复合基神经网络在水声目标分类识别中的应用  

The application of a combined basis neural network in the underwater acoustic target classification and recognition

作　　者：方世良[1] 陆佶人[1] 

机构地区：[1]东南大学无线电工程系

出　　处：《声学技术》1998年第2期54-56,62,共4页Technical Acoustics

摘　　要：本文根据不同神经网络的分类特点，提出将径向基函数网络和多层感知器网络复合构成复合基网络，用于水声信号的分类识别。试验表明，该网络的分类能力及对未训练目标的适应性优于ＢＰ网和ＲＢＦ网。In this paper, a combined basis neural network is proposed for underwater acoustic target classification. It is composed of radial basis functions and multi layer perceptrons whose characteristics on the classification are different. It has been demonstrated by the recognition experiment that the network has higher correct recognition rate and better adaptability to untrained targets than BP or RBF network.

关 键 词：目标分类 水声探测 神经网络 目标识别 分类识别 

分 类 号：TB565.2[交通运输工程—水声工程]