Application of Q-Learning Algorithm in Dribbling Ball Training of RoboCup

Abstract

RoboCup is one of the most influential Robot soccer games,and the simulation game is an important part of it. Whereas the skill of dribbling ball has a very important place in RoboCup game, by trying to introduce Q-Learning method into the training of dribbling ball, we make the agent owning the ability to learn and adaptive to the environment, so that it can gain knowledge from the environment. This paper illustrates the Q-Learning algorithm in the special area 1 vs. 1 dribbling strategy train learning method and training process, the training method is applied to everyday training and its effect was validated.