一种适用于低信噪比的数字调制方式识别方法

Digital Modulation Classification for Low SNR Environments

针对低信噪比平坦衰落信道,本文提出了一种有效的数字调制方式自动识别的方法,该方法主要采用的技术包括噪声功率估计、构造高维特征矢量、使用基于人工神经网络的分类器。噪声估计有效地抑制了噪声影响,高维特征矢量的应用提高了不同调制信号的区分度,基于人工神经网络的分类器可以实现对高维特征空间较复杂的划分。仿真结果表明该方法显著地提高了低信噪比条件下数字调制方式的正确识别概率。

An effective method of automatic digital modulation classification is proposed for low SNR flat-fading channel. The main techniques taken in the proposed method include noise power estimation, high dimension feature vector construction and the use of artificial neural network （ANN） as classifier. The noise estimation suppresses the impact of noise effectively. The use of high dimension feature vector improves the classifying capability between different modulation types. And the ANNclassifier is good at making complicated classification of high dimension feature space. Simulation results indicates thatthe proposed method obviously raises the correct classification probability in low SNR flat-fading channel.