摘要
Multi-h连续相位调制(continuous phase modulation,CPM)信号与其调制指数均值相等的Single-h CPM信号的特征具有极大相似性,难以区分。针对该问题,提出了一种基于近似熵的Multi-h CPM调制识别算法。该算法将信号按照相同调制指数为一组的方式拆分为多个子序列,通过舍弃符号间拼接产生的多余模式向量对近似熵进行修正,然后利用Multi-h CPM信号各子序列近似熵的差异性,完成Multi-h CPM信号和Single-h CPM信号的类间识别,最后利用概率神经网络完成类内识别。实验结果表明,该算法在信噪比低至11 dB时,仍可以达到90%的识别率。
Multi-h continuous phase modulation(CPM)signal is indistinguishable from Single-h CPM signal,when the latter’s modulation index is equal to the former’s average of modulation indices,for their characteristics are similar.Aiming at this problem,a Multi-h CPM modulation recognition algorithm based on approximate entropy is proposed.The algorithm splits the complete sequence of signal into multiple sub-sequences according to the same modulation index,and discards the extra pattern vectors generated by splicing between symbols,to modify the approximate entropy.Then it uses the difference of entropy between subsequences of the Multi-h CPM signal,the inter-class recognition of Multi-h CPM signal and Single-h CPM signal is completed.Finally,the intra-class identification is accomplished by using probabilistic neural networks.The experimental results show that the algorithm can achieve 90%recognition rate when the signal-to-noise ratio is as low as 11 dB.
作者
刘凯
赵梦伟
黄青华
LIU Kai;ZHAO Mengwei;HUANG Qinghua(School of Communication and Information Engineering,Shanghai University,Shanghai 200444,China)
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2020年第3期698-703,共6页
Systems Engineering and Electronics
基金
国家自然科学基金(61571279)资助课题
关键词
Multi-h连续相位调制
调制识别
近似熵
概率神经网络
Multi-h continuous phase modulation(CPM)
modulation recognition
approximate entropy
probabilistic neural networks
