Visão computacional aplicada ao monitoramento de robôs móveis em cenários de robótica educacional

Abstract

A frequent concern among students and teachers working with educational robotics is about the results gathered through practical activities. Generally, robotic scenarios are made of tasks that challenge students' knowledge, making them to setup a moving robot in order to solve certain problems. Currently, educational institutions and robotics competitions lack of mechanisms that can evaluate the robot’s behavior during the accomplishment of the proposed challenges. In this work, we propose the development of a technological solution using Computational Vision, whose purpose is to monitor and evaluate the performance of mobile robots during the execution of Robotics Pedagogic tasks. The developed system, named MonitoRE - Monitoring System for Educational Robotics, was tested on three categories of pedagogic robotics environments, obtaining promising results. Our tool uses an Absolute Location Method with descriptors based on color and shape to analyze the task environments, mapping the path taken by the robot, evaluating the achievements in the proposed tasks. The experiments carried out indicate that the adopted method is effective, performing satisfactory results in robotic monitoring. In addition, it was found that teachers and students felt more motivated, demonstrating interest in using monitored task environments, because it ease the understanding of the difficulties faced by the moving robot in completing the activities, assisting students in the teaching-learning process.