Research on shooter modeling of image guided missile based on neural network

In the human-machine-environment system of shooter-missile-battlefield, shooter is very important for image guided missile to track and attack targets in complex ground successfully. To build the model of shooter, the angle error between missile´s optical axis and line of sight, and the change rate of the angle error were set to be inputs of model. The variable describing handle movement controlled by shooter was regarded as the output of model. Based on one group of representative data of shooter, the method of principal component analysis and Bayesian-regularization BP neural network were adopted to build the model by means of neural network identification. Simulation results prove that the neural network model of shooter is of good precision and good generalization ability. The model can be applied to design of guidance and control system of the missile and the method of shooter modeling can provide reference for modeling of human in other systems.