Integrating Simulated Exogenous Environments to Support the Learning Process of the Embedded MAS Approach

A multiagent system (MAS) is a group of autonomous software entities capable of collaboration and cooperation. It is a common practice to use simulators in MAS development. However, in the case of Embedded MAS education, integrating simulators is difficult because the agent’s reasoning is tightly coupled to hardware communication interfaces. This work then presents a simulator integration methodology based on serial channel emulation that fully decouples the agent’s reasoning from the environment, allowing flexible integration of simulators. We present a report on an Embedded AI undergraduate course in two academic semesters, showing that the student’s projects completion rate increased from 33% to 100%.