Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Scenario-Adaptive MU-MIMO OFDM Semantic Communication With Asymmetric Neural Network

Chongyang Li, Tianqian Zhang, Shouyin Liu

TL;DR

This work tackles semantic communication in downlink MU-MIMO-OFDM by proposing a scenario-adaptive, asymmetric SemCom framework that places heavier computation at the base station while keeping the user equipment lightweight. The transmitter uses a CSI- and SNR-aware encoder and a neural precoder with a residual refinement, while the receiver employs a pilot-guided, DS-Conv decoder with dual outputs for image reconstruction and semantic classification. End-to-end training on 3GPP channel models (UMi, UMa, RMa) shows significant PSNR and accuracy gains over DJSCC and SSCC baselines, especially in low-SNR regimes, and demonstrates bandwidth scalability and reduced UE complexity. The results support semantic communication as a practical enhancement for 6G downlink MU-MIMO-OFDM and point toward future extensions to massive MIMO and adaptive semantic resource allocation.

Abstract

Semantic Communication (SemCom) has emerged as a promising paradigm for 6G networks, aiming to extract and transmit task-relevant information rather than minimizing bit errors. However, applying SemCom to realistic downlink Multi-User Multi-Input Multi-Output (MU-MIMO) Orthogonal Frequency Division Multiplexing (OFDM) systems remains challenging due to severe Multi-User Interference (MUI) and frequency-selective fading. Existing Deep Joint Source-Channel Coding (DJSCC) schemes, primarily designed for point-to-point links, suffer from performance saturation in multi-user scenarios. To address these issues, we propose a scenario-adaptive MU-MIMO SemCom framework featuring an asymmetric architecture tailored for downlink transmission. At the transmitter, we introduce a scenario-aware semantic encoder that dynamically adjusts feature extraction based on Channel State Information (CSI) and Signal-to-Noise Ratio (SNR), followed by a neural precoding network designed to mitigate MUI in the semantic domain. At the receiver, a lightweight decoder equipped with a novel pilot-guided attention mechanism is employed to implicitly perform channel equalization and feature calibration using reference pilot symbols. Extensive simulation results over 3GPP channel models demonstrate that the proposed framework significantly outperforms DJSCC and traditional Separate Source-Channel Coding (SSCC) schemes in terms of Peak Signal-to-Noise Ratio (PSNR) and classification accuracy, particularly in low-SNR regimes, while maintaining low latency and computational cost on edge devices.

Scenario-Adaptive MU-MIMO OFDM Semantic Communication With Asymmetric Neural Network

TL;DR

This work tackles semantic communication in downlink MU-MIMO-OFDM by proposing a scenario-adaptive, asymmetric SemCom framework that places heavier computation at the base station while keeping the user equipment lightweight. The transmitter uses a CSI- and SNR-aware encoder and a neural precoder with a residual refinement, while the receiver employs a pilot-guided, DS-Conv decoder with dual outputs for image reconstruction and semantic classification. End-to-end training on 3GPP channel models (UMi, UMa, RMa) shows significant PSNR and accuracy gains over DJSCC and SSCC baselines, especially in low-SNR regimes, and demonstrates bandwidth scalability and reduced UE complexity. The results support semantic communication as a practical enhancement for 6G downlink MU-MIMO-OFDM and point toward future extensions to massive MIMO and adaptive semantic resource allocation.

Abstract

Semantic Communication (SemCom) has emerged as a promising paradigm for 6G networks, aiming to extract and transmit task-relevant information rather than minimizing bit errors. However, applying SemCom to realistic downlink Multi-User Multi-Input Multi-Output (MU-MIMO) Orthogonal Frequency Division Multiplexing (OFDM) systems remains challenging due to severe Multi-User Interference (MUI) and frequency-selective fading. Existing Deep Joint Source-Channel Coding (DJSCC) schemes, primarily designed for point-to-point links, suffer from performance saturation in multi-user scenarios. To address these issues, we propose a scenario-adaptive MU-MIMO SemCom framework featuring an asymmetric architecture tailored for downlink transmission. At the transmitter, we introduce a scenario-aware semantic encoder that dynamically adjusts feature extraction based on Channel State Information (CSI) and Signal-to-Noise Ratio (SNR), followed by a neural precoding network designed to mitigate MUI in the semantic domain. At the receiver, a lightweight decoder equipped with a novel pilot-guided attention mechanism is employed to implicitly perform channel equalization and feature calibration using reference pilot symbols. Extensive simulation results over 3GPP channel models demonstrate that the proposed framework significantly outperforms DJSCC and traditional Separate Source-Channel Coding (SSCC) schemes in terms of Peak Signal-to-Noise Ratio (PSNR) and classification accuracy, particularly in low-SNR regimes, while maintaining low latency and computational cost on edge devices.
Paper Structure (19 sections, 10 equations, 12 figures, 3 tables)

This paper contains 19 sections, 10 equations, 12 figures, 3 tables.

Table of Contents

  1. Introduction
  2. System Model
  3. The proposed framework
  4. Asymmetric Encoder
  5. Precoding Network
  6. Lightweight Decoder
  7. Experimental Results
  8. Performance of Traditional Transmission
  9. Superiority in Low-SNR Regime
  10. Scenario Robustness
  11. Correlation with BLER
  12. Trade-off between Modulation and Coding
  13. Optimal Balance
  14. Performance of Our Framework
  15. Superiority over Baseline
  16. ...and 4 more sections

Figures (12)

  • Figure 1: Comparison of Our Proposed Image Transmission System with Traditional Image Transmission Systems
  • Figure 2: The comparison of three different network blocks
  • Figure 3: The overall architecture of the proposed scheme for wireless image transmission.
  • Figure 4: Architecture of the proposed hybrid learnable residual RZF precoding module.
  • Figure 5: Illustration of the pilot insertion and pilot-guided fusion strategy
  • ...and 7 more figures