It is a big question when will be the people able to design and implement a robot with that much intelligence, that the robot can solve such hard intellectual problems as the people can solve. RoboCup is organizing a competition every year, where people from all around the world, can challenge the others with their own made robot soccer teams. In the 2D league the main emphasize is on developing more and more precise strategies and tactics. Obviously the results achieved in this league can be easily used in the competition of the human size robots, when the robots will be able to move as dynamically as the game requires them to do.
In my work I designed the fuzzy strategy of a 2D robot soccer team in MATLAB environment, which is using the official RoboCup server. In the 2D league the game is divided into turns, all the player get information from the server about the game situation, in the beginning of their turns, and they have to make the best decision for the team, and perform it.
My strategy is based on basic conclusions from the game situation, and fuzzy decisions. During designing my team strategy I looked at examples from real life, and I used them to determine the decisions, the reactions, the behavior model of the agents, and to design the membership functions.
I wanted to define the membership functions, that they are containing all the important parts of football, however they are not too much complicated. All the players in their turns should be able to choose the best strategic decision, and to perform it.
Finally I attached the strategic level to a tactical one, so that is how the program is sending basic commands to the Java client written by Tibor Oláh , which commands are easily forwarded by the Java client to the official server. That is how my strategy can be tested, and displayed by the official RoboCup monitor.