基于多端CNN的通信信号自动调制识别研究

Research on Automatic Modulation Recognition of Communication Signals Based on Multi terminal CNN

摘要:

为有效解决通信信号自动调制识别的调制类型识别率低和调制强度识别误差高的问题，研究了基于多端CNN的通信信号自动调制识别方法。根据不同类型调制方法的作用原理，设置通信信号调制识别标准。考虑通信信号的传输过程，构建通信信号模型，利用带通采样工具采集初始通信信号，通过小波消噪、归一化等步骤，完成初始信号的预处理。利用多端CNN算法构建通信信号识别器，提取幅值、相位、频率等通信信号特征参数，通过特征匹配得出信号调制类型与强度的识别结果，实现通信信号自动调制识别。通过与传统识别方法的对比得出结论：综合考虑有、无干扰两种类型的通信信号，优化设计识别方法的调制类型识别率提高了49.6%，调制强度识别误差降低了约0.0285。

Abstract:

In order to effectively solve the problems of low recognition rate of modulation type and high recognition error of modulation intensity in automatic modulation recognition of communication signals, an automatic modulation recognition method of communication signals based on multi terminal CNN was studied. According to the principle of different types of modulation methods, the communication signal modulation identification standard is set. Considering the transmission process of the communication signal, the communication signal model is constructed, and the initial communication signal is collected by using the bandpass sampling tool. The preprocessing of the initial signal is completed through wavelet denoising, normalization and other steps. The multi terminal CNN algorithm is used to build a communication signal recognizer, extract the amplitude, phase, frequency and other communication signal characteristic parameters, and obtain the recognition results of signal modulation type and intensity through feature matching, realize automatic modulation recognition of communication signals. By comparing with the traditional recognition methods, it is concluded that the recognition rate of modulation type is improved by 49.6% and the recognition error of modulation intensity is reduced by 0.0285 by considering the two types of communication signals with and without interference.