Design of an experimental platform for the automation of the goalkeeper of a foosball table

摘要

This paper describes the process of designing, deployment, and testing of an experimental platform whose goal is to automate the movement of the goalkeeper of a foosball table. The platform imposes several challenges that span different disciplines of Electronics Engineering, including computer vision techniques for detecting and tracking the position of the ball and players in the table, estimation algorithms for predicting the movement of the ball, and control algorithms for moving the rod that holds the player. Besides, the characteristics of the application impose stringent requirements in terms of precision and processing speed, which requires an adequate choice of the software and hardware for the communication, storing, and processing of data in real-time. The developed platform will serve in teaching and research activities in the context of modern cyber-physical systems, including the experimental characterization and validation of machine learning, control, estimation and computer vision algorithms, and the integration and evaluation of specialized computing platforms for accelerating computationally intensive processes. Moreover, because of the popularity of foosball tables, this experimental setup serves as an ideal platform for outreach activities related to STEM in non-academic audiences.