Semantic Communication over Sparse Code Multiple Access Systems

Semantic communication focuses on transmitting semantic meanings, with a significant characteristic being the differentiation of the importance of information, thereby improving communication efficiency. Sparse Code Multiple Access (SCMA) is a code-domain Non-Orthogonal Multiple Access (NOMA) technique. Compared to traditional orthogonal access technologies, SCMA can support more user connections and achieve higher system capacity. In this paper, we propose a SCMA-based semantic communication framework, where the hierarchical transmission of SCMA can be effectively integrated with the importance of semantic information. The Message Passing Algorithm (MPA) is the most commonly used detection method in SCMA, but its computational complexity increases significantly as the codebook size grows. This paper introduces an improved SCMA codebook for semantic communication, which adjusts the transmission quality of each information layer in SCMA based on the importance of the semantic information, while significantly reducing detection complexity. Additionally, a low-complexity detection method and its corresponding logarithmic-domain algorithm are proposed. Simulation results show that the improved codebook outperforms traditional designs in Multi-Scale Structural Similarity (MS-SSIM), effectively recovering semantic information at low signal-to-noise ratios, while significantly reducing detection complexity.