The Confrontation Strategy And Multi-robot Cooperation Of Robot Soccer Based On Game Theory

Soccer robot system is a cross-fields science, which covers robotics, data fusion, intelligent control, image processing, communication, mechanical-electrical integration etc. Meanwhile, soccer robot system provides a standard test-bed for the research of multi-agent systems (MAS) and artificial intelligent. In robot-soccer game, how to choose the right strategy and the mutual coordination among robots in the complex and dynamic environment is the key to the decision-making system of robot. From the two key problems of the decision-making system, the confrontation strategy and the cooperation among robots are researched according to the requests of intelligence and accuracy of the decision-making system and combining the game theory in this paper.The decision-making system of robot and the algorithm of game theory are analyzed and studied relying on the soccer robot platform which is provided by the discipline groups of xi'an university of science and technology. From the angle of opposability, the confrontation strategy based on incomplete information game is proposed. Then, from the angle of cooperation, the cooperative game used in coordination and cooperation among robots is discussed. Finally, the two game models are analyzed and solved. Through the simulation results, it can be see that the incomplete information game algorithm can reduce considerably the effective strategy scope, and the cooperative game algorithm used in coordination and cooperation among can make the robot take turns with the change of situation dynamically.Owing to the instability of holding device of four-wheel robot, the holding device of soccer robot is improved. The auxiliary devices with a core of two direct current motors whose power is 12W are added to enhance the stability of holding ball and increased the probability of winning the game.Lastly, it is verified that the above-mentioned methods boosted obviously the competitive level of the whole robot team, as well as improved the holding probability and the mutual cooperation among robots.