一种神经网络最优中制导规律的工程实现方法

An Engineering Implementation Method of an Optimal Midcourse Guidance Law Using Neural Network

对于中远程导弹中制导段,提出了一种基于神经网络的最优中制导律工程实现方法,这种制导规律通过离线学习,能够在线实时工作.以垂直平面为例,分别使用BP神经网络、RBF神经网络和ELMAN神经网络形成最优中制导规律,比较了这些规律各自的优缺点.

This paper presents an optimal midcourse guidance law using neural network for midcourse phase of long-range missile, and the engineering implementation method is summarized in detail. This guidance law can operate in real time by means of off-line learning. In order to analyze this guidance law conveniently, the investigation is restricted to the vertical plane. Corresponding optimal guidance laws are formulated by means of Back Propagation（BP）neural networks, Radial Basis Function（RBF） neural networks and ELMAN neural networks, and their characteristics are compared.