The Professor’s journey#

Duckietown was designed with multiple objectives in mind: to help professors deliver great classes, publish compelling papers, and engage their communities.

If you are reading this book, you are likely mostly motivated by the teaching use case, but a Professor often has many diverse objectives (such as teaching, research, and outreach), and aligning those objectives can be beneficial.

In this page, we provide some suggestions about how the Duckietown platform can be integrated more holistically into your lab, hopefully creating a positive feedback loop and bringing compounding value over time in the different aspects of professorial life.

To do so, consider the following “journey” (although there could be many others), that begins with gaining familiarity with the platform for the purposes of teaching but culminates in the use of the platform across a range of use cases including research and outreach.

Semester 0: Evaluate the platform#

There are many freely available resources available to get started. It’s a great idea for you and/or some people from your group to try out the software, hardware, and pedagogical materials before committing to any next step.

A great place to start could be to sign up for the Self-Driving Cars with Duckietown MOOC on EdX.

Tip

Evaluate Duckietown in simulation and on real hardware before committing.

Semester 1: Start some student projects#

A great way to get more familiar with the platform before the stress of teaching a course could be to build a small city, get a few robots, and have students do some projects using it (e.g., for Bachelor’s or Master’s theses, or supporting Ph.D. students’ research).

This serves multiple goals:

  1. Mitigate technological risk: small details (space configuration, room lighting, networks, etc.) can cause big headaches. Get these out of the way before prime time.

  2. Mitigate pedagogical risk: train a few good potential future teaching assistants (TAs) for your upcoming class.

Tip

Start small, one semester ahead of your first class. Start training your staff as future teaching assistants.

Semester 2: Teaching the first class#

Teaching your first robotics class (with Duckietown or otherwise) is fun but it can also be somewhat daunting.

Tip

The best way to mitigate stress is reducing uncertainty.

Use your trained staff as TAs#

There is great value is having teaching assistants support you in teaching a class with Duckietown, especially if they have previously gained familiarity with the hardware, software and teaching materials.

If allowed by your institution, don’t shy away from the notion of student-TAs. Yes, you can have someone at the same time taking and (help) teaching the class!

Align goals and create win-win scenarios#

Students learn best by doing. Organize students in the class in groups and propose projects for them to develop new autonomous behaviors, or expand upon existing ones. Assign each group a mentor, such as a Ph.D. student or postdoc, for best results.

Ph.D. students and Postdocs students in your lab will benefit from:

  • Gaining mentoring experience;

  • Having “free labor” (the students in the groups) to test aspects of their research. Have your (previously trained) Ph.D. students help you define the projects and mentor the student groups.

Bringing a thriving group of experts, all with aligned goals, to the class will create a positive environment for your students, who will:

  1. Enjoy a high “instructors/student” ratio for this class, which will result in better course evaluations;

  2. Have the opportunity, by excelling in the class, to get much more than just a good grade for the class. E.g., particularly successful projects could result in co-authorship in publications (led by your Ph.D. students and/or Postdocs).

Peer review and community engagement#

Although we understand that grades are necessary, there are many ways to assess the learning outcomes of your students.

After all, there are no exams “in the real world”, but there definitely is peer review.

If at all possible, we recommend organizing (well in advance) a final demonstration instead of a final exam where students showcase the outcomes of their projects to the general public.

This, in our experience, produces several positive outcomes:

  • Students gain ownership of their work, resulting in improved learning outcomes;

  • Forcing students to show their results working in the real world and explaining successes and failures to a less prepared audience is a great learning experience in itself;

  • Robots and duckies attract attention! Notify your department PR team.

Select the next generation of talent#

At the end of your class, identifying the best students and offering them to join your lab for the next semesters produces a strong line-up of candidate TAs for the next iteration of the course!

Semester 3: Expansion#

As the skills of your team increase, it becomes a good time to expand your setup in size and complexity to unlock more complex scenarios.

There are several dimensions in which this expansion can happen, e.g.:

  • Number of vehicles

  • Type of vehicles (self-driving cars, quadcopters, smart city infrastructure)

  • Size and complexity of the city topography

As the complexity grows, so do the opportunities for teaching more advanced content, and exploring different research directions.

Semester 4 and beyond: Contribute back to the community#

At this stage you will be able to develop new behaviors, create new learning experiences for students, and explore different research directions, while at the same time providing engaging opportunities for outreach.

As you grow into an expert user, we ask you to consider contributing back to the community. Examples of contributions could be:

  • Share your learning experiences for other students and instructors worldwide to use.

  • Have your students and staff provide constructive feedback to the community on how to improve the platform (many things are open source, so direct contributions are welcome too!).

  • Join and become active members of the Duckietown community. Support others on Slack and Stack Overflow, as you and your students have been supported.

  • Network with fellow instructors, set up mutually beneficial collaborations.

  • Tell us about your publications using Duckietown, let us advertise them to bring you (and us) visibility.