论文部分内容阅读
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。
For low signal-to-noise flat-fading channel, an effective method of automatic identification of digital modulation is proposed in this paper. The main techniques used include noise power estimation, high-dimensional feature vector construction and classifier based on artificial neural network. The noise estimation effectively suppresses the influence of noise. The application of high-dimensional eigenvectors improves the discrimination of different modulation signals. The classifier based on artificial neural network can realize the more complicated division of high-dimensional eigenspace. Simulation results show that the proposed method can significantly improve the probability of correct recognition of digital modulation under low signal-to-noise ratio.