基于信息融合的信号调制方式识别  

Modulation Classification based on Information Fusion

作　　者：程汉文[1,2] 吴乐南[1] 

机构地区：[1]东南大学信息科学与工程学院,南京210096 [2]防空兵指挥学院,郑州450052

出　　处：《信号处理》2009年第4期625-629,共5页Journal of Signal Processing

摘　　要：本文利用信息融合技术,给出了多系统调制识别方法。单个系统利用高阶累积量构造识别特征。数据加权、特征平均、最小距离方法分别用于多系统数据层、特征层和决策层的融合。仿真结果显示多系统的调制正确识别率高于单系统的正确识别率,其中基于数据层融合的正确识别率高于基于特征层和基于决策层融合的正确识别率。说明信息融合有助于提高调制识别性能。In this paper, we propose a modulation classification method for multi-system by utilizing information fusion tech- nique. High order cumulants act as classification features in every classification system. A weighted data, feature＇ s mean and minimum distance algorithms are utilized in fusion at data, feature and decision layer respectively for multi-system. Simulations showed the correct classification rate of multi-system is higher than that of single system, and this performance based on data fusion is higher than those based on feature and on decision fusions. It demonstrates that information fusion be helpful to performance of modulation classifiers.

关 键 词：调制识别 信息融合 累积量 系统 

分 类 号：TN911.72[电子电信—通信与信息系统]