基于改进 CBR 的机器人足球对抗决策技术研究

摘要

为了进一步提高机器人足球比赛决策系统的智能性，让其能够模仿人类的思维模式进行比赛，提出了一种基于改进案例推理（ case-based reasoning ， CBR）的机器人足球对抗决策技术。该技术在传统CBR技术基础上，以“案例调整”来驱动“案例重用”，用一级案例库存储原始案例，用二级案例调整知识库存储案例调整知识；利用2级案例推理的方法，对场上局势和案例数据库进行匹配和调整，从而给执行层输出更合理的应对策略来解决当前赛场局势需求。实验结果表明，该技术比较有效、可靠。%This paper presents the intelligent confrontation decision-making system model based on two level CBR for r`obot soccer .This system model guides the robot football by imitating human thinking .Sections 1 and 2 of the full paper explain the two level CBR theory and how to build the robot football intelligent decision model which is based on two level CBR .Fig.2 illustrates the robot soccer decision-making process by using CBR technology .Fig.3 shows a soccer robot decision-making layer CBR case based structure .Section 3 verifies the feasibility of the robot soccer decision-making based on CBR with SimuroSot 5vs5 platform.Table 1 shows the specific content of the robot soccer case database on this platform .Table 2 shows the relationship between attribute weights and competition situ-ation.Table 3 shows the contents of the database strategy .