SemRAN: Semantic Radio Access Networks with Collective Knowledge Sharing in 6G and Beyond
Hao Xu, Weichao Chen, Keji Zhu, Merouane Abdelkader Debbah · 2024
Semantic communications are considered the next stage for the communications industry, particularly the unique fusion of channel coding and source coding for wireless communications, such as the joint source channel coding (JSCC) with AL/ML trained encoders/decoders. However, the critical challenge of implementing semantic communications methodology in 6G and beyond are two folds: a) The circulation of source and channel information and b) The distribution of shared knowledge both in the public domain and between peers. This paper introduces a generic semantic-ready architecture for future networks by exchanging channel information with source information in purpose-built reasoning plane functions and service-oriented integration architecture. Besides, the paper proposes a novel distributed knowledge exchange scheme resembling social network characteristics. The proposed architecture enables collaborative semantic processing between users and networks with privacy-preserving computing and storage resource configuration using coherent social knowledge bases. With extensive simulations on three datasets, our proposed architecture demonstrates advancements in decoding accuracy, transmission latency, as well as computational efficiency.