Research on RoboCup Simulation 2D Shooting Strategy Based on Digital Twin Framework

Abstract

RoboCup, as one of the main platforms for studying artificial intelligence theory, is an international academic competition. The shooting strategy is one of the most important skills in the RoboCup simulation 2D competition, which is one of the important branches of RoboCup. In the digital twin framework, this paper focuses on extracted game data from specific teams and establishes portraits of the players’ shooting strategies by factor analysis method. The main factor that determines the impact of the shooting strategy from the portraits is the shooting distance feature. Then, by optimizing the shooting strategy of YuShan2019 based on the feedback on the difference of the shooting strategy portrait of each team, excellent performance can be achieved in the actual competition, which indicates its effectiveness.