Multicell Massive MIMO Multicasting with Finite-Alphabet Inputs and Statistical CSI

Abstract

We investigate the multicast precoding design in multicell massive multiple-input multiple-output (MIMO) systems with finite-alphabet inputs. Focusing on the multicast transmission with only statistical channel state information at the base station, we derive a lower bound on the achievable ergodic rate for finite-alphabet inputs, from which we utilize the concave-convex procedure (CCCP) to devise a CCCP-based algorithm maximizing the minimum weighted achievable ergodic rate lower bound. The CCCP-based algorithm is proven to converge to a local optimum. Furthermore, exploiting the channel characteristic in massive MIMO systems, we prove that the optimal precoding vectors should be linear combination of columns of eigenmatrix of transmit correlation matrices in order to maximize the minimum weighted rate lower bound with lower computational complexity. Then, a relation-based algorithm is developed to obtain the optimal solution of the weighted max-min fairness (MMF) problem by using the duality between the MMF and quality of service problem. Numerical results demonstrate the tightness of the achievable ergodic rate lower bound and the significant performance of the proposed algorithms.