基于矩阵博弈的近距空战自主机动决策方法  

A Decision Method for Autonomous Maneuver in Close Range Air Combat Based on Matrix Game Theory

作　　者：王芳杰 黄鹏 张子俊 孙冲[2] Wang Fangjie;Huang Peng;Zhang Zijun;Sun Chong(Northwest Industrial Group Co.,Ltd.,Xi’an 710005,China;Northwestern Polytechnical University,Xi’an 710072,China)

机构地区：[1]西北工业集团有限公司,西安710005 [2]西北工业大学,西安710072

出　　处：《航空兵器》2023年第6期56-63,共8页Aero Weaponry

摘　　要：针对近距一对一场景下的无人机自主空战机动决策问题,提出了一种基于专家知识库和矩阵博弈方法的近距无人机一对一自主机动决策算法,能够在一对一近距空战中进行实时态势分析与机动决策并取得胜利。该算法构建了基于战术库的无人机自主空战触发机制,搭建了无人机动力学模型、导弹动力学模型,并基于典型空战因素与导弹攻击区分析构建了态势评价函数,将基于导弹攻击区的态势评价函数与矩阵博弈理论进行进一步的结合,构建出一套自主空战决策模型。在此基础上,设置典型场景下双方无人机均采用该算法进行博弈对抗仿真,多组仿真结果表明该算法能够有效进行实时机动决策博弈对抗,单机单步决策时间能够保持在80 ms以内。Aiming at the decision-making problem of UCAV autonomous maneuvering in close range one-to-one scenarios,a close range UCAV autonomous maneuvering decision-making algorithm based on expert know-ledge base and matrix game method is proposed,which can conduct real-time situational analysis and maneuver decision-making in one-to-one close range air combat to achieve victory.This algorithm constructs the launching mechanism of UCAV autonomous air combat based on tactical library,builds the dynamics model of fighter and missile,and builds the situation evaluation function based on typical air combat factors and missile attack area analysis.Further combining the situation evaluation function based on missile attack area with matrix game theory,a set of autonomous air combat decision model is constructed.On this basis,the algorithm is used for game confrontation simulation on both sides of UCAV in typical scenarios.Multiple sets of simulation results showed that the algorithm can effectively conduct real-time maneuvering decision-making game confrontation,and the single step decision-making time of a single machine can be maintained within 80 ms.

关 键 词：无人机 近距空战 自主决策 矩阵博弈 对抗 专家知识库 

分 类 号：TJ760[兵器科学与技术—武器系统与运用工程]