无线电信号调制方式的自动识别研究
Study on Automatic Modulation Classification of Wireless Signals
自动识别检测信号的调制方式是信号监测和解调的中间步骤，在民/军应用领域都至关重要。本文对现行的调制方式识别方法进行了系统的研究和比较，针对无线电监测的特殊要求，综合考虑了计算复杂度和识别精度，提出了以高阶累计量作为分类特征量，结合其他信号特征量以及人工神经网络作为分类器的无线电信号调制方式高效自动识别系统。首次可以区分23种调制方式，并拥有较高精度。
utomatic modulation classification (AMC) is an intermediate step between signal detection and demodulation, and plays a key role in various civilian and military applications. Thus, the authors studied and compared different modulation classification techniques in a systematic way, combined with the concern of computing complexity and precision. This paper brings out an efficient AMC based on high order cumulants (HOC) combined with several features and neural network. The system could tell 23 kinds of modulated signals apart, with correct rate up to 95%. Finally, open problems and possible directions for future research are briefly discussed
邱天爽、郭莹、唐洪、宋宪夫
无线通信电子对抗通信
调制方式识别，高阶累计量，人工神经网络
Modulation Classification High Order Cumulants Neural Network
邱天爽,郭莹,唐洪,宋宪夫.无线电信号调制方式的自动识别研究[EB/OL].(2007-11-21)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/200711-431.点此复制
评论