Comparison between MU-MIMO using Primary Pair User Scheduling and Beamforming using Interference Nulling

Abstract

This dissertation proposes a user selection algorithm for Multi-User Multi-Input Multi-Output (MU-MIMO) contexts and a nulling algorithm for beamforming that can be applied to 4G, 5G, and space communication. The proposed algorithms are implemented by simulators to compare and analyze the performance and characteristics. The user selection algorithm for MU-MIMO is as follows. The proposed user selection algorithm gives same performance compared to exhaustive search, but the algorithm offers less amount of computation. Through computer simulations, the proposed method outperforms the very well-known correlation-based method, Semi-orthogonal User Selection (SUS). The proposed method outperforms SUS, yielding a sum rate that is about 0.57 bps/Hz higher when the transmit Signal to Noise Ratio (SNR) is 10 dB with perfect Channel State Information (CSI) at Base Station (BS) and the numbers of users and transmit antennas in a cell are 100 and 4, respectively. To provide realistic simulation results, Grassmannian code is introduced and applied to MU-MIMO. Because the precoding matrix is not configured using exact inversion of channel matrix, the performances of overall MU-MIMO are degraded without perfect CSI. Nevertheless the proposed algorithm is relatively decreased compared to exhaustive search, because the proposed algorithm is based on power level. As a result, the proposed algorithm has a better performance than exhaustive search. When the nulling algorithm is processed, two algorithms based on gradient and Conjugate Gradient Method (CGM) are compared each other. The CGM-based algorithm performs better than the gradient-based algorithm when interference power is large. On the other hand, the gradient-based algorithm performs better than the CGM-based algorithm when interference power is small since the CGM-based algorithm has a weak main beam. In a case using 100 antennas, the performance of CGM based algorithm is better than the performance of gradient-based algorithm when interference power is larger than 26dB. Thus, the CGM-based algorithm will be better in satellite systems where jammers can disturb communications, and the gradient-based algorithm will be better in ground communication equipment that has relatively lower interference. Finally, simulators considering the environment of interference channel are constructed to compare MU-MIMO with beamforming. MU-MIMO turned out to perform better in the cell center; beamforming, on the other hand, is better in the cell edge. Using the previously presented results, this dissertation will facilitate improved efficiency in 4G as well as 5G and space communications.