摘要
论文为实现数字信号调制方式的自动识别,提出基于模糊C均值聚类算法和支持向量机的信号调制方式识别方法。设计了基于模糊C均值聚类(Fuzzy C-Means,FCM)的特征参数提取方法,构造聚类有效性评估函数得到不同聚类中心数下聚类有效性函数值,通过有显著差异的聚类有效性函数值来确定信号特征参数集合。利用支持向量机(Support Vector Machine,SVM)构造信号识别模型。与单独聚类方法的识别率相比,该方法提高了系统的调制识别率,尤其是在信号信噪比较低时,信号识别率明显提高。
To realize the automatic identification of digital signal modulation,the signal modulation recognition method is proposed based on Fuzzy C-Means（FCM）clustering algorithm and support vector machine（SVM）.Feature selection method is established based onFCM.The the cluster validity assessment function is constructed to get the Cluster validity function value under different cluster center number and the set of feature parameters of signals is obtained through analysis significant difference in clustering.The signal recognition model structured based on SVM.Simulation results show that classification rates of the algorithm proposed in this paper are much higher than those of clustering algorithm.Especially in low signal to noise ratio,signal recognition has improved significantly.
出处
《计算机与数字工程》
2013年第3期367-369,465,共4页
Computer & Digital Engineering
关键词
调制识别
模糊C均值聚类
特征提取
支持向量机
modulation recognition
fuzzy C-Means
feature selection
support vector vachine
