现代空战信息复杂度日益提升,传统空战行为量化指标仅能单一地从进攻角度为飞行员提供信息支撑,难以满足现代空战的决策要求。针对这一问题,文中结合空战行为过程与常见战术策略的分析,提出了战术控制距离(TCR)的新概念,建立了相应离线解算模型构建样本库。然后引入深度神经网络理论,设计了一种基于稀疏自编码网络(SAE)的深度拟合方法。仿真实验表明:SAE模型提高了解算精确性与时效性,其平方根均方误差控制在150 m以内, 满足超视距空战火控计算误差要求;TCR有效弥补了超视距空战节点信息的缺失,为飞行员提供了更加明确、直观的决策依据,具有较强的工程实用性与有效性。

With the increasing complexity of information in beyond visual range (BVR) air combat, traditional information indications can only provide support for pilots from the perspective of attack, which is difficult to meet the decision-making requirements in modern air combat. To address these problems, a new concept of tactical control range (TCR) is proposed under the consideration of escape maneuver in this paper. Firstly, the corresponding offline simulation model is given and the sample database is established. Then, fitting algorithm based on sparse auto encoder (SAE) network is designed by introducing the deep learning theory. The simulation results show that the SAE network improves the computing accuracy while greatly reducing the computing time. The simulation shows that the SAE model improves computational accuracy and timeliness, and its square root mean square error is controlled within 150 m, meeting the error requirements for BVR air combat fire control calculations. TCR can effectively makes up for the lack of information support and provide pilots with more timely decision basis in air combat, which is practical for improving the combat efficiency.

TN958. 53