Project-based learning course design for multi-agent autonomy using quadrotors

Abstract

This paper proposes a project-based learning course utilising multiple quadrotors, aiming to solidify and amplify technical knowledge and develop critical thinking capabilities. An engineering problem is provided as a surveillance mission with multiple quadrotors autonomously searching, detecting, and tracking ground vehicles. The course covers all stages of multi-agent autonomy development, from identifying system requirements, designing software and hardware, to conducting demonstrations in a drone flying arena. During the course, students improve their ability to critically formulate, solve and evaluate engineering problems as well as gain and apply technical knowledge in all aspects of autonomy such as flight dynamics, control, navigation, guidance, task allocation, situational awareness and communication. The paper details the problem design, course timeline, outcomes, and key lessons learnt from the course.