Robot Learning. 3 credits

Catalog description

As robotics becomes integral to our daily lives, demand for robots capable of tackling more complex tasks is rapidly increasing. To produce more capable robots, solely relying on manually designed robot behavior and control solutions is very limiting, and learning becomes very attractive. This course focuses on robot learning, working at the intersection of robotics, control, machine learning, and cognitive sciences, where physical dynamics, uncertainty, and limited data impose strong constraints on learning systems. After a brief introduction to robotics and machine learning basics, the course covers optimal control, model learning, reinforcement learning, imitation learning, unsupervised and self-supervised learning, and foundation models in robotics. It discusses the foundational aspects of these topics and reviews state-of-the-art approaches. The course is hands-on and project-focused. Students implement several of the introduced methods and perform semester-long research projects. Offered by Electrical & Comp. Engineering . May not be repeated for credit.