Non-communication Cooperation Strategy Between Soccer Robots

Communication-less cooperation strategy between soccer robots is researched in this thesis,and a variety of methods were proposed to solve simple or even complex cooperation between two or three robots.An engineering method is proposed to deal with the cooperation between real robots,and the game theory is also applied to solve strategy selection problem in the complex cooperation between more robots.To deal with the simplest single pass-receive cooperation between two robots,a method derived from the traditional SetPlay is generalized to specific SetPlay without communication,and public information is used to calculate the optimal receiving point to ensure the consistency of the receiving point.The consistency of the field information that collected by each robot and the rationality of the optimal receiving point are analyzed.The experimental results show that the consistency of the receiving point obtained by this method is limited by the speed and the distance between the cooperating robots,and the calculated receiving point,which is unreasonable due to the lack of the consideration of the current position of the ball receiver.At the same time this method cannot provide an effective basis to choose the passing opportunity.So this method is simple but defect significant.In order to solve the deficiency of without using the current position of the ball receiver,a color recognition method based on the front camera is proposed recognize the ball receiver.The color card on the robot body is marked as the different characteristics to distinguish from the black obstacles,and we can calculate the rough orientation of the receiver after recognizing the color card using the front camera.In this way,the receiving point is the current position of the receiver robot,so it can avoid the problem that the receiver cannot reach the receiving point in time caused by unreasonable passing opportunity.At the same time,the heuristic function is utilized to derive the choice of the optimal passing point,so the successful rate of the ball passing and interception cooperation with the interference of obstacles can be improved.The proposed method is implemented in the technical challenge of RoboCup MSL and performance well.Finally,the number of cooperation robots is extended to three,and the problem of cooperative strategy selection for multi-robots without communication is explored using the game theory.On the basis of the traditional incomplete information game theory,the establishment method of joint probability distribution in terms of the distance and the joint probability distribution correction method based on the velocity are proposed,and the strategy selection of the ball passer and receiver is solved.The experimental results show that the incomplete information game model can be used to realized the cooperative strategy problem without communication with a high successful rate and avoid the conflict between the robots.However,this method requires a relatively high accuracy of the environmental perception of the robot,so there are still some difficulties to use it to the actual complex and dynamic RoboCup MSL environment.In summary thesis,we propose a variety of methods to address the problem of multi-robot cooperative strategy without communication,and analyze the application scope and effect of the proposed methods.Some engineering methods are proposed to deal with the actual ball passing and interception problem,and game theory is also tried.