Mathematical Optimization for Precoding in MIMO Systems

Abstract

This thesis proposes new precoding techniques through the usage of mathematical optimization algorithms to enhance the precoding stage of the wireless transmission and apply it to the emerging fifth generation (5G) wireless communications networks. The precoding stage in the 5G networks is challenging because recent technologies like massive multi-input-multi-output antenna arrays (M-MIMO) and small cells (SCs) need optimized precoders’ designs that can best use these technologies. The objectives of this thesis are finding optimized precoders’ designs for 5G technologies. The first objective is to propose an optimized precoder for a transmitter equipped with M-MIMO antenna array in a single-tier network. This is done by modelling the precoder problem, solving it analytically and testing through simulations the validity of this model. The second objective is to propose an optimized precoder for a group of transmitters in a double-tier network where one transmitter is equipped with M-MIMO and the other transmitters are SCs. This is done by modelling the precoder problem by two different models: a centralized model and a distributed model. Then, the two models are solved and the optimal precoders are compared through numerical simulations. A robust precoder model is proposed in the case of distributed processing to consider the practical case of the inability of the SCs to know information about users not served by them. Numerical simulations that prove the validity of the proposed models are presented. The thesis is organized into five chapters as listed below: