Blind Multiple Measurement Vector AMP Based on Expectation Maximization for Grant-Free NOMA

We consider a new approach to perform active user detection and channel estimation in massive grant-free access without requiring prior knowledge of the wireless channels, such as information on large-scale fading coefficients. To this end, we propose a multiple measurement vector approximate message passing (MMV-AMP) with expectation-maximization (EM)-based hyperparameter update, i.e., EM-MMV-AMP. Moreover, we revisited the decision rule for active user detection for EM-MMV-AMP. The numerical results indicate that the performance of the proposed scheme is superior to those of conventional schemes.