Multi Agents Cooperation Strategy for Soccer Robots

The soccer robot system and the eeective multi agents cooperation strategy which has been applied to the BALL BREAKER team is presented. The main part of the presented strategy is composed of three role task handlers, an avoidance task handler and an arbiter. The role of each robot is deened as ooense, defence and goal keeping. Each role task handler determines the target position where the robot corresponded to each role must move to. The role of each robot is determined every sampling time by the arbiter considering the information which comes from the vision system. Thus, the role of each robot is changed dynamically. The avoidance task handler checks the possibility of the unexpected collision between robots, and changes the target position to the collision free space. The trajectory to the target position where the role task handler determines is expressed by using the 4-th oder bezier curve and PID controller is used to track the determined trajectory. The presented strategy is successfully applied to the BALL BREAKER team and the team is ranked in the second place in the FIRA 98' ASIA-PACIFIC League.