Study On Team Defense Strategy In RoboCup 2D Based On Density Clustering Algorithm

With the advent of the information age and the explosive growth of data,it is more and more urgent for people to discover the hidden rules of data from these large amounts of data.Therefore,machine learning,data mining and other big data analysis methods are increasingly paid attention to by experts and scholars.RoboCup simulation 2D is one of the oldest and oldest RoboCup World Cup subclass competition projects,which is an ideal platform for studying algorithms such as machine learning and data mining.Taking the optimization of RoboCup simulation 2D team defense strategy as the research direction,this paper proposes an adaptive DBSCAN algorithm-based defense strategy optimization method based on the distribution characteristics of data,the key area of team offense calculated by adaptive DBSCAN,and the cooperative strategy of players during team offense.The main work of this paper is as follows:Firstly,the concepts related to the RoboCup 2D simulation platform,the composition of the simulation platform and the research status at home and abroad are introduced,and the research on the offensive mode in the RoboCup 2D simulation 2D field is mainly introduced.Secondly,the classification of clustering algorithm and similarity measurement method is introduced,and the advantages and disadvantages of different clustering algorithm are illustrated,and then the adaptive DBSCAN algorithm adopted in this paper is introduced.The related concepts of adaptive DBSCAN algorithm and the rationality of automatic selection of adaptive parameters are emphasized,and the rationality of calculating the key attacking region in RoboCup 2D simulation is also discussed.Then,by extracting the relevant data of the attack state in the team log file and simulating the actual situation of 2D team attack in RoboCup,the dynamic division of the court was carried out.On the basis of the dynamic division,the team defense strategy was designed by using the calculated key area of the team attack and the cooperative strategy of the players during the team attack.Finally,the simulation experiment is carried out by modifying the underlying code according to the defense strategy.The simulation results show that the defense strategy designed by this method is reasonable and effective.This paper mainly provides a method to design reasonable and effective defensive strategies by dynamically dividing the court and using adaptive DBSCAN algorithm to calculate the key areas of the team’s attack and the players’ cooperative strategies during the team’s attack.