LDPC codes: tracking non-stationary channel noise using sequential variational Bayesian estimates

TL;DR

This paper introduces a sequential Bayesian method for tracking non-stationary channel noise in LDPC codes, utilizing probabilistic graphical models and a Gamma cluster estimator, demonstrated on real 5G data.

Contribution

The paper develops a novel Bayesian tracking approach for non-stationary noise in LDPC codes using a layered tree cluster graph and Gamma distribution, improving noise estimation accuracy.

Findings

Successfully tracks non-stationary channel noise in real-world 5G data.

Outperforms fixed-noise LDPC decoding methods.

Demonstrates the effectiveness of probabilistic graphical models for dynamic noise estimation.

Abstract

We present a sequential Bayesian learning method for tracking non-stationary signal-to-noise ratios in LDPC codes using probabilistic graphical models. We represent the LDPC code as a cluster graph using a general purpose cluster graph construction algorithm called the layered trees running intersection property (LTRIP) algorithm. The channel noise estimator is a global Gamma cluster, which we extend to allow for Bayesian tracking of non-stationary noise variation. We evaluate our proposed model on real-world 5G drive test data. Our results show that our model is capable of tracking non-stationary channel noise, which outperforms an LDPC code with a fixed knowledge of the actual average channel noise.