采用广度优先搜索邻居聚类的通信信号调制识别

Modulation Classification of Communication Signals Based on Broad First Search Neighbors Clustering

针对短波通信中无法避免的码间串扰问题,研究了聚类算法在信号调制识别中的作用,提出了一种利用广度优先搜索邻居(BFSN)聚类处理循环统计量特征的分类算法。该算法将循环统计量特征峰值作为聚类输入对象,通过BFSN聚类分析,剔除延时信号、噪声等造成的奇异类峰值,克服了多径效应产生的码间串扰影响,实现了2FSK、4FSK、BPSK、QPSK、16QAM、π/4-QPSK、π/4-DQPSK、8PSK等8种调制信号的自动识别。仿真表明,该算法聚类后提取的特征参数抗多径干扰能力强,同信道均衡方法相比识别正确率有5%的性能优势。低信噪比环境下的信号调制识别具有重要的工程应用意义。

A new modulation classification algorithm based on broad first search neighbors（BFSN） cluste鄄ring is proposed to process cyclic statistic characteristics,so as to overcome the intersymbol interference（ISI） in high frequence（HF） communications. Firstly,the BFSN method is used to cluster and analyze dif-ferent cyclic statistic characteristics,the effects of ISI are reduced by eliminating the singular peaks causedby delay signals and noise;then the signals such as 2FSK,4FSK,BPSK,QPSK,π/4-QPSK,π/4-DQPSK,8PSK and 16QAM are automatically classified by joint feature estimation. Compared with channel equilibri-um algorithms,the proposed algorithm gains an improvement of 5% in classification accuracy. The results ofsimulation show that the proposed algotithm has a good performance in mitigating multipath interference andalso high value of engineering application in low signal-to-noise ratio（SNR） condition.