Experimental Study of User Selection for Dense Indoor Massive MIMO

Multi-user massive MIMO is capable of serving at least ten users simultaneously. However, when users are closely located, their high inter-user correlation is undesired; under this condition, these densely packed users must be separated by higher layer scheduling. In this paper, a low complexity greedy user selection method combined with the incremental inter-user-interference minimization criterion is proposed. The resulting algorithm is evaluated using system level simulations that rely on the measured indoor line-of-sight channel, with 64 antennas in the base station at 2.61GHz. Measurements are carried out using four different centralized and distributed base station antenna geometries, to evaluate the user selection performance for different indoor scenarios. Our evaluation shows that in a room with 64 densely deployed users, the proposed method increases the overall system sum rate, by up to 60%. Moreover, when applying this method, the distributed deployment outperforms the collocated scenario by 18% on average.