Blind Massive Connectivity in mmWave MIMO: A Trilinear Factorization Approach via Hybrid Vector Message Passing

This paper addresses the challenge of blind massive connectivity in millimeter-wave (mmWave) multiple-input multiple-output (MIMO) systems, focusing on improving the efficiency of ultra massive machine-type communications (umMTCs) in sixth-generation (6G) wireless networks. We formulate blind massive connectivity as a Bayesian trilinear factorization problem involving the joint estimation of the angular array response, channel coefficients, and data signals. To solve the problem, we propose a novel algorithm, termed hybrid vector message passing for trilinear factorization (TF-HVMP), under a unified Bayesian inference framework. The TF-HVMP algorithm mitigates the energy leakage effect and provides superior performance in joint active user identification, channel estimation, and signal detection. Furthermore, we utilize the replica method to analyze the asymptotic performance of the trilinear factorization problem. Numerical results show that TF-HVMP closely approaches the replica bound and exhibits significant performance gains over the state-of-the-art algorithms.