Massive MIMO AF Relaying with Channel Estimation and Power Control Techniques

This paper investigates the performance of a massive multiple-input multiple-output relaying network where a group of single-antenna users communicate with a massive-antenna destination with the help of a multiple-antenna relay node. The transmission is performed based on a very simple amplify-and-forward relaying technique. To detect the signals transmitted from the users, the destination does not need to know all channels. It needs to know only the effective channel gains (the combined channel of the links from the users to the relay and the relay to the destination). Thus, we propose a channel estimation protocol at the destination where the pilot signals are first transmitted from the users to the relay, and then forwarded to the destination. The destination can estimate the effective channel gain from the received pilot signals, while the channel estimation overhead of the proposed scheme is independent of the number of destination antennas. A closed-form expression for the spectral efficiency with maximum-ratio-combining is derived. This closed-form expression enables us to suboptimally allocate power to each user and the relay to maximize the minimum of all user rates. We show that the power allocation can be solved through a sequence of geometric programs, and that the solution improves the system performance significantly.