系统工程与电子技术 ›› 2020, Vol. 42 ›› Issue (1): 223-229.doi: 10.3969/j.issn.1001-506X.2020.01.30

• 通信与网络 • 上一篇    下一篇

基于信息熵和GA-ELM的调制识别算法

李晨(), 杨俊安(), 刘辉()   

  1. 国防科技大学电子对抗学院, 安徽 合肥 230037
  • 收稿日期:2019-01-18 出版日期:2020-01-01 发布日期:2019-12-23
  • 作者简介:李晨(1995-),男,硕士研究生,主要研究方向为调制识别及机器学习。E-mail:lichen777444111@outlook.com|杨俊安(1965-),男,教授,博士研究生导师,博士,主要研究方向为信号处理及智能计算。E-mail:yangjunan@ustc.edu|刘辉(1983-),男,讲师,博士,主要研究方向为通信对抗及智能信息处理。E-mail:liuhui983eei@163.com
  • 基金资助:
    安徽省自然科学基金(1908085MF202);国防科技大学校基金(ZK18-03-14)

Modulation recognition algorithm based on information entropy and GA-ELM

Chen LI(), Jun'an YANG(), Hui LIU()   

  1. School of Electronic Countermeasure, National University of Defense Technology, Hefei 230037, China
  • Received:2019-01-18 Online:2020-01-01 Published:2019-12-23
  • Supported by:
    安徽省自然科学基金(1908085MF202);国防科技大学校基金(ZK18-03-14)

摘要:

针对当前通信信号调制识别算法在低信噪比(signal-to-noise ratio,SNR)下识别率低、训练速度慢、识别调制类型少的问题,提出了基于信息熵特征和遗传算法-超限学习机(genetic algorithm-extreme learning machine,GA-ELM)的调制识别算法。首先,提取信号的4种熵特征:奇异谱香农熵、奇异谱指数熵、功率谱香农熵和功率谱指数熵作为调制识别的特征参数;其次,采用GA-ELM作为分类器。仿真实验表明,对11种模拟、数字调制信号进行分类识别,在SNR大于4 dB时算法的总体识别率均超过98%,同时该算法训练速度快,识别系统设计简单,具有较大的应用价值。

关键词: 调制识别, 信息熵, 超限学习机, 遗传算法

Abstract:

In order to solve the problems of low recognition rate under low signal-to-noise ratio, slow training speed and few types of modulation in the current modulation recognition algorithms, this paper proposes a modulation recognition algorithm based on entropy feature and genetic algorithm-extreme learning machine (GA-ELM).Firstly, the four entropy characteristics of signals are extracted:Shannon entropy of singular spectrum, index entropy of singular spectrum, Shannon entropy of power spectrum and index entropy of power spectrum. Secondly, GA-ELM is used as the classifier. The simulation results show that the overall recognition rate of this algorithm is over 98% when the signal-to-noise ratio is more than 4 dB. At the same time, the algorithm has fast training speed, simple recognition system design and great application value.

Key words: modulation recognition, information entropy, extreme learning machine (ELM), genetic algorithm (GA)

中图分类号: