Automatic Modulation Classification of Overlapped Sources Using Multiple Cumulants

Automatic modulation classification (AMC) for overlapped sources plays an important role in spectrum monitoring and signal interception. In this paper, we propose a feature-based AMC framework for multiple overlapped sources. The framework first separates the overlapped sources via blind channel estimation and then conducts novel maximum-likelihood-based multicumulant classification (MLMC) for each of the sources. MLMC employs multiple cumulants of arbitrary orders and arbitrary lags as discriminating features and a maximum likelihood ratio test for decision making. Hence, MLMC maximizes the probability of correct classification under the condition that the selected cumulants are utilized. Moreover, both the case with perfect channel estimation and the practically more relevant case with blind channel estimations, called fast independent component analysis and natural gradient independent component analysis, are presented to facilitate the signal separation process. Extensive simulations are also conducted to verify the validity and the superiority of the proposed framework and the MLMC algorithm.