基于BP网络的群射防空火箭炮变间隔研究

Study on Variable Firing Interval of MRLS Using BP Neural Network

通过动力学仿真为BP网络建立训练样本,从而使其能够模拟变发射间隔与起始扰动的非线性动力学关系。在对BP网络训练后对变间隔的问题进行研究,并进行优化,从而得出了一组较好的发射间隔值,使得相应的起始扰动较小。动力学的仿真表明BP网络优化的结果比较可靠,从而为确定群射防空火箭炮的发射间隔提供参考。

The BP neural network is trained to be able to simulate the non-linear dynamical connection between the variable Cuing interval and the initial disturbance with swatch from the dynamical simulation. The variable interval was studied using the trained neural network and optimized, and a group of interval value was obtained with the smaller initial disturbance. The simulation of dynamical gives some evidence that optimization result credible, which can be referenced for the interval of the MRLS used in anti air-strike battle.