认知无线电Underlay模式下MQAM信号的调制识别

Modulation Identification of MQAM Signals in Underlay Cognitive Radios

针对传统的认知无线电Underlay中时频重叠MQAM信号调制识别方法性能低的问题,提出了一种采用时频分析图像纹理特征的时频重叠信号调制识别方法。首先对接收到的时频重叠MQAM信号做频率切片小波变换得到时频分析图像,并选取纹理差异明显的切片部分进行灰度化处理,然后通过提取时频分析图像的灰度-梯度共生矩阵特征,最后利用径向基函数神经网络分类器有效地实现了时频重叠MQAM信号调制方式的识别。仿真结果表明:在信噪比为4dB下,所提出的方法的平均识别率可达到95%以上;在信噪比大于0dB时,所提方法的识别性能优于基于高阶累积量的识别方法。

A novel method of modulation identification based on texture features of timefrequency analysis images is proposed to address the poor performance problem of traditional methods of modulation identification of time-frequency overlapped MQAM signals in underlay cognitive radios.A time-frequency analysis image is obtained by using a frequency slice wavelet transform to the time-frequency overlapped MQAM signals,and slices with obvious texture difference are selected for grayscale processing.And then,the features of the grayscale-gradient co-occurrence matrix of the time-frequency analysis image and the RBF neural network classifier are used to achieve the modulation recognition of the time-frequency overlapped MQAM signals.Simulation results show that the average recognition rate of the proposed method reaches more than 95% when the SNR is 4 dB,and its recognition performance is better than that of the method based on high-order cumulants when the SNR is more than 0 dB.