一种短波信道中数字信号调制方式识别算法

Algorithm for modulation recognition in shortwave channel environment

短波天波信道存在多径效应及冲击和高斯噪声。针对短波信道特点,提出先利用零记忆非线性系统抑制冲击噪声,后分段盲均衡,基于最大后验概率识别的分类方法。零记忆非线性系统可抑制冲击噪声,而让信号无失真通过。分段均衡减弱信道时变和参数估计误差对均衡器影响。由于冲击噪声的瞬时性和信道时变性,使每段信号均衡效果并不一样,采用最大后验概率的方法选用均衡效果最好的一段信号识别。仿真表明,在典型短波信道且含有冲击噪声情况下,信噪比10dB时,平均识别概率大于90%,该算法适合在实际中应用。

The shortwave channel is a kind of time-variant multipath channel with impulsive noise and Gauss noise. The method of zero-memory non-linearity system and segmental blind equalization is proposed for modulation recognition. An ideal zero-memory non-linearity system transforms the impulsive noise to Gauss noise, leaving the signal intact. Segmental equalization relieves the influence on equalizer caused by time-variant channel and error of parameters estimation. The results of equalization have much difference because of remain impulsive noise and time-variant channel. The optimal equalized segmentation determining the modulation type is selected by the maximum posterior probability. Simulation results show average successful rates are more than 90% in shortwave channels with impulsive noise when SNR is above 10dB.