摘要
针对传统调制识别方法稳定性差及在小信噪比情况下识别率低的问题,提出了高斯过程分类调制识别方法。从调制信号中提取11种特征参数,采用高斯过程方法对其进行拟合,由此构造出高斯过程调制识别器,对11种模拟和数字调制信号进行同时识别,并与基于支持向量机的识别器进行比较。仿真结果表明,提出的方法识别率高且稳定性强,对各种调制信号均有很强的适应性,其性能明显优于支持向量机识别器,且在小信噪比情况下优势更为明显,为基于信号特征的调制识别提供了新的思路。
Aiming at the poor stability and low recognition rate of traditional modulation recognition methods under low signal to noise ratio（SNR） , a method of modulation recognition based on Gaussian process for classification （GPC） was proposed. Eleven character parameters were extracted from modulation signals and fitted by Gaussian process method to construct a recognizer of GPC to recognize analog and digital modulation signals simultaneously, which was compared with the recognizer based on support vector machine （SVM）. The simulation results show that the proposed method has high recognition rate, strong stability, and strong adaptability for all the modulation signals, whose performance is much better than the SVM recognizer with an obvious superiority under low SNR, providing a new thought for modulation recognition based on signal characters.
出处
《计算机仿真》
CSCD
北大核心
2015年第10期14-18,共5页
Computer Simulation
基金
国家自然科学基金资助项目(61379104)
关键词
调制识别
非合作通信
机器学习
高斯过程分类
Modulation recognition
Non - cooperative communication
Machine learning
Gaussian processfor classification
