3.0

Average Course Rating

This course provides a comprehensive, hands-on introduction to the design and implementation of modern learning-based robotic systems. The primary objective of this course is to equip students with the theoretical skills and practical tools necessary to succeed as world-class robot learning engineers and scientists. This course involves equal amounts of in-classroom teaching and outside-the-classroom practical work. Students will work in small groups (approx 3 students per group) to build a physical SO-101 arm from a provided equipment kit, and develop key software stack components to enable robot learning from data. The first two modules of the course will cover the ‘nuts and bolts’ of robot learning and data curation strategies, including a variety of teleoperation methods. The remainder of the course will then delve into learning ‘policies’ that determine how a robot acts in response to sense data. The course will cover state-of-the art policy learning approaches, including generalist techniques such as vision-language-action (VLA) models, in detail. Required Course Background: machine learning or deep learning, and robotics. Students may receive credit for EN.601.498 or EN.601.698, but not both.