论文部分内容阅读
为了解决同信道多信号的调制识别问题,提出了一种基于广义自回归(GAR)建模的调制识别方法。该方法利用观测数据的GAR模型参数估计各个待识别信号的短时平均中心频率和短时平均带宽,把一个多信号的调制识别问题转化为多个单信号的调制识别,并利用信号的短时平均中心频率和短时平均带宽的统计量作为特征输入到分类器,完成各个信号的调制类型识别。计算机仿真结果表明,当待识别信号在频域没有重叠或者部分重叠时,该方法都是有效的。
In order to solve the problem of modulation identification of co-channel multi-signal, a modulation identification method based on Generalized Autoregression (GAR) modeling is proposed. The method estimates the short-term average center frequency and the short-term average bandwidth of each signal to be identified by using the GAR model parameters of the observed data, converts a multi-signal modulation identification problem into multiple single signal modulation identification, and utilizes the signal short-time The average center frequency and short-term average bandwidth statistics are input to the classifier as a feature to complete the modulation type identification of each signal. Computer simulation results show that this method is effective when the signals to be identified do not overlap or partially overlap in the frequency domain.