论文部分内容阅读
研究了一种基于高阶累积量和神经网络的干扰识别算法。该方法把卫星通信中常见的各种干扰信号的归一化高阶累积量作为分类特征参数,应用神经网络对特征参数进行分类训练,将接收干扰信号的归一化高阶累积量输入已训练的神经网络进行干扰类型的识别。试验结果表明:该算法在低干信比的情况下具有较高的识别准确率。
An interference identification algorithm based on high order cumulant and neural network is studied. In this method, the normalized higher-order cumulants of all kinds of interference signals commonly used in satellite communication are taken as the classification characteristic parameters. The neural network is used to classify the characteristic parameters and the normalized higher-order cumulants of the received interference signals are trained Neural network for interference type identification. Experimental results show that the proposed algorithm has high recognition accuracy at low signal-to-interference ratio.