基于加权表决的决策层融合多系统调制识别  

Weighted voting based on decision fusion for modulation classification of multisystems

作　　者：程汉文[1] 陈亮[2] 吴乐南[3] 

机构地区：[1]防空兵指挥学院,河南郑州450052 [2]江苏大学机械工程学院,江苏镇江210013 [3]东南大学信息科学与工程学院,江苏南京210096

出　　处：《系统工程与电子技术》2010年第2期342-345,共4页Systems Engineering and Electronics

摘　　要：利用信息融合技术研究多系统调制识别问题,提出一种加权表决融合算法。该算法首先定义系统对识别结果的贡献度,用来描述多个系统的联合对识别性能的提升,并将其作为分配权重的依据。然后将这种多系统识别问题建模为合作对策模型,利用合作对策的Shapley值来决定各系统的权重。最后通过所提出的加权表决算法,对多个系统的调制方式判决结果进行决策层融合。计算机仿真表明该算法有助于提高调制识别的性能。Information fusion technique is utilized for modulation classification and a weighted voting algorithm is presented.Firstly,this algorithm defines a contribution to classification results to describe performance improvement of multisystems and depends on it to assign weights for each system.And then the modulation classification of multisystems is modeled as cooperative games and the weights are decided by Shapley value of cooperative games.Finally,by using the weighted voting algorithm presented,decision fusion results of modulation classification is processed.Computer simulation demonstrates an effective improvement in recognition performance via this algorithm.

关 键 词：调制识别 信息融合 加权表决法 合作对策 

分 类 号：TN911[电子电信—通信与信息系统]