Implementação de estratégias de comportamento para agentes em uma simulação 3D de futebol

Abstract

Aiming to promote research on fields such as artificial intelligence and robotics, RoboCup promotes annual robot soccer competitions. These competitions demand work on different knowledge fields, being the development of motor skills as important as the creation of behaviors capable of controlling the robots. Looking to gather experience applicable to controlling physical robots, we utilize computer simulation to create behaviors capable of being followed by a team of simulated players, capable of autonomously participating in a soccer match. Working with the SimsPark simulation platform, we have a three-dimensional environment capable of replicating an approximation of the challenges faced in the real robots competition. The behaviors were implemented by use of a framework made available by the German team MagmaOffenburg, a framework which provides the agents with basic motor skills. Three roles were defined for the creation of the behaviors, each of which adequate to the objectives identified for these roles in a full soccer match. These behaviors create a team capable of fulfilling the basic functions expected in a soccer match.