论文部分内容阅读
利用观测样本的高阶累积量特征,在累积量域内构造信号分类特征,采用支持矢量机作为分类器,在高维空间实现对GSM、CDMA和OFDM信号的分类。该算法通过提取观测样本的累积量的识别特征矢量以区分不同的信号类型,并将特征向量映射到高维空间中加以分类,首先在理论上分析了算法的正确性,并通过仿真实验进行了验证,结果表明,算法具有较好的推广能力,在较大的信噪比范围内对三种信号均有较高的识别率。
Using the high-order cumulant features of the observed samples, the signal classification features are constructed in the cumulants domain, and the support vector machine is used as the classifier to classify the GSM, CDMA and OFDM signals in the high-dimensional space. The algorithm extracts the identification feature vector of the cumulants of observed samples to distinguish different signal types and maps the feature vectors into high-dimensional space for classification. Firstly, the correctness of the algorithm is analyzed theoretically and simulated by simulation The results show that the algorithm has a good popularization ability, and has high recognition rate for all three signals in the range of large signal-to-noise ratio.