Performance comparison of linear and nonlinear precoding for massive MIMO

Abstract

Massive Multiple-Input Multiple-Output (MIMO) answers the exponentially increasing demand for comprehensive fixed broadband and broadcast wireless communication services. Massive MIMO is a pivotal technology in the 5G and beyond (5GaB) wireless communication systems. This paper compares linear precoding algorithms such as Minimum Mean Square Error (MMSE), Neumann Series Approximation (NSA) with nonlinear precoders such as Tomlinson-Harashima Precoder (THP), and Lattice Reduction (LR) algorithm, the Lenstra-Lenstra-Lov'asz (LLL) precoder. The comparison was conducted using Bit-Error Rate (BER) and signal-to-noise ratio (SNR) performance measures. Simulated results prove that nonlinear precoders outperform linear precoding in high SNR regions.