【摘要】 无人作战飞机如何根据战场形势对空战态势迅速做出准确感知和评估,自主进行决策完成作战任务已成为研究的重点和热点;深度学习凭借其在图像和语音识别及其它诸多领域的巨大成就,已经引起了世界科学科技工作者的广泛兴趣。基于深度Q网络模型算法提出一种空战决策模型,用于获取具有时序特征的空战数据集,预处理后由随机森林模型对输入属性变量进行学习分类,即进行特征重要性排序,从而获取影响空战态势较为本质的特征表达,为无人作战飞机的自主智能决策提供支持。更多还原

【Abstract】 Unmanned Combat Aerial Vehicle(UCAV) how to make accurate perception and evaluation of the air combat situation rapidly according to the battlefield situation,and how to make decisions independently to accomplish the combat task has become the focus and hot spot of the research.Deep learning has attracted great interest of scientists and engineers in the world,owing to its great achievements in image and speech recognition and many other fields.Based on the depth Q network model algorithm,an air combat decision model is proposed to obtain the air combat data set with timing characteristics.After pretreatment,the input attribute variables are classified by the random forest model,that is,the order of importance is obtained.Therefore,the essential expression of the air combat situation can be obtained to provide supports for autonomous intelligent decision making of the unmanned combat aerial vehicle.更多还原