Blind Modulation Classification for MIMO systems using Expectation Maximization

In this paper, we propose a blind modulation classifier for multiple-input multiple-output (MIMO) systems. The assumption of unknown channel matrix and noise variance has not been considered prior to this work. For each modulation candidate, the channel parameters are jointly estimated via expectation maximization (EM). The resulting estimation is used for the likelihood evaluation of the corresponding modulation candidate. Classification decision is reached using the maximum likelihood (ML) criterion. Classification performance is validated in simulated fading channel with white Gaussian noise. The proposed classifiers achieves robust classification accuracy in most scenarios for BPSK, QPSK, and 16-QAM modulations.