摘要
针对多项式拟合和插值法解算空空导弹攻击区所存在的问题,提出了采用纯BP神经网络和BP网络-插值方法计算空空导弹攻击区。仿真获得纯BP神经网络和BP网络-插值法的界外发射概率、失去发射机会概率,结果表明这两种方法都有较高的拟合精度和速度。最后,分析了两种方法的优缺点,提出了根据作战需求和火控系统的配置情况确定选用何种方法的原则。
According to polynomial fitting and interpolation method resolving existence problem of air to air missile impact area, presents two methods of pure BP neural network and the interpolation of BP neural network. The simulation gains outside emission probability and loss of emission chance probability. Simulation results show they can also greatly increase the accuracy of fitting and velocity. The two methods have their own merits, but they also have their own shortage. In the air, which method is selected is decided by the capacity of air fire control system and combat demand.
出处
《制导与引信》
2005年第4期20-24,共5页
Guidance & Fuze
关键词
空空导弹
攻击区
神经网络
精度
air to air missile
impact area
neural network
accuracy
