Learning distributed algorithms by programming robots

Abstract

The learning process of theoretical concepts such as the model of a distributed environment and different distributed algorithms together with their execution and correctness requires time and is often considered by students a hard and non-challenging issue. In this paper we suggest adopting a more practical approach based on real implementations of distributed algorithms with the help of robots. A learning-by-doing approach can, in our opinion, help students acquiring a deeper knowledge of the model and of the algorithms, and can also stimulate them, and let them improve their teamwork skills. In this paper, we present a specific case study of a practical project, run for two consecutive years at the University Cà Foscari of Venice, inside an International Master of Computer Science course of Advanced Algorithms. The students for their final exam had to work in groups and their task was to design and implement a distributed algorithm to solve an assigned problem, using a Lego Mindstorm EV3 robot and a Makeblock mBot robot. In this paper, we discuss the positive effects of such a non-traditional teamwork approach by analyzing the teacher’s perception, the feasible impact on the students’ grades, and the students’ involvement and positive feeling, highlighted by the results of some questionnaires proposed at the beginning and the end of the projects. We finally discuss the limits of such an approach and possible improvements.