基于支持向量机与谱特征的数字信号识别

Digital Modulation Recognition using Support Vector Machines and Spectral Characters

由于已有的调制识别方法存在提取的特征参数多、计算量复杂等不足,现提出一种将支持向量机与谱特征结合的方法,该法不仅提取的特征参数少,对噪声不敏感,计算量小,而且还通过设计分级的支持向量机分类器提高了识别率.仿真实验表明,这种方法比普通的阈值法在识别率方面有明显的改善.

Most modulation recognition methods are limited by computational complexity of extracting characteristic parameters,and a method of combining support vector machine and the spectral characteristics are developed. The method not only uses less characters which are not sensitive to noise and calculated simply, but also improves the recognition rate through design of hierarchical support vector machine classifier. Simulation results show that recognition rate is more significantly improved by this method than by the conventional way.