Base station antenna configuration for robust MU-MIMO

This paper proposes a design methodology of base station antenna configuration for multi-user (MU) multiple-input, multi-output (MIMO) systems to be robust against channel feedback errors. In the proposed scheme, the base station antenna configuration controls the covariance matrices of MIMO channels so as to keep the space of pre-coding weights stable against the channel variation. Specifically, in case of 2-user scenario with 4-element linear array antenna for base station and 2-element terminal antenna, the optimal configuration is found by dividing antenna elements into two groups, where two groups of 2-antenna pairs are separated by 6A to realize diversity and multiplexing gains. Optimum antenna spacing for each group is 0.4A for enhancing the robustness against channel feedback errors and maximizing the system capacity. Moreover, user combinations for MU-MIMO are adaptively selected to maximize system throughput. However, if the optimal antenna configuration is employed, user combinations with larger angular differences tend to be selected and it degrades fairness among the users. Thus, we also introduced a generalized proportional fairness (GPF) based scheduling to keep fairness among users. Computer simulations confirm that optimal base station antenna configuration reduces system capacity deterioration against channel feedback errors dramatically, and it maintains the effectiveness of MU-MIMO system in terms of improving system capacity regardless of channel feedback errors.