Publication: AUTOMATIZACIÓN DE UNA ‘TABLE FOOTBALL’ MEDIANTE PROCESAMIENTO DE IMÁGENES Y ALGORITMOS NEURO-DIFUSOS
No Thumbnail Available
Date
2018
Authors
ATUAN TOLEDO, RICHARD MICHAEL
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
El estudio que se presenta a continuación, corresponde al primer paso de un proyecto quepretende automatizar el funcionamiento en uno de los equipos de una "Table Football", teniendocomo motivación la creación de una plataforma de desarrollo experimental, donde sepuedan implementar y probar algoritmos o tecnologías con fines académicos y pedagógicos.El presente trabajo buscar sistematizar, en un ambiente simulado, el movimiento de uno delos arqueros con el fin de bloquear los tiros rivales a portería.En esta primera etapa del proyecto, se implementa un algoritmo computacional que logra,por medio de una simulación, estimar la posición del arquero dentro del juego. Mediante procesamientode imágenes se obtiene el lugar donde se encuentra la pelota en juego, mientrasque un algoritmo Neuro-Difuso (ANFIS) permite estimar dónde se dirige esta, estableciendoasí, la posición donde el arquero debería ubicarse para bloquear el paso de la pelota. Además,se simula el movimiento del portero en cuestión mediante un motor DC cuya posiciónse controla a través de un controlador PID. Cabe destacar que al ser una primera etapa, laestructura física del proyecto se encuentra en desarrollo, por lo que se opta por un trabajo desimulación con datos teóricos (motor DC) y prácticos (Imágenes/Vídeos).En este trabajo se plantean soluciones a problemas tales como, la obtención de imágenesy el reconocimiento de la pelota dentro de ellas, predicción de la posición futura del balón yel control del movimiento en un motor DC, especificando las consideraciones técnicas tantopara la obtención de imágenes como para su implementación. Finalmente, se integran dichassoluciones propuestas en un ambiente simulado desarrollado para estudiar el comportamientodel arquero. Esto permitió evaluar, vía simulación, la utilidad de los algoritmos propuestospara ser usados en el sistema real.
The study presented in this document, corresponds to the first step of a project that aimsto automate the operation on one of the teams of a "Table Football", having as a motivationthe creation of a plataform for experimental development, where you can implement andtest algorithms or technologies with academic and pedagogical purposes. This work seeks tosystematize, in a simulated environment, the movement of the goalkeeper, in order to blockrivals shots to the goal.In this first stage of the project, a computational algorithm is implemented that achieves,through a simulation, estimate the position of the goalkeeper within the game. By imageprocessing is obtained the place on the field where the ball is, while a Neuro-Diffuse algorithm(ANFIS) allows to know where the ball is going, establishing the position where thegoalkeeper should be placed. In addition, the movement of the goalkeeper is simulated by aDC motor whose position is controlled by a PID controller. It should be noted that, as it isa first stage, the physical structure of the project is in development, so the study opted for asimulation work with theoretical data (DC motor) and practical data (Images / Videos).In this work there are solutions to problems such as, image acquisition and identificationof the ball inside these, prediction of the future position of the ball and control of movement ina DC motor are proposed, specifying the technical considerations both for obtaining imagesand for their implementation. Finally, these proposed solutions are integrated into a simulatedenvironment developed to study the behavior of the goalkeeper. This allowed to evaluate, viasimulation, the usefulness of the algorithms proposed to be used in the real system.
The study presented in this document, corresponds to the first step of a project that aimsto automate the operation on one of the teams of a "Table Football", having as a motivationthe creation of a plataform for experimental development, where you can implement andtest algorithms or technologies with academic and pedagogical purposes. This work seeks tosystematize, in a simulated environment, the movement of the goalkeeper, in order to blockrivals shots to the goal.In this first stage of the project, a computational algorithm is implemented that achieves,through a simulation, estimate the position of the goalkeeper within the game. By imageprocessing is obtained the place on the field where the ball is, while a Neuro-Diffuse algorithm(ANFIS) allows to know where the ball is going, establishing the position where thegoalkeeper should be placed. In addition, the movement of the goalkeeper is simulated by aDC motor whose position is controlled by a PID controller. It should be noted that, as it isa first stage, the physical structure of the project is in development, so the study opted for asimulation work with theoretical data (DC motor) and practical data (Images / Videos).In this work there are solutions to problems such as, image acquisition and identificationof the ball inside these, prediction of the future position of the ball and control of movement ina DC motor are proposed, specifying the technical considerations both for obtaining imagesand for their implementation. Finally, these proposed solutions are integrated into a simulatedenvironment developed to study the behavior of the goalkeeper. This allowed to evaluate, viasimulation, the usefulness of the algorithms proposed to be used in the real system.
Description
Catalogado desde la version PDF de la tesis.
Keywords
NEURO-DIFUSO , PROCESAMIENTO DE IMAGENES , TABLE FOOTBALL , TACA-TACA