基于AR模型与神经网络的核爆与闪电电磁脉冲信号识别

 对平稳随机信号功率谱估计的AR模型，分别利用自相关函数法和Burg算法求该模型系数，作为核爆炸和闪电电磁脉冲信号的特征值;采用BP神经网络作为分类器以及不同的隐含层数和隐含层节点数，对核爆和闪电电磁脉冲实测数据进行识别研究。结果表明：AR参数模型法对两类信号特征值提取是非常有效的，采用Burg算法来求AR模型参数，其特征值提取效果优于自相关函数法。

Recognition of NEMP and LEMP signals based on auto-regression model and artificial neutral network

Auto-regression (AR) model, one power spectrum estimation method of stationary random signals, and artifical neutral network were adopted to recognize nuclear and lightning electromagnetic pulses. Self-correlation function and Burg algorithms were used to acquire the AR model coefficients as eigenvalues, and BP artificial neural network was introduced as the classifier with different numbers of hidden layers and hidden layer nodes. The results show that AR model is effective in those signals, feature extraction, and the Burg algorithm is more effective than the self-correlation function algorithm.