Non-Orthogonal Multiple Access Enhanced Multi-User Semantic Communication

TL;DR

This paper introduces NOMASC, a non-orthogonal multi-user semantic communication system that supports diverse source modalities, employs an asymmetric quantizer, and uses neural networks for efficient detection, outperforming existing methods especially at low-to-medium SNRs.

Contribution

It proposes a novel NOMA-based multi-user semantic communication system with an asymmetric quantizer and neural network-based detection, enhancing performance and robustness over existing approaches.

Findings

Outperforms other methods at low-to-medium SNRs

Supports diverse source modalities in multi-user scenarios

Demonstrates high robustness under various conditions

Abstract

Semantic communication serves as a novel paradigm and attracts the broad interest of researchers. One critical aspect of it is the multi-user semantic communication theory, which can further promote its application to the practical network environment. While most existing works focused on the design of end-to-end single-user semantic transmission, a novel non-orthogonal multiple access (NOMA)-based multi-user semantic communication system named NOMASC is proposed in this paper. The proposed system can support semantic tranmission of multiple users with diverse modalities of source information. To avoid high demand for hardware, an asymmetric quantizer is employed at the end of the semantic encoder for discretizing the continuous full-resolution semantic feature. In addition, a neural network model is proposed for mapping the discrete feature into self-learned symbols and accomplishing intelligent multi-user detection (MUD) at the receiver. Simulation results demonstrate that the proposed system holds good performance in non-orthogonal transmission of multiple user signals and outperforms the other methods, especially at low-to-medium SNRs. Moreover, it has high robustness under various simulation settings and mismatched test scenarios.