基于语义重要度的不等错误保护数据传输机制

摘要：语义通信是一种智能、简约的通信新范式。从系统组成和理论基础两方面，对语义通信和传统通信进行对比，指出其关键区别在于背景知识的引入。将基于信源信道联合设计的语义通信（JSCC-SC）分为面向数据重建和面向任务执行两类，并对其网络架构进行比较。提出一种信源信道分离设计的语义通信系统。该系统通过训练好的神经网络识别数据的语义重要度，并基于语义重要度构造不等错误保护（UEP）传输机制。通过仿真实验证明了其能够提升重要数据的传输可靠性，即降低重要数据传输的误码率，在保持复杂性的同时，进一步提升无线通信的可靠性与有效性。

关键词：语义通信；语义信息论；6G；智能通信

Abstract: Semantic communications is a new intelligent and simple communication paradigm. Semantic communications are compared with traditional communications in terms of system composition and theoretical basis. It is pointed out that the key difference lies in the introduction of background knowledge. The existing semantic communications based on the joint source channel coding (JSCC-SC) research is classified into two categories: data-reconstruction-oriented and task-execution-oriented. Then their network architectures are compared. A data importance-aware semantic communications system with separated source and channel coding is proposed. The system learns the semantic importance of data through a pre-trained neural network, and constructs a transmission mechanism with unequal error protection (UEP) based on the semantic importance. Simulation shows that it can improve the transmission reliability of important data, that is, reduce the bit error rate of important data transmission, and further improve the reliability and effectiveness of wireless communication while maintaining complexity.

Keywords: semantic communication; semantic information theory; 6G; intelligent communication