利用高阶累积量的时频重叠MPSK信号调制识别

针对多信号的互相干扰,在时频重叠情况下调制识别困难的问题,提出一种基于高阶累积量的识别方法。由于不同数字调相信号的高阶累积量存在差异;并利用高阶累积量对噪声的不敏感性,从中选取四个特征参数。采用决策树的分类方法,对任意方式组合而成的十种相移键控三信号进行制式识别。实验结果表明,在信噪比不低于0 dB的情况下,可以有效地识别十种时频重叠三信号。与现有的算法对比,该算法在提高信号识别率的基础上,同时也增加了可识别时频重叠信号的个数。

Modulation Identification of Time-frequency Overlapped MPSK Signal Using High Order Cumulant

Targeting at mutual disturbance caused by multiple signals and difficult modulation identification under the circumstances of time-frequency overlap, an identification method based on high-order cumulant is raised. There exist high-order cumulant divergences of different digital phase-modulated signals, and high-order cumulant is not sensitive for noises. By taking advantages of these two characteristics and selecting four feature parameters, classification method of decision tree is adopted to undertake pattern identification of ten types of phase-shift keying three signals. The experiment results show that, under the circumstance that noise-signal ratio is not less than 0dB, ten types of time-frequency overlapping three signals can be identified effectively. This algorithm not only improves the signal identification rate, but also increases the numbers of identical time-frequency overlap signals.