Teaching by Practice the Basis of Consensus for Multi-Agent Systems

Abstract

This paper provides in depth feedback on a new course on consensus for multi-agent systems developed for graduate students. The considered pedagogic approach gives a simple theoretical basis to the students through very few lectures while putting a strong effort into practice and applications. The paper describes how the transition to full remote teaching has been achieved due to the Covid-19 pandemic, especially regarding practical work sessions. Initially planed with aerial robots, practical work sessions have been modified to use a remote online platform for multi-robot experiments, Robotarium from Georgia Tech. To keep lectures and practical work sessions attractive to the students, especially in this remote learning context, a motivating example has been proposed to students, mixing arts and sciences, and consisting in re-creating a choreography from a kite contest video with a multi-robot system.