基于非监督学习神经网络的自动调制识别研究与实现  

ON AUTOMATIC MODULATION RECOGNITION BASED ON UNSUPERVISED LEARNING NEURAL NETWORKS AND ITS IMPLEMENTATION

作　　者：徐毅琼[1] 葛临东[1] 王波[1] 叶健[1] 

机构地区：[1]解放军信息工程大学信息工程学院,河南郑州450002

出　　处：《计算机应用与软件》2011年第1期79-81,95,共4页Computer Applications and Software

基　　金：国家高技术研究发展计划项目(2006AA01Z146)

摘　　要：以非监督学习神经网络为主要研究对象,描述自组织网络的基本模型,分析传统自组织网络的训练算法,提出了一种基于自组织特征映射SOFM(Self-Organizing Feature Map)神经网络的通信信号自动调制识别方法。方法改进了训练算法中的学习率函数和邻域函数,提高了算法的收敛速度和性能,并将其应用在通信信号调制识别中。仿真实验检验基于SOFM神经网络的调制识别方法的性能,并与后向反馈(BP)神经网络加以比较,结果表明SOFM神经网络的调制识别方法具有较高的识别精度,改进后的训练算法提高了识别的有效性。This paper focuses on the unsupervised learning neural networks.Firstly,the basic structure of self-organised neural network is described.Then the traditional training algorithm of self-organised neural network is analysed,and the automatic modulation recognition method for communication signals based on self-organised feature map（SOFM） neural network is presented.The method improves the learning rate function and neighbourhood function of the training algorithm,enhances the convergence speed and performance of the algorithm,and has been applied in the modulation recognition of communication signals.Simulations test checks the performance of SOFM neural network based modulation recognition method,and compares it with the back-propagation（BP） neural network.Results illustrate that the modulation recognition method based on SOFM neural network has higher recognition precision,and the improved training algorithm has ameliorated its effectiveness of recognition.

关 键 词：调制识别 自组织特征映射神经网络 后向反馈神经网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]