基于循环谱和稀疏表示的多信号调制识别*1

针对难以识别时频重叠的多信号问题,提出一种不用分离混合信号即可识别信号类型的新方法。该方法针对各种调制循环谱的不同,用稀疏表示提取各信号的特征,最后根据提取的特征利用支持向量机对信号进行识别分类。经理论推导和仿真实验得出:该方法对噪声具有一定的鲁棒性,在较低信噪比条件下仍能保持较好的识别性能,在信噪比为-4 d B时,对单信号和混合信号的正确识别率分别可达到93.5%和90.67%。

Multi-signals Modulation Recognition Based on Cyclic Spectrum and Sparse Representation

For the modulation recognition problem of several time-frequency overlapping modulation signals, a new method that can identify the signals mixed signal types without separation is proposed. For different cyclic spectrum of different modulation signals, the features can be extracted by using sparse representation, and finally according to the extracted feature ,support vector machine is used to recognition and classification. By theoretical analysis and experimental simulation can be obtained: this method has a certain robustness of noise, it still has a good recognition performance at low SNR. When the SNR is -4dB , the correct recognition rate of single and mixed signals can respectively reach 93.5% and 90.67%.