A low complexity deterministic algorithm for the antenna/user selection in multi-antenna systems

When a MIMO wireless channel matrix H is rank deficient, it is often required to select a submatrix Hp ϵ H such that the columns of Hp lie close to the optimal subspace that constructs H. Main motivation for the selection of Hp is to gain the same spectral efficiency benefits as that of H, but at a lower computational and RF hardware cost. In this contribution, we first present an analogy between the antenna and user selection problems. Secondly, a matrix volume maximization criterion that is applicable to both problems have been presented. We have proposed a low complexity deterministic algorithm for the selection of antennas/users such that the volume of Hp is maximum. We have shown that under particular channel propagation setup, smaller subsets can provide the same spectral efficiency as that of the full MIMO channel. We used a 3D extension of the WINNER channel model at center frequency 2.53 GHz to evaluate the performance of the proposed algorithm.