Linear Precoding for the MIMO Multiple Access Channel with Finite Alphabet Inputs and Statistical CSI

TL;DR

This paper develops a low-complexity linear precoder design for MIMO MAC with finite alphabet inputs and statistical CSI, achieving significant performance improvements in large system limits.

Contribution

It introduces an asymptotic analysis and optimal precoder structures for finite alphabet inputs under statistical CSI, with an efficient iterative design algorithm.

Findings

Proposed precoder outperforms existing designs in simulations.

Significant reduction in computational complexity.

Achieves notable performance gains in large system scenarios.

Abstract

In this paper, we investigate the design of linear precoders for the multiple-input multiple-output (MIMO) multiple access channel (MAC). We assume that statistical channel state information (CSI) is available at the transmitters and consider the problem under the practical finite alphabet input assumption. First, we derive an asymptotic (in the large system limit) expression for the weighted sum rate (WSR) of the MIMO MAC with finite alphabet inputs and Weichselberger's MIMO channel model. Subsequently, we obtain the optimal structures of the linear precoders of the users maximizing the asymptotic WSR and an iterative algorithm for determining the precoders. We show that the complexity of the proposed precoder design is significantly lower than that of MIMO MAC precoders designed for finite alphabet inputs and instantaneous CSI. Simulation results for finite alphabet signalling indicate that the proposed precoder achieves significant performance gains over existing precoder designs.