This is the webpage for the 2020 Autonomous Vehicles class at l’Université de Montréal which is informally known as “Duckietown”.

For students interested in taking the class

I wish I could take everyone, but due to space and other considerations there are only 18 spots available for the class this year. If you would like to be one of them, please fill out the application form. Please include anything that might help you stand out. This class is a collaborative learning experience and I’m looking for a good mix of student backgrounds.

I invite all to come to the first class which will be held virtually on zoom on Wed Sept. 2 @ 11:30

Decisions about admissions will be made following the first class and the successful applicatants will be notified by Friday Sept. 4 by email.

Once you have been admitted you should contact Céline Bégin, she will have the list of admitted students. If you are worried that you may not be admitted please register for a backup course.

There are no “prerequisites” for the course but a strong programming background, as well as any knowledge of basic probabilities, computer vision, or control will all help.

If you are not sure whether or not you would like to take the class I would encourage you to:

The AI Driving Olympics Competition

Again this year we are running the class in conjunction with the “AI Driving Olympics” competition at NeurIPS 2020. Submitting entries to the competition will be an element of the class. Last one of the students from the class won the “LFV” challenge.

What makes this a special class?

This class is unique in many ways.

Class philosophy

The best engineers and scientists are the ones who have solid theoretical foundations, as well as practical experience in the domain of interest.

In autonomous robotics, it is important to get the “feeling” of what actually makes a robot work, how the success or failure depends on subtle interaction between many hardware and software components.

To this end, it is necessary to study a complete system like Duckietown. The materials might be cheap, the appearance might be playful, but the complexity of behaviors and representations is comparable to those of deployed robotic systems.

This class is a collaborative learning experience about robotics. The fundamental theme of the class is that embodied systems are a special brand of AI system that has special real-world constraints. The only way to really feel those constraints is to experiment with the physical system. Clearly, this is a little more complicated given the COVID situation, but we have been preparing for this…

A Personalized Experience

Each student gets their own personal Duckiebot to build and love. This will be a little more logistically complicated this year, but we will figure it out!

A Global Class

This class is offered at the same time at several other institutions, including:

  • Toyota Technological Institute at Chicago and the University of Chicago, led by Prof. Matthew Walter.

and several others… You can join the global slack community to see what’s going on around the world.

Your code can live on

This class is a living thing, it will never be the same twice. The best projects from each year will be added to or replace existing code in the duckietown codebase. If you do a good job, your project will be what next year’s students use and try and improve upon.