.
• Install the duckietown-shell following the official guide.

Getting Started

Install The Environment

Clone this repository:

git clone https://github.com/FaMoSi/Duckietown-AI-Driving-Olympics.git

Change into it:

cd Duckietown-AI-Driving-Olympics

Install all the dependencies using the requirements.txt file:

pip install -r requirements.txt

That's all, everything is installed and you can run all the scripts in this project!

Run

• RL Expert: follow the Reinforcement Learning instructions.
• IL Duckiebot: follow the Imitation Learning instructions.

Author

• Simone Faggi
• Email: simone.faggi@yahoo.it