Efficient EP Detectors Based on Channel Sparsification for Massive MIMO Systems

Dong, Yuanyuan; Hua, Li; Zhang, Zhenyu; Wang, Xiyuan; Dai, Xiaoming

doi:10.1109/lcomm.2019.2962449

Cited by 16 publications

(17 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…are priors associated to coded binary vector labels ci that map to constellation symbols xi, ∀i ∈ 1, Nt . This factorization is used in [3,10,11] and referred to as scalar factorization. The scalar FG drawn from (2) is shown in Fig.…”

Section: Factor Graph Representation With Matrix Decompositionmentioning

confidence: 99%

“…Assuming H known at the receiver, previous works [10,11] use QR Decomposition (QRD) [9] to split the channel into two matrices, with H = QR, where Q is a unitary matrix and R is an upper triangular matrix. Then, the observation model can be updated as follows:…”

Section: Factor Graph Representation With Matrix Decompositionmentioning

confidence: 99%

“…1. Authors from [3] show the application of EP on scalar FG without decomposition and authors from [10] show the application of EP on scalar FG with QRD. There are several possible types of iteration in the graph, the first one between equalization nodes and variable nodes xi is called inner iteration, the second between equalization nodes and demapping nodes is called auto iteration, and the last between equalization nodes and prior probabilities nodes is called turbo iteration.…”

Section: Expectation Propagation Over Gkd or Svd Factor Graphsmentioning

confidence: 99%

“…Several matrix decompositions can be used, like the QR decomposition (QRD), the Golub-Kahan bidiagonal Decomposition (GKD) [9], all the way to the Singular Value Decomposition (SVD) to alter/modify the underlying graph. Prior works [10,11] applied QRD and corresponding pre-processing using MPAs. The QRD produces a fairly similar FG compared to its initial form, except fewer messages need to be exchanged but the overall structure remains the same.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Expectation Propagation on Factor Graphs Based on Matrix Decomposition

Mekhiche

Cipriano²,

Poulliat

2023

ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In the context of the Gaussian linear model, recent works have studied factor graph modification using QR decomposition that enables the derivation of scalar Expectation Propagation (EP) based detectors. In this paper, we investigate on new factor graph representations induced by the use of the Golub-Kahan bi-diagonal Decomposition (GKD) and of the Singular Value Decomposition (SVD). New EP messages induced by the GKD or SVD underlying graphs are derived, that can be both scalar or vector messages. Complexity and performance of the resulting algorithms are studied for digital communications applications.

show abstract

Section: Factor Graph Representation With Matrix Decompositionmentioning

confidence: 99%

Section: Factor Graph Representation With Matrix Decompositionmentioning

confidence: 99%

Section: Expectation Propagation Over Gkd or Svd Factor Graphsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Expectation Propagation on Factor Graphs Based on Matrix Decomposition

Mekhiche

Cipriano²,

Poulliat

2023

ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…Other works [3], [5] proposed to use low complexity MPA to improve the performance-to-complexity trade off. Low complexity EP applied on modified FG has also been studied in MIMO context [7], [8]. The proposed algorithms in this article achieve a more effective scheduling of the EP message exchange applied to less connected scalar factor graphs thanks to a QRD pre-processing suited for SCMA.…”

Section: Introductionmentioning

confidence: 99%

SCMA Detection in MIMO Systems with Low Complexity EP using QRD

Mekhiche,

Cipriano,

Poulliat

2023

2023 12th International Symposium on Topics in Coding (ISTC)

View full text Add to dashboard Cite

This article presents new low complexity message passing based detection algorithms for Sparse Code Multiple Access (SCMA) in Multiple Input Multiple Output (MIMO) systems thanks to the combination of sparser Factor Graph (FG) representation via QR Decomposition (QRD) and enhanced message-passing scheduling. The proposed algorithm achieves better performance to complexity trade-off than state-of-the-art message-passing-based detectors. Our results are presented based on complexity and error rate performance analysis.

show abstract