The Technology Research On Aircraft Passive Sonic Detect And Identification

In modern warfare, with electronic jamming, anti-radiation missiles, stealth technology and other technological development, as a main part of the airborne early warning system, radar systems is more vulnerable to be attacked. The aircraft passive sonic detects and identification technology is an effective complement for radar detection technology. It uses sonic signal of battlefield targets as a study object. Because it isn't send signals, so it is quite concealed. The aircraft passive sonic detects and identification technology is divided into two subsystems. Those are location and identification systems. The emphasis of this paper is design and simulation of identification subsystem.This paper studies sonic signal's time domain and frequency domain characteristics of two types targets, the targets is helicopters and fighter planes, and used of based on wavelet decomposition method for extracting energy characteristic details as feature extraction. The paper gives a detailed theory and algorithm process. Base on detailed energy scaling feature, this paper design two types of neural network classifiers: BP neural network classifiers and SOM neural network classifier. The Paper studies training algorithms of BP neural network and SOM neural network, and explores how options training parameters of the two classifier, use training samples done recognition experiment for two classifier, and compare recognition results. The results show that BP neural network classifiers and SOM neural network classifiers can reach a higher recognition rate, and use SOM neural network to target passive sonic recognition technology is feasible.