A Scalable Precoding Scheme based on Grassmannian Codebook for MU-MIMO

Abstract

This thesis presents a precoding scheme based on Grassmannian codebook with low feedback rate in MU-MIMO systems For a spatial-multiplexing system the codebook-based precoding is attractive for its low feedback rate requirement In this context the codebook design based on Grassmannian line packing has been proposed for multi-user MIMO systems However Grassmannian codebook requires to operate at sufficiently high signal-to-noise ratio (SNR) and spatially uncorrelated channels To address the aforementioned issues the solution proposed in this work consists of two parts To reduce the feedback rate requirement for MU-MIMO precoding the channel state information (CSI) is quantized However the BER tends to be saturated in high SNR region due to quantization noise And also when the channel independence is corrupted companding is used in Grassmannian codebook precoding to make BER lower On the other hand searching the optimal codeword in the multi-user scenario incurs high computational complexity To remedy the difficulty a low-complexity and efficient searching method is proposed based on genetic algorithm (GA) Simulation results are presented to demonstrate the efficacy of the proposed precoding method for MU-MIMO systems

Date of Award	2017 Aug 2
Original language	English
Supervisor	Kuang-Hao Liu (Supervisor)