Teaching
Physical AI course, ROS2 part

Physical AI course, ROS2 part

Matsuo-Iwasawa Lab provides several deep learning-related courses for all students even outside of the University of Tokyo. I am one of the lecturers for the Physical AI course, which is a course for students to learn the fundamentals of Robotics and Robot Learning from both theoretical and system engineering perspectives.

I was in charge of the ROS2 part where I started from the very basic knowledge of ROS2 and built up to the actual imitation learning pipeline, enabling students to experience the data collection with ROS2, converting asynchronous rosbag data to synchronous data at a target frequency (rosbag2lerobot), training the policy using LeRobot, and deploying the policy in Gazebo.

I made the materials publically available (Japanese):

Please refer to the course website for more details (Japanese): Physical AI