基于遗传神经网络的战场声目标识别

通过小波包分析提取战场声信号不同频带上的能量作为特征向量,设计了遗传神经网络作为战场声目标识别的分类器,克服了传统BP神经网络容易陷入局部最小点的缺陷,通过Matlab仿真对比实验结果,证明遗传神经网络能够提高战场声目标的识别率。

The Recognition of Battlefield Acoustic Targets Based on Genetic Neural Network

Through extracting the energies on different frequency bands of battlefield acoustic signals as feature vectors by using wavelet packet analysis,this paper designs the genetic neural network as the classifier for the recognition of battlefield acoustic targets for overcoming the defects of traditional BP neural network that is easy to fall into partial minimum point,and based on the results of Matlab simulation contrast test,proves that genetic neural network can improve the recognition rate in battlefield acoustic targets.