Digital Wireless Semantic Communication Enabled by the Discrete Semantic Knowledge Base and Updating
Yongcheng Shen, Zhicheng Bao, Lin Li, Cong Li, Haoming Zhang, Baoling Liu, Dong Chen · 2024
Semantic Knowledge Base (SKB) based on semantic communication attracts increasing attention, since SKB can provide shared knowledge to codecs and improve communication efficiency. However, it has become one of the major challenges for efficient semantic knowledge base communication systems to combat channel noise when facing varying data sources. To solve this problem, this paper proposes a Digital Semantic Communication Enabled by the Discrete Semantic Knowledge Base and Updating (DSC-DSKB) for overcoming channel noise and enhancing communication robustness. Specifically, this paper defines a discrete codebook as a Discrete Semantic Knowledge Base (DSKB) and introduces digital quantization in the Communication process to build the DSC-DSKB. When the DSKB distribution type of the receiver and the transmitter differ, the receiver should update the DSKB to make up the performance gap. Therefore, we propose an update strategy for DSKB, which updates the knowledge base data through the index knowledge heatmap. The experimental results demonstrate that the DSC-DSKB significantly enhances communication performance while keeping the same channel bandwidth as the traditional approach. Meanwhile, the update approach for DSKB effectively maintains high performance while reducing the amount of data delivered.