Rastreamento de atletas em partidas de futebol utilizando filtragem estocástica

Abstract

Brazil is considered by many the country of football, due to the passion of the Brazilians for the sport and the production of players recognized worldwide. Sport, however, is constantly evolving, and thereat, the use of technology becomes much more common. Two factors may be related to this implementation of the technology: human errors during soccer matches, and the need for tactical evolution of the clubs. Thinking about this, the interest in joining sport and technology has gradually increased in the sports field. This work aims to promote the integration of engineering, mainly the control theory, with sport. Using MATLAB tsoftware, one of the proposals was to implement a player trajectory generator applied to soccer matches. The program simulates possible cases of play, which the athlete goes through a certain path, diverting from opponents in different formations. Therefore, it is possible to define tactical trainings, demonstrating to the athletes possible plays that can occur during the match, and, allowing to verify the energetic expenditure of each athlete, how much it goes during the game and its average speed. Another proposal was to generate noisy measures to be used in stochastic filters. Such filters have the function of reducing or eliminating random noise and other uncertainty of measurements of magnitudes carried out over time. With the simulator of trajectories and the filters implemented computationally, performance tests were made in relation to the estimates obtained by the filters.