基于高效自适应聚类算法的调制识别研究  

Study on efficient adaptive clustering algorithm based modulation identification

作　　者：叶健[1] 吴月娴[1] 葛临东[1] 

机构地区：[1]解放军信息工程大学,河南郑州450002

出　　处：《计算机工程与设计》2007年第3期506-508,共3页Computer Engineering and Design

基　　金：总装重点科研基金项目(6130320)

摘　　要：提出了一种基于星座聚类的通信信号调制识别新方法。该方法将星座图形状作为调制识别的特征,运用聚类算法EAFCM(efficient adaptive fuzzy C-means)重建接受信号的星座图。基于模糊C-均值(FCM)聚类算法的自适应高效聚类算法EAFCM不仅克服了模糊C-均值聚类算法需要预先确定聚类参数c、对初始中心敏感等不足,而且具有良好的抗噪声性能。将该方法应用到对PSK/QAM信号的调制识别,实验结果表明该方法是实际有效的。A novel modulation recognition method is proposed, adopting the constellation shape as classification characteristics. A new clustering method, named EAFCM, based on fuzzy C-means algorithm, is used for rebuilding the constellation. It not only overcomes the sensitivity to initial centers as FCM, but also performances well with signals of lower SNR. This method is applied to modulation recognition of PSK/QAM signals and the experiments show its feasibility.

关 键 词：调制识别 EAFCM算法 自适应 减法聚类 聚类有效性分析 星座图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]