基于数据驱动的自学习防空火力控制技术

Self-learning Air Defense Fire Control Technology Based on Driven-data

立足打赢未来智能化战争的作战需求,提高目标探测识别、运动参数估计的实时性和准确性,解决自主拦截决策建模难度大、决策结果稳定性差等问题,通过引入数据挖掘、深度学习、神经网络等人工智能技术,重点开展基于大数据的多类型目标状态空间模型分析、多探测模式下的目标融合识别技术、基于卷积神经网络的射击诸元修正等,研制一套自学习防空火力控制系统,有效弥补传统防空火力控制技术在时敏目标状态空间模型、大闭环校射、协同信息处理、控制决策等环节的不足,提升火控系统自修正、自学习能力,为武器装备向无人化智能化方向发展提供技术支撑。

Aiming at the operation requirements of winning the future intelligentized war and improving the real-time and accuracy of target detection identifaction and movement parameter estimation and such problems as the difficult modeling of autonomous interception decision-making and poor stability of decision-making results,such artificial intelligence techniques as data digging,deep learning,neural network,etc are introduced.The state space model analysis of multi-type targets based on big data,the target fusion identification technology in multi-detection mode,the fire data correction based on convolutional neural network,etc are focusly carried out,a set of self learning air defense fire control system is developed.To effectively make up for the shortcomings of such links as state-space model of time sensitive target,big closed loop fire calibration,coordinative information processing,control decision-making,etc.The capability of self correction and self learning of fire control system are improved and the technical support for the weapon equipment to develop in unmanned intelligent direction is provided.