Neural Enhancement of Factor Graph-based Symbol Detection

Luca, Schmid,; Schmalen, Laurent

doi:10.1109/spawc51304.2022.9834025

Cited by 3 publications

(4 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The BCJR algorithm is applied on a shortened channel with an impulse response of length 4 for complexity reasons. parametrization S = B = 1, P (1,1) = H H , P (1,1) p = P 1 and P…”

Section: Numerical Resultsmentioning

confidence: 99%

“…The message passing algorithm is based on one central message update rule for the VNs and FNs, respectively. In the logarithmic domain 1 , the message updates are…”

Section: Factor Graphs and Marginalizationmentioning

confidence: 99%

“…Due to the band structure of G, many of the factor nodes I n,m (c n , c m ) represent the identity function and can thus be omitted, leading to a sparse graph representation. Note that there exist different factorizations of p(y|c), which consequently lead to distinct graphs [27]. However, the specific structure of the factor graph in Fig.…”

Section: System Model and Preliminaries A System Modelmentioning

confidence: 99%

“…101001899) and in part from the German Federal Ministry of Education and Research (BMBF) within the project Open6GHub (grant agreement 16KISK010). Parts of this paper have been presented at the IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Oulu, Finland, July 2022 in paper [1].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Low-Complexity Near-Optimum Symbol Detection Based on Neural Enhancement of Factor Graphs

Luca

Schmalen

2022

IEEE Trans. Commun.

Self Cite

View full text Add to dashboard Cite

We consider the application of the factor graph framework for symbol detection on linear inter-symbol interference channels. Based on the Ungerboeck observation model, a detection algorithm with appealing complexity properties can be derived. However, since the underlying factor graph contains cycles, the sum-product algorithm (SPA) yields a suboptimal algorithm. In this paper, we develop and evaluate efficient strategies to improve the performance of the factor graph-based symbol detection by means of neural enhancement. In particular, we consider neural belief propagation and generalizations of the factor nodes as an effective way to mitigate the effect of cycles within the factor graph. By applying a generic preprocessor to the channel output, we propose a simple technique to vary the underlying factor graph in every SPA iteration. Using this dynamic factor graph transition, we intend to preserve the extrinsic nature of the SPA messages which is otherwise impaired due to cycles. Simulation results show that the proposed methods can massively improve the detection performance, even approaching the maximum a posteriori performance for various transmission scenarios, while preserving a complexity which is linear in both the block length and the channel memory.

show abstract

“…The BCJR algorithm is applied on a shortened channel with an impulse response of length 4 for complexity reasons. parametrization S = B = 1, P (1,1) = H H , P (1,1) p = P 1 and P…”

Section: Numerical Resultsmentioning

confidence: 99%

“…The message passing algorithm is based on one central message update rule for the VNs and FNs, respectively. In the logarithmic domain 1 , the message updates are…”

Section: Factor Graphs and Marginalizationmentioning

confidence: 99%

Section: System Model and Preliminaries A System Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations