| 自动识别数字调制方式的一种方法 |
| |
| 引用本文： | 谭晓波,张杭.自动识别数字调制方式的一种方法[J].电子信息对抗技术,2010,25(4):3-7,34. |
| |
| 作者姓名： | 谭晓波 张杭 |
| |
| 作者单位： | 解放军理工大学通信工程学院,南京,210007 |
| |
| 摘 要： | 提出了一种对经符号成形的数字通信信号进行调制方式自动识别的方案,该方案将数字已调信号的瞬时特征与功率谱特性相结合,设计了一组对信号信噪比不敏感的特征参数。通过提取经升余弦滤波成形的2FSK等六种常用的数字调制信号的5个特征参数,分别采用决策树和神经网络分类器进行调制方式自动识别。仿真表明,当信噪比为11dB时,采用决策树分类器对除2ASK外的5种信号的识别率在95%以上,神经网络分类器此时对所有信号的识别率达到了98%以上,证明了该方案的有效性和可行性。
|
| 关 键 词： | 调制识别 符号成形 特征提取 决策树分类器 神经网络 |
An Approach of Digital Modulation Automatic Recognition |
| |
| Authors: | TAN Xiao-bo ZHANG Hang |
| |
| Affiliation: | TAN Xiao-bo,ZHANG Hang(Institute of Communication Engineering,PLAUST,Nanjing 210007,China) |
| |
| Abstract: | A new automatic modulation recognition scheme for the signals symbol shaped is proposed,which based on the instantaneous time domain feature and spectrum characteristic.A set of feature parameters are designed which are not sensitive to signal-to-noise ratio(SNR).Five features of 2FSK and other five digital modulated signals shaped by raised cosine filter are extracted,decision-tree and neural network classifiers are used for automatic modulation recognition.Simulation proves that except 2ASK signal,all the... |
| |
| Keywords: | modulation recognition symbol shaped feature extraction decision-tree classifier neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
