基于神经网络的战场被动声／地震动目标识别方法

目的在于研究人工神经网络在战场声／地震动目标识别应中的有效算法。通过建立战场目标声／地震动特性探测与分析系统，在总结目标特性规律，分析传统BP算法固有缺陷的基础上，采用改进的算法对分类器进行训练，典型战场目标信号样本检验表明，该方法具有良好的识别分类效果，利用基于神经网络的分类来实现对战场声／地震目标的识别分类是可行的。

Passive Sensing with Acoustics/Seism on the Battlefield Using Artificial Neural Network

This paper is intended to explore the effectiveness of battlefield targets classification and identification according to the characteristics of acoustic and seismic signals using ANN (Artificial Neural Network). Acoustic/seismic test and analysis system for typical battlefield targets is developed and properties of targets are acquired. Aiming at solving the drawback of extremely slow convergence speed of normal BP algorithm, a reformed BP algorithm is adopted to train the classifier. It is demonstrated that the reformed BP algorithm has higher correct identification rates for acoustic and seismic signals of battlefield targets according to signal sample experiments of typical targets, and the ANN classifier is suitable for the classification of battlefield targets.