| 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。 |
| 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 |
| DOI：10.3969/j.issn.1001-893X. |
| Received:April 11, 2005Revised:October 21, 2005 |
| 基金项目:安徽省自然科学基金 |
|
| Modulation Identification of Digital Signals Based on Wavelet Transform |
|
| () |
| Abstract: |
| 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. |
| Key words: digital modulation signals,modulation identification,wavelet transform,artificial neural network(ANN) |
