The Olin College course “A Computational Introduction to Robotics” (CompRobo) serves as a tour through some of the most important ideas at the heart of modern robotics. The course utilizes a project-based learning pedagogy that allows students to build mastery of key concepts while also allowing for a great deal of student choice and autonomy. The major focal points of the course are state estimation, localization, computer vision, decision-making, and societal implications of embodied systems.

Robot Details and Documentation

A picture of a Neato robotic vacuum with a custom remote control interface based on Raspberry Pi

The documentation describes both how to connect to the the physical robot or a simulator and how to build your own customized Neato.

Student Facing Documentation

Teaching Team Documentation

RoboBehaviors and Finite State Machines Project

The Neato robot in a simulated world tracking a barrel

The first project, RoboBehaviors and FSMs (finite state machines) provides a scaffolded assignment for students to get up to speed with important concepts in ROS through implementing compelling behaviors on a robot. The project emphasizes the establishment of good practices such as debugging techniques and visualization.

Supporting Documents

Robot Localization Project

The Neato robot in the simulated Guantlet world with a small number of laser scans collected to localize.

The Robot localization project is a scaffolded assignment for students to learn about the particle filter algorithm. Along the way they will learn some basics of Bayesian inference and some new ROS tools and workflows.

Supporting Documents

The Broader Impacts of Robots

A collage of robots including AUV Sentry, Asimo, Waymo car, a coffee maker, factory assembly arms, and a Boston Dynamics Spot.

By virtue of being embodied, robots can literally change the world. Daily in-class activities and a 3-part assignment examine the impacts and implications of robotic systems and algorithmic choices made in their design.

Supporting Documents

Machine Vision Project

An example of a keypoint matching algorithm working on indoor images

The machine vision project is an open-ended project on using computer vision in the context of robotics.

Supporting Documents

Final Project

A series of 3 images, moving clockwise starting on the top left, two Neatos coordinating paths around one another, an RViz diagram for a Neato moving quickly through unknown space, and a 6DOF manipulator playing chess.

The final project is an open-ended project that lets students explore a robotics topic and algorithms in depth.

Supporting Documents

In-class Activities

Note: see Site-wide TOC for an easy to navigate outline of each day’s activities Note: Subject to change as the semester unfolds!

RoboBehaviors and Finite State Machines (+ Broader Impacts of Robots Discussions)

State Estimation and Localization (+ Robot Application Contexts Discussions)

Machine Vision (+ Influences on Robotics Development Discussions)

Final Project (+ Implications of Robots Discussions)

Bonus Materials (e.g., Recitations)

Conclusion and Learning More

CompRobo serves as a fun, hands-on introduction to key ideas in robotics algorithms and toolsets. Despite the fact that the course is successful at Olin, we realize that everyone’s institutional context is different. To connect with folks at Olin College to learn more about this module or determine how you might build off of this at your own institution, e-mail Olin’s External Programs and Partnerships to start the conversation.