Sequence Spreading-Based Semantic Communication Under High RF Interference

TL;DR

This paper tackles robust semantic communication in industrial environments with high RF interference by integrating sequence spreading (SS) with SemCom and introducing a novel Signal Refining Network (SRN). The approach enables scalable multi-user SemCom and reduces dependence on costly online end-to-end training by using offline pre-training and an SRN that refines post-despreading signals. Experiments on GM-based high-RFI channels and the europarl text dataset show that the SRN-assisted SS-based SemCom achieves notable gains, including a 25% BLEU improvement and a 12% increase in semantic similarity over E2E training with the same bandwidth. The method demonstrates strong SU and MU performance under diverse SNR and SF settings, offering practical pathways for robust, scalable SemCom in industrial deployments.

Abstract

In the evolving landscape of wireless communications, semantic communication (SemCom) has recently emerged as a 6G enabler that prioritizes the transmission of meaning and contextual relevance over conventional bit-centric metrics. However, the deployment of SemCom systems in industrial settings presents considerable challenges, such as high radio frequency interference (RFI), that can adversely affect system performance. To address this problem, in this work, we propose a novel approach based on integrating sequence spreading techniques with SemCom to enhance system robustness against such adverse conditions and enable scalable multi-user (MU) SemCom. In addition, we propose a novel signal refining network (SRN) to refine the received signal after despreading and equalization. The proposed network eliminates the need for computationally intensive end-to-end (E2E) training while improving performance metrics, achieving a 25% gain in BLEU score and a 12% increase in semantic similarity compared to E2E training using the same bandwidth.

Figures (7)

• Figure 1: The considered SS-based deep SemCom system model with $M$ users.
• Figure 2: The proposed SRN for SS-based Deep SemCom systems.
• Figure 3: Visualization of the GM-modeled noise and high RFI. The graph illustrates the impulsive and fluctuating nature of RFI commonly observed in industrial environments.
• Figure 4: Performance comparison of baseline SS-based SemCom, SRN-assisted SS-based SemCom, and E2E-trained DeepSC SemCom systems across different $SF$ values with $D = 1$ and $E_s/N_0 = 15$ dB under Rayleigh fading and high RFI channel. For E2E-trained DeepSC, we increase the number of transmitted symbols by a factor of $SF$.
• Figure 5: Illustration of SRN's signal estimation quality (the real part) in the SU scenario with $E_s/N_0 = 15$ dB and $SF = 1$.