一种基于双DQN的空战干扰样式选择方法

An Air Combat Jamming Style Selection Method Based on Double DQN

针对深度Q网络(Deep Q Network,DQN)的干扰样式选择方法,提出了一种针对空战场景的双深度Q网络(Double Deep Q Network,Double DQN)干扰样式选择方法。以机载火控雷达空-空模式下的典型工作模式和我方干扰样式集为基础建立干扰矩阵作为算法输入,结合Double DQN原理提出了干扰样式选择方法。通过仿真实验对比了Double DQN和DQN算法的性能,并对部分参数设置进行探索,给出了两种空战态势下的最优干扰样式选择路径。仿真结果表明,该方法能够在满足实时性的条件下根据干扰效果在空战态势下选择出最优干扰样式。

For the jamming style selection method of deep Q network(DQN),a Double Deep Q Network(Double DQN)jamming style selection method for air combat scene is proposed.Based on the typical working mode of the airborne fire control radar in air-to-air mode and our jamming style set,the jamming matrix is established as the algorithm input.Then a method for selecting jamming styles is proposed according to the principle of Double DQN.Finally,the performance of Double DQN algorithm and DQN algorithm is compared through simulation experiments,some parameter settings are explored,and the optimal jamming style selection path under two air combat situations is given.The results show that the method can select the optimal jamming style in air combat situations according to the jamming effect under real-time conditions.