论文部分内容阅读
在分析提升小波应用在调制模式自动识别的基础上,提出了一种新的特征提取方法。该方法首先利用最优估计理论获得小波的最佳预测系数,根据最佳预测系数进行分解提取特征值,最后利用支持向量机分类器进行信号的分类识别。在求解支持向量机的参数优化问题中,提出了一种通用的解决方案,利用带惯性权重的粒子群算法来确定其最优系数。新方法提取的特征值经计算机仿真研究证明,该算法具有较好的工程应用性和有效性。
Based on the analysis of lifting wavelet application in automatic recognition of modulation mode, a new feature extraction method is proposed. Firstly, the best prediction coefficient of wavelet is obtained by using the best estimation theory, then the eigenvalues are extracted according to the best prediction coefficient, and then the classification and recognition of signals are carried out by using SVM classifier. In solving the parameter optimization problem of SVM, a general solution is proposed, which uses the PSO with inertia weight to determine its optimal coefficient. The eigenvalues extracted by the new method have been proved by computer simulation that the algorithm has good engineering applicability and validity.