【作者】 杜青； 刘剑飞； 刘娟； 乔延华；

【Author】 Du Qing Liu Jianfei Liu Juan Qiao Yanhua Information Engineering college of Hebei University of Technology,Hebei 300401

【机构】 河北工业大学信息工程学院；

【摘要】 信号调制方式的自动识别是未来软件无线电必须具备的功能之一,本文以模式识别为基础,介绍了模拟调制信号特征参数的提取,将小波神经网络应用于调制识别。在仿真中将应用于调制识别的小波神经网络算法与人工神经网络算法(ANN)做了比较,结果表明小波神经网络自动调制识别方法识别率高,解决了传统神经网络隐层节点个数难以确定的问题；当 SNR=10dB 时,正确识别概率已经达到93%,适于在工程中应用。更多还原

【Abstract】 Automatic recognition of modulation type of a signal is one of the essential functions of future Software Radio.Based on pattern recognition,the paper introduces the way of extracting characteristic features,and proposes modulation recognition based on wavelet neural network.The simulation compares the wavelet neural network with the ANN(artificial neural network)algorithm in modulation recognition.The results show that the proposed modulation recognition algorithm has the higher success rates,and resolved the problem which traditional neural network hide layer node numbers were difficultly determined.The success rates are over 93% when the signal noise rates are not lower than 10 dB.The proposed scheme is fit for engineering applications.更多还原