Duckietown

What is Duckietown?

Duckietown started as a class at MIT in 2016 and all of the pieces of the experience (the slides, the demos, the platform, the software) were made openly available and other institutions began to take interest. The platform has since been used at several universities around the globe, including NCTU, Tsinghua, ETH, Université de Montréal and many others. Now it is a worldwide initiative to realize a new vision for AI and robotics education. In 2018 the Duckietown Foundation, a non-profit foundation, was born, and its aim is to develope and promote the Duckietown project.

AI Driving Olympics:

The Duckietown Foundation also hosts the AI Driving Olympics (AI-DO) twice per year, at the famous machine learning conferences ICRA and NeurIPS. The AI-DO is a set of competition with the objective of evaluating the state of the art for ML/AI for embodied intelligence. The main objective of our team is to take part in the next AIDO competition!

Collaboration with PMDS:

We asked the Polimi Data Scientists to join forces and work together to create an heterogeneous team composed of both Computer Science and Automation Egineers. We think that this is a great opportunity for all the students who want to face a new challenge and learn new things!

Our project:

We are looking for people interested in the world of autonomous driving, robotics and AI, in order to create a team capable of working with the Duckietown Platform and facing challenges (like the next AIDO competiotion)! The purpose is to teach a small robotic car to navigate on a track, which is no easy task. The main objective of the team is to take part in the AIDO competition.

If you are interested in autonomous driving and robotics, please fill the forms below and you might get selected to be part of the team! The team will work in a Linux enviroment, using mainly technologies like Python, ROS, Docker and Reinforcement Learning.

Here you will find usefull resources made by our team to help you get started with autonomous driving and reinforcement learning with python.

Setting up your headless Duckietown environment in Google Colab

Using the given example

Make an autonomous agent with Reinforcement Learning using an example from the Duckietown repo.