Physical Robots for Teaching Mobility & Manipulation using ROS in Remote Learning

Even though remote learning has been present and available in a myriad of topics before pandemic times, robotics remote learning had the limitation of interacting with robotic platforms through simulation alone. With COVID-19, postgraduate education was forced to move to remote learning. Birk et al. conducted a reasonable practice for online teaching of a robotics course at Jacobs University Bremen. Although their lectures covered most of the robotics areas, they used the pre-recorded videos to teach and did not hold labs to demonstrate the operations on real robots. The pivot created a paradigm shift for robotics courses traditionally taught in-person where students had the opportunity to experience interaction with robotic platforms in simulation and with the physical platforms. To address the learning challenges and with remote robotics, we present two case studies where physical robots were used to teach concepts in navigation and manipulation. This included instances of long-distance robotic teleoperation as well as autonomous behaviors. The Robot Operating System (ROS) plus Gazebo, and RViZ were used for teleoperation and autonomous routines with the Turtlebot3 and Kinova Gen3 lite robot platforms. Our results demonstrate how concepts in robots such as SLAM can be taught in both simulation and with physical robots plus how students learn how to mitigate the errors and uncertainty produced in the real world with both mobile and manipulator robots. Furthermore, the graduate-level courses were designed to be accessible to students from interdisciplinary backgrounds such as human-centered design engineering, architecture, and engineering. The benefit of this design is to enable students from interdisciplinary backgrounds to learn how to make both technical and design contributions to robotics projects that can lead to a more holistic, user-centered approach to robotics.