两个新的特征参数在数字调制识别中的应用  

The application of two new characteristic parameters in digital modulation recognition

作　　者：张国战[1] 李旭[1] 

机构地区：[1]中国电波传播研究所青岛分所,山东青岛266107

出　　处：《中国电子商情（通信市场）》2010年第2期171-175,共5页

摘　　要：首先在综合分析了各种基本数字调制信号的瞬时幅度、瞬时频率、瞬时相位特性的基础上，提出了4个特征参数（其中两个特征参数是新提出来的，对噪声不敏感。），形成了比较完备的能够区别2ASK、4ASK、2FSK、4FSK、8FSK、2PSK、4PSK、8PSK和16QAM调制方式的特征矢量空间；构建了基于BP神经网络的数字调制分类器。并且对四个特征参数在不同的信噪比下的变化和分类器识别成功率进行了MATLAB仿真，仿真表明四个特征参数对噪声不敏感；在信噪比不低于5dB下，识别成功率大于96％，从而说明了两个新的特征参数的实用性。First comprehensive analysis of the various basic digital modulation signal of the instantaneous amplitude, instantaneous frequency, instantaneous phase characteristics based on the proposed four characteristic parameters （two characteristic parameters of the new proposal and is not sensitive to noise.）, formed a relatively complete to distinguish 2ASK, 4ASK, 2FSK, 4FSK, 8FSK, 2PSK, 4PSK, 8PSK and 16QAM modulation of the feature vector space; build a BP neural network based on digital modulation classifier. And the characteristic parameters of the four different SNR and classification success rate of recognition was MATLAB simulation, the simulation shows that the four characteristic parameters is not sensitive to noise; the signal to noise ratio of not less than 5dB, the identification success rate of greater than 96%, which shows the characteristics of the two new parameters of practicality.

关 键 词：人工神经网络 数字调制模式识别 特征矢量 

分 类 号：TN911.3[电子电信—通信与信息系统]