边际谱和多重分形在调制模式识别中的应用

Title:

Marginal spectrum and multifractal theoryand its application in modulation recognition

作者:

秦立龙^1,2, 王振宇²

1. 国防科学技术大学 电子科学与工程学院, 湖南 长沙 410073;
2. 解放军电子工程学院 通信对抗工程系, 安徽 合肥230037

Author(s):

QIN Lilong^1,2, WANG Zhenyu²

1. School of Electronic Science and Engineering, National University of Defence Technology, Changsha 410073, China;
2. Department of Communication Countermeasure Engineering, Electronic Engineering Institute, Hefei 230037, China

关键词:

调制识别; 边际谱; 分形理论; 支持向量机

Keywords:

modulation recognition; marginal spectrum; fractal theory; support vector machine

分类号:

TP18;TN911.7

DOI:

10.3969/j.issn.1673-4785.201301031

文献标志码:

A

摘要:

为了提高数字信号调制模式识别在低信噪比下的正确率,根据对边际谱和多重分形理论原理的分析,提出了一种新的基于多重分形理论的特征提取方法。该方法首先引入HHT变换求得样本的边际谱,不同调制模式的边际谱具有明显的差异性,可以利用分形的方法提取边际谱的分形维数作为调制识别的特征参数。最后利用支持向量机分类器进行信号的分类识别。并在求解支持向量机优化问题中,利用通用的粒子群算法确定了最优系数。计算机仿真研究证明,新方法提取的特征能有效地提高识别正确率,具有较好的工程应用性。

Abstract:

Through the analysis of the marginal spectrum and multifractal theory, a new feature extraction method based on multifractal theory was proposed to improve the accuracy of the digital modulation recognition under the low signal-to-noise ratio. First, the Hilbert-Huang transform was put forward to obtain the marginal spectrum of the samples. There are differences among different modulation modes. The fractal dimensions of the sample after Hilbert-Huang transform were calculated by the fractal method. Next, the feature was extracted. Finally, the identification task was solved by using SVM classification machine. In order to determine the optimal coefficient of the support vector machine, a universal particle swarm optimization algorithm was used. The computer simulation results showed that the performance of this feature extracted by the new algorithm efficiently improves the accuracy of modulation recognition and could be feasible to use in engineering applications.