Joint precoding and antenna selection in massive mimo systems

Abstract

This thesis presents an overview of massive multiple-input multiple-output (MIMO) systems and proposes new algorithms to jointly precode and select the antennas. Massive MIMO is a new technology, which is candidate for comprising the fifth-generation (5G) of mobile cellular systems. This technology employs a huge amount of antennas at the base station and can reach high data rates under favorable, or asymptotically favorable, propagation conditions, while using simple linear processing. However, massive MIMO systems have some drawbacks, such as the high cost related to the base stations. A way to deal with this issue is to employ antenna selection algorithms at the base stations. These algorithms reduce the number of active antennas, decreasing the deployment and maintenance costs related to the base stations. Moreover, this thesis also describes a class of nonlinear precoders that are rarely addressed in the literature; these techniques are able to generate precoded sparse signals in order to achieve joint precoding and antenna selection. This thesis proposes two precoders belonging to this class, where the number of selected antennas is controlled by a design parameter. Simulation results show that the proposed precoders reach a lower bit-error rate than the classical antenna selection algorithms. Furthermore, simulation results show that the proposed precoders present a linear relation between the aforementioned design parameter that controls the signals’ sparsity and the number of selected antennas. Such relation is invariant to the number of base station’s antennas and the number of terminals served by this base station.