Position Prediction of Opponent Players by SIRMs Fuzzy Models for RoboCup Soccer 2D Simulation

Abstract

In RoboCup soccer, it is important to appropriately react to the movement of the opponent players. Particularly in the situation of one-to-one defense, predicting how the opponent player moves necessary because the player facing the opponent player is the only one who handles the defense situation. In this paper, we propose a prediction technique of the position of the opponent player in the one-to-one defense situation using SIRMs fuzzy models in the RoboCup soccer simulation 2D league. By conducting a series of computational experiments, we investigate the prediction accuracy of the SIRMs fuzzy models that are trained using a training set. The training set contains the input field state vector along with its corresponding target signal. The trained SIRMs fuzzy models are used in one-to-one defense during the soccer games.