Sparse Layered MIMO with Iterative Detection

TL;DR

This paper introduces a sparse layered MIMO scheme that enhances spectral efficiency and error performance using non-orthogonal transmission and SVD precoding, along with a low-complexity detector and analytical error bounds.

Contribution

The paper proposes the novel SL-MIMO scheme combining non-orthogonal transmission with SVD precoding and develops a low-complexity detector and analytical error bounds.

Findings

SL-MIMO outperforms X- and Y-codes in error rate.

SL-MIMO achieves 5.6 dB and 4.7 dB gains in 6x6 MIMO with 64-ary constellation.

Analytical bounds validate the diversity gain of SL-MIMO.

Abstract

In this paper, we propose a novel transmission scheme, called sparse layered MIMO (SL-MIMO), that combines non-orthogonal transmission and singular value decomposition (SVD) precoding. Nonorthogonality in SL-MIMO allows re-using of the eigen-channels which improves the spectral efficiency and error rate performance of the system through enhancing the coding gain and diversity gain. We also present a low-complexity message-passing (MP) detector for the proposed SL-MIMO system which performs quite close to maximum likelihood (ML). The joint moment generating function (MGF) of the ordered eigenvalues is calculated and used to derive a closed-form upper bound on the average word error probability (AWEP) of the SL-MIMO system, and this derived expression is then used to analyze the diversity gain of the system. We use our analytical results to design sub-optimal codebooks to minimize the error rate of the SL-MIMO system. Simulation results in 4x4 and 6x6 multiple-input multiple-output (MIMO) systems with 4-ary, 16-ary, and 64-ary constellations show that our proposed SL-MIMO scheme outperforms competing approaches such as X- and Y-codes in terms of system error rate performance. SL-MIMO has 5.6 dB advantage compared to X-codes and 4.7 dB advantage compared to Y-codes in 6x6 MIMO system with a 64-ary constellation.