Research on action select strategy of robot soccer based on Petri net and position prediction

This paper proposes a robot soccer action select strategy based on Petri net and position prediction. It predicts the positions of robots and the ball according to information of newly-generated frames from the vision system. Then the paper proposes a model of action select strategy of goalkeeper based on the results of data preprocess. The method is modeled, analyzed and simulated by Petri net and the properties of boundness, safety and reversibility are proved to show the correctness of the model. The framework of realization of the system is also given and it has been tested by program. The results show the practicability and validity of the method.