基于深度学习的通信信号调制方式识别

Communication Signal Modulation Recognition Based on Deep Learning

针对采用传统机器学习算法对通信信号调制识别方法中的计算复杂度高、准确率低以及人工提取特征步骤繁琐等问题,提出一种基于深度神经网络通信信号调制识别模型。模型可以直接识别经过采样之后的通信信号类别,且具有识别准确率高、通用性强、抗噪声性能好及处理流程简便等特点,有效解决了传统算法无法实现自动提取特征的缺陷。通过大量实验以及对通信信号特征的准确分析,采用卷积神经网络和循环神经网络等网络的组合设计,构建了一个识别准确率较高且端到端的通信信号识别模型。

Aiming at the problems such as high complexity, low accuracy and cumbersome manual extraction of features by traditional machine learning algorithm,a kind of communication signal modulation recognition model based on deep neural network is proposed. The model can directly identify the sampled communication signal category,and has such characteristics as high recognition accuracy,strong versatility,good anti-noise performance,and simple processing flow. At the same time,the method of using neural network for identification can effectively solve the problem that the traditional algorithm cannot realize automatic characteristic extraction. Based on a large number of experiments and accurate analysis of communication signal characteristics,a combination of convolutional neural networks and cyclic neural networks is used to construct a communication signal recognition model with high recognition accuracy and end-to-end characteristics.