An Investigation on Intelligent Relay assisted Semantic Communication Networks

Abstract

With the development of sixth generation (6G) networks, semantic communication is treated as an emerging technology to provide more intelligent communication between two parties for enhance the accuracy and efficiency of communications. Whereas, it demands to merge all physical layer blocks in the conventional communications, which heavily deteriorates the spectrum sparsity problem of wireless communication networks. For this sake, we develop a novel intelligent relay assisted semantic communications, by enhancing communication efficiency, which creates a paradigm for text transmission fusion in 6G networks. In this contribution, the semantic communication system we have designed combines traditional deep learning methods with intelligent semantic relays, based on which, the protocols about minimizing the semantic errors by recovering the meaning of sentences for addressing channel variations issues. Meanwhile, our designed system could be applied in the case of wireless channel deterioration or knowledge background mismatch at the transmitter and receiver, and is also a paradigm for future applications in one-to-many and many-to-many semantic communication. At last, we propose a range of research directions and open challenges for boosting the implementations of relay assisted semantic communications.