PA-SLNR Multi-user MIMO Precoding Based on GSVD

Abstract

In this work, a linear precoding scheme for Multiuser Multiple Input Multiple Outputs (MU-MIMO) system is introduced. The proposed method optimizes Per-Antenna Signal to Leakage plus Noise Ratio (PA-SLNR) performance criteria, and makes use of the Generalized Singular Value Decomposition (GSVD) computation to develop a simple precoding algorithm. However, the linear technique available in the literature so far uses Per-User Signal to Leakage plus Noise Ratio (PU-SLNR) performance criteria and develop a closed form precoding matrix solutions based on Generalized Eigenvalue Decomposition (GEVD) and simple algorithm based on GSVD. The PU-SLNR objective function neglects to take the intra-user antennas interference cancellation into account. In this work, a new cost function based on PA-SLNR is defined, GSVD based solution is explained and simple algorithm is developed to compute the precoding matrices for multiple users with multiple antennas. Simulation results show that the proposed precoding method has a better BER and output SINR outage performance than the conventional methods