Matrix Decomposition for Massive MIMO Detection

TL;DR

This paper analyzes matrix decomposition algorithms like QR, Cholesky, and LDL for massive MIMO detection, comparing their complexity and suitability for 5G systems to guide system and VLSI design choices.

Contribution

It provides a detailed complexity analysis of classical matrix decomposition methods applied to massive MIMO detection, highlighting their advantages and limitations.

Findings

QR, Cholesky, LDL-decomposition algorithms have varying computational complexities.

Comparison with approximate inversion methods shows trade-offs in accuracy and efficiency.

Insights assist system designers in selecting suitable detection algorithms for 5G MIMO systems.

Abstract

Massive multiple-input multiple-output (MIMO) is a key technology for fifth generation (5G) communication system. MIMO symbol detection is one of the most computationally intensive tasks for a massive MIMO baseband receiver. In this paper, we analyze matrix decomposition algorithms for massive MIMO systems, which were traditionally used for small-scale MIMO detection due to their numerical stability and modular design. We present the computational complexity of linear detection mechanisms based on QR, Cholesky and LDL-decomposition algorithms for different massive MIMO configurations. We compare them with the state-of-art approximate inversion-based massive MIMO detection methods. The results provide important insights for system and very large-scale integration (VLSI) designers to select appropriate massive MIMO detection algorithms according to their requirement.