CBRRoboSoc: An Efficient Planning Strategy for Robotic Soccer Using Case Based Reasoning

Abstract

Over the past decade the field of robotic soccer has been the focus of intensive research. This paper proposes an efficient planning strategy for robotic soccer using case based reasoning (CBRRoboSoc) to build prototypes for a team of two soccer playing robots in the RoboCup small size league. The proposed CBRRoboSoc system develops an efficient strategy to plan both offense and defense team behavior. The system along with its field simulation, prediction and command generation phases, employing Bayesian classifier, help the robots to plan individual moves, recognize game states, as well as to model the world of the playing field. Our model supports game play selection in key game situations, which should in turn significantly advantage the team. Further more by incorporating case base for offense team, simulation of four keep away trials illustrate optimized offense behavior.