基于支持向量机的典型宽带电磁干扰源识别

由于民航周围电磁环境复杂, 一旦产生电磁干扰(electromagnetic interference，EMI), 就不易被排查, 特别是随机性较强的宽带干扰。对此, 提出一种基于支持向量机(support vector machine, SVM)的干扰源识别方法。通过实时测量干扰信号的频谱数据, 并分析其特点, 选择包络因子、频谱能量、频谱峰值、均值和方差5个特征向量, 用主成分分析法降低数据冗余程度, 最后采用SVM来判断干扰源类型。仿真结果证明, 所提算法能有效识别6类典型机场宽带干扰源, 识别精度可达98.33%。

Typical wideband EMI identification based on support vector machine

Due to the complex electromagnetic environment around civil aviation, once the electromagnetic interference (EMI) is produced, it is not easy to be investigated, especially the random strong wideband interference. For wideband, an interference source recognition method based on support vector machine (SVM) is proposed. By measuring the spectral data of the signal in real time and analyzing its characteristics, five features of the evenlope factor, energy, peak value, mean and variance are selected as feature vectors, and principal component analysis is used to reduce data redundancy, finally, the type of the interference source is determined by SVM. Simulation results show that the identification algorithm proposed in this paper can effectively identify 6 types of wideband interference, and the identification accuracy is up to 98.33%.