Develop simulation and trial test Q-Learning system for soccer robot system

Purpose of my research is to make an automatic control system for soccer robots by the reinforcement learning . We would like to use the radio controller soccer robots for our research. We have a program agent for trying to movement many times in the simple soccer problem. But this simulation takes very long time for learning by Q-learning . In this paper, we propose a small simulation field and try to reinforcement learning. Finally, we would like to control soccer robots by this result.