Energy Efficient Multi-Way Massive Mimo AF Relaying with Imperfect CSI and Optimal Power Allocation

In this paper, a multi-way massive multiple-input and multiple-output (MIMO) communication system is considered, where multiple users equipped with single-antenna send their data through one multiple-antenna amplify-and-forward (AF) relay. First, the relay estimates the channels of all users and applies a maximum-ratio combining (MRC) process in a multiple-access (MAC) stage by assuming imperfect channel state estimation (CSI) at the relay. Then, a maximum-ratio transmission (MRT) is employed at the relay to transmit signals in a broadcast (BC) stage. The asymptotic performance of signal-to-interference-plus-noise ratio (SINR) is derived under three power-scaling scenarios. Based on the asymptotic SINR outcomes, the optimal power allocation (OPA) algorithm is formulated to provide an energy-efficient system. Besides, the spectral efficiency (SE) and energy efficiency (EE) of three asymptotic scenarios are studied, and a trade-off between SE and EE is exhibited under imperfect CSI. Finally, simulation results illustrate that the transmit power at each user and/or relay is proportional to the inverse of the antennas number at the multi-way massive MIMO (MWMM) relay. Also, it reveals that the OPA algorithm significantly outperforms the UPA under imperfect CSI.