摘要
自主空战机动决策是基于数学优化、人工智能等方法,模拟飞行员空战决策,自动生成飞行控制指令的过程。它在空战仿真、飞行员辅助决策和无人战斗机自主飞行等领域广泛运用。根据求解思路的不同,综合论述了两类自主空战机动决策方法:针对基于对策的空战决策方法,阐述了从追逃对策到双目标对策的发展脉络和内在联系,并着重分析了两类重要的双目标对策模型——矩阵对策和影响图对策;针对基于人工智能的空战决策方法,系统论述了基于专家系统的机动决策、基于遗传学习系统的机动决策、基于人工免疫系统的机动决策和基于神经网络的机动决策,明确了各种决策的建模方法、适用条件、改进途径等问题。总结了各个空战机动决策方法的优点及不足,指出了自主空战机动决策的进一步研究思路。
Based on the mathematical optimization and artificial intelligent algorithms,autonomous air combat maneuver decision can be used to simulate the pilots′ decisions,during the air combat,and provide control instructions,automatically.It is widely used in various fields including air combat simulation,polit assistant system,UCAV control and so on.According to their solving methods,they can be categorized into two kinds of maneuver decision,based on game theory and artificial intelligent,respectively.For the former kind of methods,this paper explains the nature connections between pursuit-evasion game and two-target game,as well as their progress during these years,and particularly analyses two important two-target game models which are marix game and influence diagram game.For the other kind of methods,this paper discusses maneuver decision based on expert system,genetic algorithms,artificial immune system and artificial neural systems respectively,and points out the modeling methods,boundaries of validity and possible improvement way for each algorithms.In the end,this paper summarizes the advantage and disadvantage of the two kinds of maneuver decision algorithms and provides new way for further research.
出处
《航空计算技术》
2012年第1期27-31,共5页
Aeronautical Computing Technique
基金
航空科学基金项目资助(20115185004)
航空装备预先研究基金项目资助(10102020101)
关键词
空战
机动决策
对策
人工智能
算法
air combat
maneuver decision
game
artificial intelligence
algorithm
