Semantic-Forward Relaying: A Novel Framework Towards 6G Cooperative Communications

TL;DR

The paper addresses the inefficiency and fragility of traditional relaying in 6G by introducing semantic-forward (SF) relaying, which transmits semantic features rather than raw data to reduce payload and improve robustness. It develops a turbo-style joint source-channel coding framework that iteratively exchanges extrinsic information between channel and semantic decoders, leveraging side information stored in a common knowledge base to achieve lossless recovery under achievable rate constraints. Theoretical analysis yields a rate region with $R_0 ≥ I(X;Y)$, $R_2 ≥ I(Y;U|V)$, and $R_1 ≥ H(X|U,V)$, showing how semantic side information reduces the required rates; practical implementation uses CNN-based semantic encoders/decoders in a bit-wise joint coding system evaluated on image transmission with LDPC codes. Simulation results demonstrate that SF relaying can reduce link payload while maintaining or improving recovered image quality, even under adverse channel conditions, highlighting its potential for robust, semantic-aware 6G cooperative communications.

Abstract

This letter proposes a novel relaying framework, semantic-forward (SF), for cooperative communications towards the sixth-generation (6G) wireless networks. The SF relay extracts and transmits the semantic features, which reduces forwarding payload, and also improves the network robustness against intra-link errors. Based on the theoretical basis for cooperative communications with side information and the turbo principle, we design a joint source-channel coding algorithm to iteratively exchange the extrinsic information for enhancing the decoding gains at the destination. Surprisingly, simulation results indicate that even in bad channel conditions, SF relaying can still effectively improve the recovered information quality.

Figures (4)

• Figure 1: The principle of SF Relaying.
• Figure 2: The system structure of joint source-channel coding for SF relaying.
• Figure 3: Euclidean distance with diverse $\rho$.
• Figure 4: Comparison of the reconstructed image qualities.