J4 ›› 2014, Vol. 41 ›› Issue (2): 44-50.doi: 10.3969/j.issn.1001-2400.2014.02.008

• 研究论文 • 上一篇    下一篇

用于调制识别的二维累积量特征统计模型

刘沛1;水鹏朗1;郭永明2;李宁2   

  1. (1. 西安电子科技大学 雷达信号处理国家重点实验室,陕西 西安  710071;
    2. 国家无线电频谱管理研究所,陕西 西安  710061)
  • 收稿日期:2013-01-08 出版日期:2014-04-20 发布日期:2014-05-30
  • 通讯作者: 刘沛
  • 作者简介:刘沛(1983-),男,西安电子科技大学博士研究生,E-mail: liupei1983519@163.com.
  • 基金资助:

    国家科技重大专项资助项目(2010ZX03006-002-03)

Analysis of the statistical model with the  two-dimensional cumulant feature applying to modulation classification

LIU Pei1;SHUI Penglang1;GUO Yongming2;LI Ning2   

  1. (1. National Key Lab. of Radar Signal Processing, Xidian Univ., Xi'an  710071, China;
    2. National Institute of Radio Spectrum Management, Xi'an  710061, China)
  • Received:2013-01-08 Online:2014-04-20 Published:2014-05-30
  • Contact: LIU Pei

摘要:

高阶累积量是一种用于数字调制方式识别的重要特征.笔者在高斯白噪声信道下,构造了一种用于识别线性数字调制方式的二维归一化四阶累积量特征,并推导出该特征近似服从高斯分布.为了验证该模型与特征样本服从的统计模型一致,根据贝叶斯准则在二维特征平面上构造最大似然分类器,并从理论上推导出二元调制方式识别问题的平均分类正确率,它与仿真实验得到的平均分类正确率吻合得很好,证明了该方法的正确性.

关键词: 累积量, 调制识别, 统计模型, 高斯分布, 最大似然分类器

Abstract:

Higher order cumulants are the key features for implementing digital modulation classification. However, few available literatures focus on the statistical model of cumulant features. A two-dimensional normalized fourth-order cumulant feature is proposed to classify linear digital modulation in the additive white Gaussian noise channel, and then it is derived that the two-dimensional feature asymptotically obeys Gaussian distribution. In order to show the correctness of the proposition, a maximum likelihood classifier is formed in the two-dimensional feature domain according to the Bayesian criterion. The average probability of correct classification of the binary class problem is theoretically determined, which is consistent with the result obtained by simulations, thus justifying the correctness of the proposed theoretical results.

Key words: cumulants, modulation classification, statistical model, Gaussian distribution, maximum likelihood classifier

中图分类号: 

  • TN929.5