Research And Exploitation Of RoboCup Robot Decision-Making System

In recent years, robot soccer competition is developed rapidly which is a High-Tech activity. Soccer robot relates to many fields and it provides a suitable experimental platform for the research on single agent, multi robots and other related fields.Robot soccer system of RoboCup includes four components such as robot, communication subsystem, decision-making subsystem and vision subsystem. In order to win in game of soccer robot, each component of soccer robot system need be very good capability. Our key research is decision-making subsystem, and we introduce about level structure of single soccer robot .Besides, we also point out that robot could study independently and cooperate with each other. In order to solve these questions, we add to learning model base on three-layered structure.Effective shooting is the ultimate goal of designing decision-making subsystem. In the competition, robots will be hindered by the enemy robot, so it is necessary to avoid enemy robots. At present, better method of avoidance is Limit-Cycle which can avoid the movement of robot to reach the target, but this method is based on static ball. Actually, the ball often collides with wall or robots so that bear of ball can not be determined. In order to solve this problem, this paper presents LC-LWR method which can forecast the ball trajectory and determine the location of ball in a time. So robot can avoid obstacles to find ball.The simulation results show that LC-LWR method is effective.This research tests on the robot platform of North China University of Technology.