Graph Neural Network-based Joint Equalization and Decoding

Jannis Clausius,Marvin Geiselhart,Daniel Tandler,Stephan ten Brink

2024-01-29

Abstract:This paper proposes to use graph neural networks (GNNs) for equalization, that can also be used to perform joint equalization and decoding (JED). For equalization, the GNN is build upon the factor graph representations of the channel, while for JED, the factor graph is expanded by the Tanner graph of the parity-check matrix (PCM) of the channel code, sharing the variable nodes (VNs). A particularly advantageous property of the GNN is the robustness against cycles in the factor graphs which is the main problem for belief propagation (BP)-based equalization. As a result of having a fully deep learning-based receiver, joint optimization instead of individual optimization of the components is enabled, so-called end-to-end learning. Furthermore, we propose a parallel flooding schedule that further reduces the latency, which turns out to improve also the error correcting performance. The proposed approach is analyzed and compared to state-of-the-art baselines in terms of error correcting capability and latency. At a fixed low latency, the flooding GNN for JED demonstrates a gain of 2.25 dB in bit error rate (BER) compared to an iterative Bahl--Cock--Jelinek--Raviv (BCJR)-BP baseline.

Information Theory,Signal Processing

What problem does this paper attempt to address?

The paper attempts to address the problem of how to use Graph Neural Networks (GNN) for Joint Equalization and Decoding (JED) in communication systems to improve performance and reduce latency. Specifically:

1. **Equalization Problem**: Traditional equalization methods based on the Belief Propagation (BP) algorithm, while having lower complexity and latency, are suboptimal in performance, especially when dealing with severe Inter-Symbol Interference (ISI). The paper proposes using GNN for equalization. By constructing GNN on a Factor Graph, it can better handle the ISI problem and has stronger robustness to loops in the graph.

2. **Joint Equalization and Decoding (JED)**: Traditional JED methods usually adopt an iterative approach, such as combining the BCJR algorithm with BP decoding, but this method has high complexity and large latency. The paper proposes a GNN-based JED method. By connecting the factor graphs of equalization and decoding to form a joint factor graph, it achieves end-to-end learning, thereby improving the overall performance of the system.

3. **Low Latency and High Performance**: To further reduce latency and improve performance, the paper also proposes a parallel flooding schedule method. This method significantly reduces latency while improving error correction performance by updating the nodes and edges in the graph in parallel, while maintaining the same resource consumption.

In summary, the main goal of the paper is to improve the equalization and decoding process in communication systems by introducing GNN, providing a more efficient and robust solution, especially in scenarios dealing with ISI and low latency requirements.