基于极端学习机的数字通信调制识别  

Digital communication modulation recognition based on extreme learning machine

作　　者：宋丽辉[1] 张靓[1] 

机构地区：[1]燕京理工学院,河北廊坊065201

出　　处：《激光杂志》2016年第3期119-122,共4页Laser Journal

基　　金：杰出应用人才培养计划研究与实践(SZ141191)

摘　　要：调制识别是数字信号通信系统中的一项关键技术,针对当前通信调制识别方法识别率低的难题,在分析当前数字通信调制识别研究现状的基础上,提出一种基于极端学习机的数字通信调制识别方法。首先提取数字通信信号的特征参数,并对特征组合在一起得到特征向量,然后将特征参数作为极端学习的输入建立数学通信调制类型的分类和识别模型,最后采用数字通信调制信号进行仿真实验。结果表明,当数字信号的信噪比低时,本文方法可以获得88%以上的数字信号识别率,可以满足数字通信系统的实际应用要求。Modulation recognition is a key technology in digital communication system, because current communication modulation recognition rate is low, this paper put forward a modulation recognition of digital communication method by using extreme learning machine, based on the analysis of the digital modulation recognition research on the current situation. The characteristic parameters of the typical digital communication signal is extracted, and the features are normalized, and then the parameters is taken as input of extreme learning machine and establish communica- tion modulation classification and recognition model, lastly, the simulation experiment is carried out by using digital communication modulation signals. The results show that when the digital modulation signal to noise ratio is relatively low, the proposed method can get more than 88% recognition rate, can satisfy the practical application requirement of digital communication system.

关 键 词：通信系统 调制识别 极端学习机 特征参数 

分 类 号：TN971[电子电信—信号与信息处理]