基于联合参数估计的最大似然调制识别算法

ML-based modulation identification algorithm based on joint parameter estimation

提出一种针对延迟信号相位差的联合参数估计最大似然调制识别算法,通过对相邻信号相位差的分布建立等均值高斯混合分布模型,同时完成了延迟信号相位差最大似然调制识别中所需的频差和噪声似然参数的估计。此算法无需信号先验知识,可通过EM算法同时估出所需的相关参数,极大地降低了最大似然调制识别算法中参数估计的复杂度,可应用于MPSK信号的调制识别,也可用于MQAM信号的相位方式识别。

A new ML-based modulation identification algorithm for the phase angle difference between original signal and delayed signal is proposed. A two-state Gaussian mixture distribution model is established to fit the distribution of phase angle difference of the adjacent signals, and the parameters of Gaussian mixture distribution are solved by EM algorithm at the same time, then the parameters of mean and variance are used for ML-based modulation identifi- cation. Experimental results show that the parameters can be effectively estimated without knowing the modulation type, and the algorithm also can gain high probability of correct classification in the MPSK modulation identification or the phase mode identification of MQAM signals.