5)x]]\ Dq Hdv\ Dqg H[suhvvlyh Wrro Iru Prghoolqj Wkh Frjqlwlyh Od\hu Lq 5rer&xs6rffhu

The idea of robot playing soccer has been developed since early 90s Chen et al. (2003). Soccer environment is a dynamically changing environment which requires individual skill as well as team skill and therefore is an interesting research field on Artificial Intelligence and robotics. Prolog is a programming language that represent logic reasoning. Is is a perfect tool to represent human reasoning, so it seems to be a good choice for implementing the cognitive layer of soccer players that is a simulation of human behaviour related to this game. For example, applying the rule “if the goal keeper is not at the goal then kick to ball”. But many of the most important decisions that are made by soccer players deal with non-crisp issues. They are related to fuzziness (e.g. “if other player of my team is FAR fromme then don’t pass him/her the ball”), uncertainty (e.g. “if I CAN get the goal then kick the ball”), or incompleteness (e.g. “if I cannot see the position of a player, by default I’m not going to pass him the ball”). In this workwe are going to provide a programming framework to Robot Soccer programmers to model robot control in an expressive but simple way. We propose the possibility of using fuzzy concepts for this modelization and we are going to provide some conclusions about this tool based on a bench of practical experiments that we have done and that we describe here. In the rest of this section we introduce RoboCupSoccer field (section 1.1) and we discuss some previous fuzzy approaches in logic programming (sections 1.2 and 1.3. In section 2 we describe our framework, RFuzzy enumerating the features that characterize its expressivity for modelling problems in general. From the following section we focus on the Robot Soccer use of our tool. Section 3 describes the environment for our experimentation (the general arquitecture at section 3.1 and the particular Prolog code arquitecture at section 3.2). We have described in detail our experiments in section 4. We provide information about the decision making analysis that we have done (section 4.1.1) and the action execution analysis (section 4.1.2). We finally conclude at section 5.