Non-Orthogonal Unicast and Multicast Transmission for Massive MIMO With Statistical Channel State Information

Abstract

We investigate non-orthogonal unicast and multicast (NOUM) transmission for massive multiple-input multiple-output systems, where only statistical channel state information of all user terminals is available at the base station. We adopt a weighted sum of the achievable ergodic unicast rate and multicast rate as the design objective. We first show the closed-form eigenvectors of the optimal unicast and multicast transmit covariance matrices, respectively, which reveals the optimality of beam domain transmission and simplifies the large-dimensional matrix-valued NOUM transmission design into a beam domain power allocation problem. Via invoking the concave-convex procedure, we, then, propose an efficient iterative beam domain power allocation algorithm with guaranteed convergence to a stationary point. In addition, we derive the deterministic equivalent of the objective in each iteration to further reduce the optimization complexity. Simulation results show that the proposed NOUM transmission can provide a significant performance gain in terms of the achievable ergodic unicast-multicast rate region over the conventional orthogonal approach.