基于AP聚类的数字信号调制体制识别方法

针对高斯噪声信道下MASK、MFSK和MPSK信号的类间识别问题,提出了一种基于短时傅里叶变换（Short Time Fourier Transform,STFT）和仿射传播聚类（Affinity Propagation Clustering,AP）相结合的信号类间识别方法。通过对在高斯信道下3类信号时域和频域特征的联合分析,提取出信号的时频特征。通过仿射传播聚类算法对信号进行聚类,通过信息迭代更新,可以快速、自动地找到聚类中心和聚类数目。仿真结果表明,在信噪比（SNR）较低的情况下仍能达到很好的分类效果。

A Modulation Recognition Method Based on AP

In this investigation,we propose a promising inter-class recognition method based on short time Fourier transform（ STFT） and affinity propagation clustering（ AP） to solve the problem of recognition among MASK,MFSK and MPSK modulated signal. The inter-class recognition method proposed consists of two parts. Firstly,based on signature analysis of these three types of signal in both time and frequency domain,we extract the time-frequency feature,and secondly,we cluster the signal by affinity propagation clustering technology. This method could find the clustering center and amount automatically and quickly by updating the information iteratively. Experimental simulations verify the superior performance of our presented inter-class recognition scheme even in lowSNR region.