摘要
针对吸气式高超声速导弹在纵向平面内以极限过载进行的机动,考虑导弹的工作动压窗口、攻角窗口,基于大量二分法仿真得到的数据,用导弹机动初始高度、马赫数以及质量作为神经网络的输入,用导弹以极限过载进行机动的机动时间的最大值作为神经网络的输出,对神经网络进行训练。训练好的神经网络即可以通过机动初始条件来预测导弹以极限过载进行机动的机动时间的最大值。在此基础上,把以极限过载进行机动的机动时间的最大值作为上界,随机选取以极限过载进行机动的机动时间,设计了动压约束的纵向机动突防方法,并通过仿真验证了该方法的可行性。
The neural network is trained for the maneuvering of air-breathing hypersonic missile at extreme overload in the longitudinal plane,based on a large number of dichotomies simulation data;considering the dynamic pressure window and the attack angle window;setting the missile maneuvering initial height,Mach number and mass as neural network inputs,and the maximum value of the maneuvering time that the missile maneuvering with the extreme overload as neural network output.The trained neural network can predict the maximum maneuver time by the initial conditions.On the basis,longitudinal maneuvering penetration method is designed by setting the maximum value of the maneuvering time as the upper bound and random selecting the time that the missile maneuvering with the extreme overload.The feasibility of this method is verified by simulation.
出处
《空天防御》
2018年第2期14-17,共4页
Air & Space Defense
关键词
吸气式高超声速导弹
极限过载
神经网络
二分法
纵向机动突防
air-breathing hypersonic missile
extreme overload
neural network
dichotomy
the maneuvering penetration in the longitudinal plane