基于小波变换的数字通信信号识别  

Modulation Identification of Digital Signals Based on Wavelet Transform

作　　者：薛磊[1] 刘小秋[1] 

机构地区：[1]解放军电子工程学院,合肥230037

出　　处：《电讯技术》2006年第3期52-56,共5页Telecommunication Engineering

基　　金：安徽省自然科学基金资助项目(050420101)

摘　　要：小波变换对瞬态信息具有较强的检测能力。不同的数字通信信号在码元变化时呈现不同的瞬态信息。分别对幅度未归一化和幅度归一化的3种数字信号(ASK、FSK、PSK)进行小波变换,提取变换后包络方差与均值平方之比作为分类的特征参数,最后利用人工神经网络进行分类识别。仿真结果表明,在低信噪比(5dB)时该算法仍具有很高的识别率。Wavelet transform （WT） is suitable for detecting transient signals. Instantaneous features of digital signals will be presented when the symbols change. This paper discusses the wavelet transform of the signals with and without the amplitude normalization, then extracts the ratio of the variance to the square of mean as the key feature, and finally an artificial neural networks （ANN） classifier is proposed to identify three digital modulation signals： ASK, FSK and PSK. Simulations show that the algorithm is very effective at low SNR of 5 dB.

关 键 词：数字调制信号 调制样式识别 小波变换 人工神经网络 

分 类 号：TN911.72[电子电信—通信与信息系统]