Two-Stage LASSO ADMM Signal Detection Algorithm For Large Scale MIMO

TL;DR

This paper introduces a two-stage LASSO ADMM algorithm for large scale MIMO detection, combining machine learning techniques and optimization tools to improve accuracy and efficiency over existing methods.

Contribution

It proposes a novel two-stage LASSO ADMM detection algorithm with interference cancellation for large scale MIMO systems, enhancing detection performance.

Findings

Effective in various modulation schemes

Outperforms traditional detection algorithms

Works well with both uncoded and coded data

Abstract

This paper explores the benefit of using some of the machine learning techniques and Big data optimization tools in approximating maximum likelihood (ML) detection of Large Scale MIMO systems. First, large scale MIMO detection problem is formulated as a LASSO (Least Absolute Shrinkage and Selection Operator) optimization problem. Then, Alternating Direction Method of Multipliers (ADMM) is considered in solving this problem. The choice of ADMM is motivated by its ability of solving convex optimization problems by breaking them into smaller sub-problems, each of which are then easier to handle. Further improvement is obtained using two stages of LASSO with interference cancellation from the first stage. The proposed algorithm is investigated at various modulation techniques with different number of antennas. It is also compared with widely used algorithms in this field. Simulation results demonstrate the efficacy of the proposed algorithm for both uncoded and coded cases.